Ecommerce Skills

View saved version →

---
name: checkout-ux
description: Use when reviewing a cart, checkout, or payment form — e.g. "review my checkout", "reduce checkout drop-off", "guest checkout or force an account", "too many form fields", "Apple Pay", "shipping or tax surprise at the end", "coupon field killing conversion", "payment form on mobile", "people abandon at payment".
---

# Checkout UX

Audit the cart, checkout, and payment step — not the product page. Let someone who already wants the item pay without being blocked, surprised, or sent away to hunt for a code.

## When to use

- A cart, checkout, or payment form (URL, screenshot, or walkthrough)
- Guest vs forced account, wallets, coupon field, shipping/tax total, payment errors
- "Review my checkout" or "reduce checkout drop-off"

## When not to use

- Product-page merchandising (social proof, sold-out on the PDP) → `product-page`
- How the list price is written → `pricing`; ticket *price* → `event-pricing`
- Cart-abandonment *email* timing → `post-purchase`
- Ranking store-wide conversion levers → `increase-conversion`

## Inputs

Collect, or state your assumption: the flow (steps vs one page; mobile vs desktop); whether an account is required before the form; the field list; when shipping, tax, and the total appear; coupon field; wallets (Apple Pay, Google Pay, Shop Pay, Link) and their placement; returning-customer path; error states (decline, inventory, 3-D Secure).

## Procedure

1. **Do not block payment behind account creation.** If checkout opens on Login / Register, replace Register with Continue as guest; account optional after payment. At one large retailer this raised the number of customers purchasing 45%. † Asking people to register at the *start of browsing* is a different experiment (long-run purchases up, short-run sales unhurt) and is not a reason to gate the pay form.
2. **Cut fields to what you must have to ship and take payment.** One name field; collapse Address line 2 and company; default billing = shipping; infer city/region from postal code where possible. Typical large-site checkouts ask for more. † Show format hints on restricted fields. Mark required fields more strongly than a tiny asterisk — a color cue beat asterisks, but pair it with text.
3. **Errors after the field or after submit, next to the field.** Not while they type, not only at the top. Immediate in-flow errors were ignored ("completion mode"); ISO-style instant feedback performed worst. After submit, messages beside the bad field beat a dump at the top or bottom.
4. **Steps vs one page: mixed. Field count matters more than page count.** A short labeled sequence is fine; twenty fields on one page is not a win. Do not claim one-page always converts better. †
5. **Show item, shipping, tax, and the total before the pay click.** Drip pricing (base now, fees later) makes people pick a lower base that is often a higher total, then stick with it dissatisfied. Posting tax-inclusive prices cut grocery demand 8%, which is why hiding tax until the register looks tempting; a last-second jump is still a surprise. One clear shipping line is fine; stacking many surcharges reverses the benefit. Depth: `pricing`.
6. **Hide the coupon box behind "Have a code?".** Prompting people who *don't* have a code lowered fairness, satisfaction, and purchase completion. Auto-apply known codes when you can. Depth: `promotions`, `discounts`.
7. **Express wallets above the form, especially on mobile.** Apple Pay, Google Pay, Shop Pay, and Link skip address and card fields. Not a vendor requirement. No peer-reviewed wallet lift on file. †
8. **Trust next to pay: guarantee copy and a cost breakdown beat a badge wall.** One third-party assurance seal raised conversion in a randomized field experiment (9,098 sessions). More than two seals lost effectiveness; they helped more for small retailers and new shoppers. Lock-icon clutter is not a finding. Depth: `store-audit`.
9. **Mobile: numeric keyboard on card/postcode/phone, large tap targets, wallets first.** †
10. **Returning customers: remembered shipping after they choose to identify**, not a login wall before the form. "Already have an account?" is a link, not a gate. †
11. **Recovery copy for decline, inventory, and 3-D Secure.** Don't clear the form. Say what failed and the next step. 3-D Secure is required in some regions — recover through it. If the item sold out during checkout, say so. †

Rank by distance from the rule and whether the miss sits on the pay path. Ship the top two or three, measure, then take the next.

## Output

| # | Issue | Change | Measured effect (study context) | Source |
|---|---|---|---|---|
| 1 | Login/Register before the form | Continue as guest | +45% customers purchasing (one large retailer) † | Spool |
| 2 | Always-on coupon box | "Have a code?" | Lower completion when prompted with no code | Oliver & Shor 2003 |

Close with what the flow already does right. Effects are from the cited contexts, not predictions.

## Common mistakes

- Treating this as a product-page audit — social proof and sold-out variants stay on `product-page`.
- Forcing an account "to capture the email" before payment — take the order first.
- A visible coupon field "because some people have codes" — the people without codes are the ones it hurts.
- Revealing shipping and tax only on the last click — drip pricing; people who stay are less satisfied.
- A wall of lock icons — one recognized seal can help an unknown store; more than two lost force.
- Declaring one-page checkout the winner — mixed; field count is the lever.

## Quick reference

| Check | Rule |
|---|---|
| Account | Guest continue; optional account after pay |
| Fields | Ship-and-pay only; collapse optional; format hints |
| Errors | After field/submit, beside the field |
| Pages vs steps | Mixed; fewer fields, not fewer pages |
| Total | Item + shipping + tax before pay |
| Coupon | Apply-on-demand, not an always-on box |
| Wallets | Above the form, especially mobile |
| Trust | Guarantee + breakdown; at most one or two seals |
| Returning | Opt-in identify, not a login wall |
| Failures | Keep the form; say what happened |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: digital-product-pricing
description: Use when pricing or promoting a digital product sold online — an ebook, online course, template, printable, download, or gated content/lead magnet — e.g. "how much should I charge for my course", "$97 or $100", "should I discount pre-orders", "how do I label the free PDF", "should I let people pay what they want for the ebook", "pricing my Notion template".
---

# Digital product pricing

Digital products have no cost signal from materials or shipping, so the price digits carry more of the quality judgment than for a physical good. This skill is the delta versus `pricing`, `promotions`, and `discounts`. Memberships and recurring plans → `subscription-pricing`.

## When to use

- Choosing the digits of a one-off digital price (rounded vs just-below)
- Pricing a freemium→paid upgrade or wording a lead magnet
- Designing a pre-order or launch offer, or deciding whether to let buyers pick a price

## When not to use

- Recurring plans and memberships → `subscription-pricing`
- Working out *how much* people will pay → `price-research`
- Shared digit, gift, "$0", deadline, and depth rules in isolation → `pricing`, `promotions`, `discounts`

## Inputs

Collect, or state your assumption: product and candidate price; individual vs known publisher; whether the buyer has already used a sample or free tier; launch date; results-driven vs pleasure-driven.

> **Classify the product first.** *Hedonic* = bought for enjoyment, pleasure, or the experience (chocolate, fashion, decor, travel, spa, perfume). *Utilitarian* = bought to do a job, judged on function and value (detergent, appliances, insect repellent, tools, accounting software). If it's mixed, decide by what the buyer is optimizing for at the moment of purchase.

## Procedure

Only the rules that change for digital products. Shared catalogs live in the Depth skills.

1. **Unknown or individual seller → rounded.** No materials or shipping to signal quality, so a rounded price ($100, not $97) is one of the few quality cues available. Just-below reads as price-led. Depth: `pricing`.
2. **Buyer already knows the quality → just-below.** After a free sample or a freemium upgrade prompt, quality is already experienced and a low-price ending is fine — and the left digit must actually drop. Depth: `pricing`.
3. **Distant pre-orders: gift, not discount.** Launch 3+ weeks out → bonus module or extra pack at no cost, not a percentage off. Within about a week, either works. Results-driven courses tolerate only short pre-launch windows. Depth: `promotions`.
4. **Lead magnets: "for $0", never "FREE".** Depth: `promotions`.
5. **Pick-your-price, never pay-what-you-want.** 3–4 fixed options; most pick the lowest, so set it deliberately. PWYW is for donations. Depth: `promotions`.

Shared, unchanged: sale-price just-below, same-left-digit upgrade tiers, deadline not a seat cap, discount ≤ 40% (quality is hard to judge before buying a file). Depth: `pricing`, `discounts`.

## Worked example

Individual creator, LinkedIn-growth course, plan: $97, "FREE download" checklist, pre-orders 6 weeks out at 20% off, pay-what-you-want ebook.

- $97 → **$100** (unknown seller, quality is the decision driver). Keep a just-below figure for launch-week only.
- "FREE download" → **"Download it for $0"**.
- 20% off, 6 weeks out → **bonus module at no cost**; keep the window short for a utilitarian course.
- Pay-what-you-want ebook → **three fixed options**.
- Never "first 50 spots"; use a dated deadline.

## Output

| Item | Value |
|---|---|
| Price | $100 (rounded — individual seller, quality signal) |
| Sale/upgrade price | $79.99 during launch week only |
| Pre-order offer | Bonus module at no cost for orders before launch |
| Lead magnet wording | "Get the checklist for $0" |
| Participative pricing | Ebook at $9 / $15 / $25, not pay-what-you-want |
| Why | one line per choice |

## Common mistakes

- $97 "because creators do it" — for an unknown seller it signals price-led; round it.
- Keeping a pre-order discount for a launch weeks away — a gift wins.
- Calling a lead magnet "FREE" — write "for $0".
- Pay-what-you-want with a floor — still pay-what-you-want; use fixed options.

## Quick reference

| Situation | Do | Depth |
|---|---|---|
| Individual / unknown seller | Rounded | `pricing` |
| Freemium / sample already used | Just-below | `pricing` |
| Pre-order, 3+ weeks out | Gift, not discount | `promotions` |
| Lead magnet | "For $0" | `promotions` |
| Let buyers choose | 3–4 fixed options | `promotions` |
| Discount / urgency | ≤ 40%; dated deadline | `discounts` |

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: discounts
description: Use when deciding how to word, size, or display a price discount, sale price, coupon, or markdown for an online store — e.g. "should this be 25% off or $15 off", "how do I show the sale price", "is 60% off too much", "what kind of coupon code should we send", "how do I write the promo banner".
---

# Discounts

Turn a planned discount into the exact numbers and wording that make it look biggest and most credible. The same discount can be expressed several ways; people react to the size of the numeral they see, not the arithmetic, and very large discounts read as a quality warning.

## When to use

- Choosing between percentage-off and amount-off
- Writing the sale badge, strikethrough price, banner, or email subject for a markdown
- Sizing a discount (how deep is too deep)
- Choosing the format of a coupon code (store-wide vs product-specific)
- Deciding whether to add a time limit

## When not to use

- Setting the full price itself (digits, endings) → the `pricing` skill
- Promotions that aren't price cuts (gifts, bundles, pre-orders, special days) → the `promotions` skill

## Inputs

Collect, or state your assumption:

- Full price `P` and either the sale price `S` or the discount `D` (`D = P − S`). `P` must be a price you actually charged for a reasonable period before the sale — if it wasn't, there is no discount to frame.
- Currency
- Scope: one product, one brand/category, or the whole store
- Where it appears: product page, banner, coupon code, email
- Reason for the sale (seasonal, clearance, launch, customer birthday) — if none, say so
- Audience gender mix, if known (affects price color only)

## Procedure

1. **Check the reference price is genuine.** Showing a struck-through or "was" price that was never really charged is deceptive pricing (US FTC guides, UK CMA, and the EU rule that a "was" price must be the lowest price of the prior 30 days). Confirm `P` was the real selling price before continuing; if you're unsure of the local rule, say so.
2. **Compute the three numbers.** `D = P − S`; `off% = D / P`; `higher% = D / S`. Example: P = 90, S = 60 → D = 30, off% = 33%, higher% = 50%.
3. **Check the depth.** If `off% > 40%`, warn: discounts this deep signal low quality, and at 60%+ they tend to reduce sales, especially for lesser-known brands or products whose quality is hard to judge before buying. Propose capping at 40% and adding a gift or bundle instead (see `promotions`). Exception: explicit clearance or shutdown, stated as the reason.
4. **Pick amount vs percentage by the numeral of the full price.** If `P ≥ 100` in the currency's units → show the **amount** off ("$120 off"). If `P < 100` → show the **percentage** off. The rule is about the size of the printed number, so it holds in any currency.
5. **If percentage: recalculate against the sale price.** State the discount as "was X% higher" where `X = higher%`, because `D / S > D / P`. Cap: if `higher% > 100%`, people ignore the "100" and read only the remainder, so fall back to plain "`off%` off" or the amount.
6. **Keep the price ending consistent.** The sale price should end the way the full price ends ($26.75 → $16.75, not $16.99). If `S` doesn't, adjust `S` to the nearest value with the same ending and recompute step 2. Always show the full price next to the sale price so the subtraction is easy.
7. **Coupons.** Store-wide or most-of-store coupon → **percentage** off ("everything 15% off"), which invites "the more I spend, the more I save". Product- or brand-specific coupon → **amount** off ("$2 off Gillette razors"), which makes the loss of not using it concrete. Send store-wide coupons to segments likely to buy new items, not to hoard. The evidence doesn't give a coupon size; apply the depth check from step 3 to the coupon, and if a coupon can stack on top of a marked-down item, apply it to the combined discount.
8. **Phrase the condition before the reward.** "Spend $100 and get $20 off", not "Get $20 off when you spend $100". The offer then reads as a reward, not a restriction. †
9. **Time limit, if any.** A time limit is fine and should be short (24–48 h), tied to a stated reason (seasonal, birthday), and carry a timezone. Online, time limits mostly raise awareness and visits rather than immediate sales, so don't rely on the deadline alone. Never limit by quantity ("only 500 left"): people who miss out get angry and switch brands; time limits don't cause this.
10. **Price color.** If the audience is mostly men *and* the purchase is routine (the effect disappears when people scrutinize the price, as for expensive or important buys), show *all* prices in red; men judged red prices as much better value, women were unaffected. Mixed red/black backfires, so it's all or nothing.

## Output

| Item | Value |
|---|---|
| Full price | $90.00 |
| Sale price | $60.00 |
| Display | ~~$90.00~~ **$60.00** — was 50% higher |
| Badge / subject line | "Cyber Monday: was 50% higher — ends Monday 23:59 PT" |
| Coupon (if any) | Store-wide → `CYBER15` = 15% off everything |
| Warnings | none / depth > 40% / higher% > 100% |
| Why | one line per choice, citing the rule |

## Common mistakes

- Choosing "$30 off" on a $90 item because "a dollar figure feels concrete" — under 100, the percentage is the bigger numeral. Use 33% off, or better, "was 50% higher".
- Rejecting a 60% discount only on margin grounds — the bigger problem is that it signals poor quality and can cut sales.
- Rounding the sale price to .99 when the full price ends in .00 or .75 — keep the endings the same.
- Treating a countdown as the main sales driver online — it drives clicks and opens, not purchases.
- Using "only N left" for urgency — quantity scarcity angers the people who miss out.
- Striking through a price that was never actually charged — that's deceptive pricing, not a discount.

## Quick reference

| Situation | Do |
|---|---|
| Full price ≥ 100 | Amount off |
| Full price < 100 | Percentage off, stated as "was X% higher" (X = D/S) |
| D/S > 100% | Plain "% off" or amount |
| Discount > 40% | Warn; cap or swap for a gift/bundle |
| Store-wide coupon | % off |
| Product/brand coupon | Amount off |
| Sale price ending | Same as full price |
| Wording | Condition first, then reward |
| Urgency | Short time limit with a reason and timezone; never a unit cap |
| Mostly male audience, routine purchase | All prices red |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: event-pricing
description: Use when pricing or promoting tickets, classes, workshops, bookings, or any event with dates and limited seats — e.g. "how should I price the workshop", "early-bird discount or something else", "should I show sold-out dates", "only 12 seats left banner", "what goes on the ticket page", "how do I sell out this cohort".
---

# Event pricing

Events are experiences: buyers decide on feelings, sold-out dates signal quality, and the wrong kind of scarcity turns the people who miss out against you. This skill is the delta versus `pricing`, `promotions`, `discounts`, and `product-page`. Ticket *price digits* still follow `pricing`; the payment form follows `checkout-ux`.

## When to use

- Setting a ticket, class, or booking price
- Designing pre-sales, early-bird offers, and how dates or seats are shown

## When not to use

- Shared digit, gift, deadline, and depth rules in isolation → `pricing`, `promotions`, `discounts`
- Product-page media and price position in isolation → `product-page`, `product-visuals`
- Finding the price level people will pay → `price-research`
- Checkout/payment fields for the ticket sale → `checkout-ux`

## Inputs

Collect, or state your assumption: the event, ticket price, and date(s); how far away the first date is; how many dates or slots and how many are sold out; any complementary item you can give; whether it is bought for enjoyment or for a functional outcome.

Most events — workshops, concerts, retreats, cooking classes — are hedonic. Training bought for a functional result (exam prep, compliance) is the utilitarian exception.

## Procedure

Only the rules that change for dated, capacity-limited events.

1. **Rounded ticket price.** Experiences are bought on feelings, so $150 beats $149. Just-below only for a promotion price. Depth: `pricing`.
2. **Show sold-out dates — up to 30% of them.** Mark them instead of hiding them; beyond ~30% people abandon. This is the event form of the sold-out-variant rule. Depth: `product-page`.
3. **Deadline, not a seat-count, as the promotional lever.** "Early pricing ends 30 Sept, 23:59 CET" is fine. "Only 12 seats left!" as a marketing hook is not: people who miss a quantity-capped *offer* get angry and switch. Showing *accurate* remaining capacity as inventory (a seat picker, "2 dates left" in a list) is fine. Depth: `discounts`.
4. **Pre-sales: gift, not early-bird discount, when the event is 3+ weeks out.** A bonus session, materials, or the recording at no cost. Within about a week, either works. Hedonic events tolerate long windows; functional ones should keep them short. Depth: `promotions`.
5. **Human-labeled recommendations** for experiences ("others also booked"). Depth: `store-audit`.

Shared, unchanged: discount ≤ 40% and amount-vs-% by the 100 threshold; invented relevant special days; video and gaze-away on the ticket page; price below the image. Depth: `discounts`, `promotions`, `product-visuals`, `product-page`.

## Worked example

Pottery studio, 8-week wheel class, plan: $149, 25% early-bird, "Only 12 seats left", sold-out Saturday hidden, algorithm "recommended" row.

- **$150**, rounded.
- Early-bird 8 weeks out → **glazing-masterclass recording at no cost**, dated deadline, no seat-count banner.
- Show the sold-out Saturday, marked, if sold-out dates stay in the 10–30% band.
- **"Others also booked"** instead of the algorithm label.

## Output

| Item | Value |
|---|---|
| Ticket price | $150 (rounded; hedonic) |
| Dates shown | All 8; 2 marked sold out (25%) |
| Pre-sale (8 weeks out) | Recording at no cost by a dated deadline |
| Urgency copy | Deadline + reason + timezone; no seat-count promo |
| Rejected | 25% early-bird, "Only 12 seats left", hidden sold-out dates |
| Why | one line per choice |

## Common mistakes

- Keeping the early-bird discount because it "gives a real reason to book now" — 3+ weeks out, a gift converts better.
- Dropping "12 seats left" only because it might be inaccurate — accurate or not, unit caps as a promo anger the people who miss out. Use a deadline. Accurate inventory in a picker is a different thing.
- Hiding sold-out dates — they are the strongest quality signal on the page; show them, up to 30%.
- Pricing at $149 to look cheaper — it signals a bargain, not an experience.

## Quick reference

| Situation | Do | Depth |
|---|---|---|
| Ticket price | Rounded | `pricing` |
| Sold-out dates | Show, marked; 10–30% | `product-page` |
| Promo urgency | Deadline, never a seat-count hook | `discounts` |
| Pre-sale, 3+ weeks out | Gift, not early-bird | `promotions` |
| Recommendations | Human label | `store-audit` |

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: increase-aov
description: Use when someone wants bigger orders — "raise average order value", "AOV is flat", "get people to buy more per order", "should we do a free-shipping threshold or a bundle", "how do we sell three instead of two" — for any online store or product page.
---

# Increase AOV

Turn "AOV is $34, make it bigger" into a ranked backlog. Leave the how-to and the papers in the Depth skills. Rank **direct** findings (measured on order size) ahead of **adjacent** (measured on sales or returns) and **indirect** (measured on willingness to pay or deal attractiveness) at equal fit. Every number is what a cited study measured in its own context; do not turn it into a forecast for the user's AOV.

## When to use

- The user names AOV, basket size, units per order, or "sell more per customer" as the goal
- Deciding between a free-shipping threshold, a membership, a bundle, or a coupon

## When not to use

- More visitors buying at all → `increase-conversion`; second orders → `increase-repeat-purchases`
- Cart/checkout field design (guest, wallets, coupon-box distraction) → `checkout-ux`
- The exact wording of a discount → `discounts`; the promotion mechanic in depth → `promotions`

## Inputs

Collect, or state your assumption:

- Products, prices, typical basket (units and value), complements sold
- Whether the main product is naturally bought in multiples
- Current CTA, shipping policy, coupons, returns, any bundle and its wording
- Order frequency (for a membership) and margin (for discount depth)

## Procedure

Walk each check; note whether the setup complies, and if not, what changes. Include applicable levers the user didn't mention; list the rest as considered.

**Direct**

1. **Quantity CTAs.** Multi-unit product → Buy 1 / Buy 2 / Buy 3 instead of "Add to cart". Depth: `product-page`.
2. **Paid free-shipping membership with a minimum order.** A plain threshold without a membership isn't in the evidence — say so if recommending one. Depth: `post-purchase`.
3. **Free, easy returns.** Higher spend per order and more variety. Depth: `post-purchase`.

**Adjacent** (measured on total sales or returns, not basket size)

4. **Coupon type.** Store-wide → percentage; product-specific → amount off. Depth: `discounts`.
5. **Bundle framing.** Wanted product as the gift, "at no cost", not a "special bundle price". Depth: `promotions`.

**Indirect**

6. **Token-fee upgrades** rather than free. Depth: `promotions`.
7. **Flat rate above pay-per-use** where usage varies. Depth: `pricing`, `subscription-pricing`.
8. **Simple add-on pricing.** Rounded, same ending as the base, no service fee. Depth: `pricing`.
9. **Complements around the focal product** in the shot (raises focal-product purchase likelihood, not accessory sales). Depth: `product-visuals`.
10. **Intuitive units** ("about 15 cups", not "250 g"). Depth: `product-copy`.
11. **Rank.** Direct first; then by measured effect and distance from the rule; cheap changes (CTA, wording, coupon type) ahead of programs that need billing or logistics.

## Output

| # | Finding | Change | Measured effect (study context) | Class | Depth |
|---|---|---|---|---|---|
| 1 | Single "Add to cart" on a multi-unit product | Buy 1 / 2 / 3 | Sales up to +28% (37 experiments) | Direct | `product-page` |
| … | | | | | |

Then: what already complies. Each effect was measured on its own; don't add them up. Ship the top two or three, measure, then take the next.

## Common mistakes

- Inventing a lift for a free-shipping threshold ("+10–20%") — the evidence is for a paid membership with a minimum; say so.
- "Special bundle price" wording — gift framing is the measured lever.
- A dollar-off coupon for a store-wide promotion — percentage puts people in "the more I spend" mode.
- Defaulting the bundle in the cart instead of quantity buttons — the measured lever is the CTA format.

## Quick reference

| Lever | Depth | Direct? |
|---|---|---|
| Buy 1 / 2 / 3 | `product-page` | Direct |
| Membership + minimum | `post-purchase` | Direct |
| Free returns | `post-purchase` | Direct |
| Store-wide % coupon | `discounts` | Adjacent |
| Gift-framed bundle | `promotions` | Adjacent |
| Token-fee upgrade | `promotions` | No |
| Flat rate / simple add-ons | `pricing` | No |

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: increase-conversion
description: Use when someone wants more visitors to buy — "conversion rate is low", "how do I get more people to check out", "what should we fix first to lift conversions", "our product page isn't converting", "rank what would move conversion" — for an online store or product page. Cart and payment-form work goes to checkout-ux.
---

# Increase conversion

Turn "conversion is 1.4%, make it higher" into a ranked backlog. This skill does not audit a checkout page and does not restate the task skills. Walk the levers below, note whether the setup already complies, and order the misses by measured effect × distance from the rule. Every number lives in the Depth skill; do not convert it into a forecast for the user's rate.

## When to use

- The user names conversion, checkout *rate*, or "people don't buy" as the goal
- The user shares a store or product page and asks what to fix first

## When not to use

- Cart, checkout, or payment-form drop-off → `checkout-ux`
- Basket size → `increase-aov`; second orders → `increase-repeat-purchases`
- A single element in depth (a photo brief, a price, a discount) → the Depth skill named in that step

## Inputs

Collect, or state your assumption:

- The page or store: URL (fetch it), screenshot, or element-by-element description
- Whether drop-off is on the product page or at cart/payment (if payment, stop and use `checkout-ux`)
- Product, price, variants, sold-out share, review snapshot
- Company size; whether the product is high-tech; any real sustainability practice

> **Classify the product first.** *Hedonic* = bought for enjoyment, pleasure, or the experience (chocolate, fashion, decor, travel, spa, perfume). *Utilitarian* = bought to do a job, judged on function and value (detergent, appliances, insect repellent, tools, accounting software). If it's mixed, decide by what the buyer is optimizing for at the moment of purchase.

## Procedure

Walk each check. Include levers the user didn't mention if they apply; list the ones that don't apply in one line at the end.

1. **Checkout and cart, if they have one.** Guest vs forced account, field count, wallets, total/shipping surprise, coupon-field distraction. Often the largest conversion hole and out of scope for this router. Depth: `checkout-ux`.
2. **Product-page mechanics.** Price below the image; view or purchase counts; sold-out variants visible at 10–30%; quantity CTAs on multi-unit products; options in steps of 2–5. Depth: `product-page`.
3. **Media.** Hedonic → video/GIF, gaze away, first-person hand, shine, spacing. Utilitarian → stills are enough. Depth: `product-visuals`.
4. **Copy.** Three best claims, plain language, intuitive units. Depth: `product-copy`.
5. **Trust.** Cost breakdown; say you're small if low-tech; extraordinary warranty if the category is known but you aren't. Depth: `store-audit`.
6. **Offer.** "For $0" not "free"; wanted product framed as the gift; token fee for upgrades. Depth: `promotions`.
7. **Price and discount.** Quality-led or hedonic → rounded; price-led or promotion → just-below. Discounts ≤ 40%. Depth: `pricing`, `discounts`.
8. **Urgency.** Deadline with a reason and timezone; never a unit cap. Depth: `discounts`.
9. **Reviews.** 4.0–4.5 sells more than a perfect average; reply to every review. Depth: `reviews`.
10. **Rank.** Compare like with like: the studies report sales, conversion, intent, and willingness to pay. A WTP percentage is a weaker surrogate for conversion than a measured sales lift. Cheap, no-inventory fixes ahead of shoots and builds at equal effect. Each effect was measured on its own; don't add them up.

## Output

| # | Finding | Change | Measured effect (study context) | Depth |
|---|---|---|---|---|
| 1 | Forced account before payment | Guest continue | Spool: +45% customers purchasing (one large retailer) † | `checkout-ux` |
| 2 | Price above image | Move below | +35.2% sales (liquor store) | `product-page` |
| … | | | | |

Then: what already complies, and the caveat that effects are from the cited contexts, not predictions. Ship the top two or three, measure, then take the next.

## Common mistakes

- Auditing the checkout page inside this skill — hand it to `checkout-ux`.
- Inventing a lift ("30–50% relative") — quote the study's number and its context only.
- Ranking price position last as cosmetic — it is one of the largest measured effects on the page.
- Treating a 4.9 as an asset to protect — 4.0–4.5 sells more.

## Quick reference

| Check | Depth |
|---|---|
| Cart / payment form | `checkout-ux` |
| Price, social proof, CTA, variants | `product-page` |
| Photos and video | `product-visuals` |
| Claims and units | `product-copy` |
| Cost breakdown, size, warranty | `store-audit` |
| Gifts, "$0", upgrades | `promotions` |
| Digits and markdowns | `pricing`, `discounts` |
| Review average and replies | `reviews` |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: increase-repeat-purchases
description: Use when someone wants more customers to come back and buy again — "repeat rate is low", "improve retention", "get a second order", "customers buy once and disappear", "what should our post-purchase flow do" — for any online store.
---

# Increase repeat purchases

Turn "only 12% buy again" into a ranked backlog. Leave the how-to and the papers in the Depth skills. Several effects show up in retention over years, not in next month's numbers — say so. Every number is what a cited study measured in its own context; do not turn it into a forecast for the user's repeat rate.

## When to use

- The user names repeat rate, retention, lifetime value, or "second order" as the goal
- Reviewing what happens after checkout: emails, box contents, returns, loyalty

## When not to use

- First purchase → `increase-conversion`; order size → `increase-aov`
- Cart and payment-step UX → `checkout-ux`; abandonment *email* timing stays here via `post-purchase`
- Review solicitation and complaint handling in depth → `reviews`; referral mechanics → `referrals`

## Inputs

Collect, or state your assumption:

- Product and how long it takes to judge (instant vs weeks)
- Post-purchase emails: what is sent, when, first question asked
- Box contents (inserts, codes, packaging)
- Return and refund policy; loyalty program (opt-in or automatic, tiers, reward type)
- Abandoned-cart email timing; whether the store replies to reviews

## Procedure

Walk each check; include applicable levers the user didn't mention; list the rest as considered.

1. **Loyalty: automatic, single tier, threshold reward.** Enroll everyone; one rule ("buy 10, get 1 at no cost"). Effect is in multi-year retention. Depth: `post-purchase`.
2. **Feedback request with a positive open-ended first question.** "What did you enjoy most?" before any rating. Depth: `reviews`.
3. **Ask for reviews at 9–13 days**, later for products that take weeks to judge. Measured on review volume, not repeat purchase; it is the delivery vehicle for step 2. Depth: `reviews`.
4. **Handwritten thank-you note, nothing attached.** A discount on the note cancels the effect. Depth: `post-purchase`.
5. **Instant refund when the return ships.** Carrier scan, not after inspection, for items that don't need inspection. Depth: `post-purchase`.
6. **Free, easy returns.** Depth: `post-purchase`.
7. **Free-shipping membership with a minimum order.** Depth: `post-purchase`.
8. **Reply to every review** — measured on ratings and volume, not repeat. Depth: `reviews`.
9. **Abandoned-cart reminder at 24–72 h**, never within the hour. Depth: `post-purchase`.
10. **Referrals as the next step** once repeat customers exist: friend's benefit as the headline. Depth: `referrals`.
11. **Rank.** Note which effects are long-horizon (loyalty) versus immediate (email timing, refund timing). Cheap changes ahead of program builds at equal effect.

## Output

| # | Finding | Change | Measured effect (study context) | Horizon | Depth |
|---|---|---|---|---|---|
| 1 | Opt-in tiered points | Auto-enroll, single "buy 10, get 1" | +29.5% customer value over 5 years | Long | `post-purchase` |
| … | | | | | |

Then: what already complies. Each effect was measured on its own; don't add them up. Ship the top two or three, measure, then take the next.

## Common mistakes

- Inventing a lift ("could double repeat rate") — quote the study's number and its context.
- Building a replenishment sequence first — reasonable, but not in this evidence.
- Making the loyalty program visible but still opt-in — the finding is automatic enrollment.
- Keeping the discount on the thank-you note — it cancels the note's effect.
- Judging the loyalty program on next quarter's numbers — the effect is in multi-year retention.

## Quick reference

| Lever | Depth | Horizon |
|---|---|---|
| Auto-enroll single-tier loyalty | `post-purchase` | Long |
| Positive open-ended first question | `reviews` | Immediate |
| Review ask at 9–13 days | `reviews` | Immediate |
| Handwritten note, no promo | `post-purchase` | Immediate |
| Refund on carrier scan; free returns | `post-purchase` | Immediate / ongoing |
| Abandoned-cart email 24–72 h | `post-purchase` | Immediate |
| Referral ask, friend's benefit first | `referrals` | Next |

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: post-purchase
description: Use when designing or reviewing what happens after checkout in an online store — packaging inserts and thank-you notes, shipping subscriptions and free-shipping thresholds, return and refund policy, loyalty programs, and cart-abandonment or retargeting timing — e.g. "when should the abandoned-cart email go out", "should we offer free returns", "set up a loyalty program", "what goes in the box".
---

# Post-purchase

Design shipping, returns, loyalty, and re-engagement so each order raises the odds of the next. The wins are timing and sincerity: a reminder sent too early does worse than no reminder, and a coupon attached to a thank-you note cancels the note's benefit.

## When to use

- Reviewing a post-purchase or retention plan
- Abandoned-cart timing, return/refund/shipping policy, packaging inserts, loyalty program

## When not to use

- Cart layout, checkout fields, wallets, or payment-step errors → `checkout-ux` (this skill owns abandonment *email* timing, not the checkout surface)
- Asking for reviews or handling complaints → the `reviews` skill
- Referral mechanics → the `referrals` skill
- Site-wide trust signals (warranties, cost transparency) → the `store-audit` skill

## Inputs

Collect, or state your assumption:

- Abandonment-email timing; what's in the box (inserts, codes)
- Return policy: who pays, when refunded, drop-off options; whether items need inspection
- Loyalty structure (tiers, opt-in vs automatic, reward type)
- Average order value; free-shipping threshold

## Procedure

1. **Abandoned-cart reminder at 24–72 hours.** Never within the first hour: that soon, people become *less* likely to buy than with no reminder. The motive fades over days and is gone within a week, so 24–72 h is the window.
2. **Handwritten thank-you note, nothing attached.** A handwritten note in every order (photocopy works; personalization optional). No discount code or promotion on or with it — that reads as a sales tactic and cancels the effect. Strongest on loyal customers. Optionally scent the packaging for memorability.
3. **Free-shipping subscription with a minimum order.** If order frequency supports it, sell a membership for free delivery (members spent more and bought a wider range; a separate uncited report puts it at over 3× †). Require a minimum order (e.g. $20), or members split purchases into many small orders.
4. **Free, easy returns.** Customers expect free, no-questions returns; the long-term result is higher spend per order and broader baskets. Include a label (or print-at-home), accept the original packaging, offer drop-off points. Shoppers benchmark you against Amazon, and returns/refunds are where others fall short. †
5. **Instant refund when the return ships.** Refund on the carrier scan, not after inspection: satisfaction and loyalty rise, return rates don't. If abuse is a concern, limit to items that don't need inspection (books, sealed consumables, low-value goods) and keep the inspection step for items that must be checked for use or damage (electronics, cosmetics, apparel).
6. **Loyalty: automatic, single tier, threshold reward.** Enroll everyone automatically. One rule ("buy 10, get 1 free") is enough — the study didn't test tiers, so don't add them for their own sake: +29.5% customer value over five years, showing up in retention, not the next quarter. Making the reward variable ("your 11th purchase is 20% or 80% off") drove more repeat behavior in a separate opt-in study — its "fewer sign up" trade-off doesn't arise under auto-enrollment, and the combination of the two designs is untested.

## Output

The user's plan as a table: element, keep/change, replacement, rule + number. Then the sequence in order (confirmation → box contents → return policy text → loyalty enrollment → abandonment timing).

## Common mistakes

- A 30-minute abandoned-cart email as "standard" — under an hour it backfires.
- A 10% code on the thank-you card — it undermines the note's sincerity; send offers separately.
- Free returns as "something to test" — they're the expected standard.
- Refund after inspection — refund on ship-back; inspect only where needed.
- Tiers because they look sophisticated — the evidence is for one automatic threshold reward; tiers weren't tested.
- Free shipping with no minimum — members fragment orders.

## Quick reference

| Element | Do |
|---|---|
| Abandoned cart email | 24–72 h after; never < 1 h |
| Thank-you note | Handwritten (copy is fine), no promo attached |
| Shipping subscription | Paid membership + minimum order threshold |
| Returns | Free, label included, drop-off options, same packaging |
| Refund timing | Instant on carrier scan (limit to no-inspection items if abuse is a risk) |
| Loyalty | Auto-enroll, one tier, "buy N get 1", variable reward |
| Packaging | Consider a scent for memorability |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: price-research
description: Use when someone needs to find out how much customers will actually pay before setting or changing a price — e.g. "what's the most profitable price for this", "should I survey my list about pricing", "how do I test willingness to pay", "is Van Westendorp the right method", "we're just copying competitor prices".
---

# Price research

Find the profit-maximizing price with two one-question surveys and a correction step, instead of guessing, copying competitors, or averaging "what would you pay?" answers. The method combines an open question (which people overstate) with a randomized yes/no question (which is unbiased but coarse) to get an accurate demand curve cheaply.

## When to use

- Pricing a new product, plan, or course before launch
- Deciding whether to raise or lower an existing price
- The user proposes a "how much would you pay?" survey, Van Westendorp, or competitor matching

## When not to use

- Choosing the digits or presentation of a price already decided → the `pricing` skill
- Unable to get ~500 completed responses (30 per closed price × 7, the same again for the open survey, plus drop-off): the method needs volume; suggest a price A/B test instead

## Inputs

- Product description, images, or a demo (realism matters)
- Unit cost, reachable audience size, currency
- The lowest and highest price you can imagine charging
- Channels to recruit respondents (site, newsletter, customers, panel)
- Strategy: maximize profit now, volume/reach, or premium positioning

## Procedure

1. **Why not the usual methods.** Averaging open "how much would you pay?" answers overstates by a large margin (often ~50%). Asking "would you buy at $X?" alone is biased by the price acting as a quality cue. Competitor matching ignores your own demand. Van Westendorp gives an "acceptable range", not the profit-maximizing point. The lab-standard auction method (BDM) is accurate but requires selling the real product and is too costly for most teams.
2. **Build the price ladder.** Pick 7 prices with equal steps from the lowest to the highest plausible price ($20, $40 … $140). More steps need more respondents. †
3. **Write two surveys with identical product descriptions.** Survey A (open): "What is the most you would pay for [product]? (in $)". Survey B (closed): "Would you buy [product] for $X? Yes / No", where X is one ladder price chosen at random per respondent. Use a tool that supports random-subset questions (Qualtrics does) and a random link splitter to send 50% to each survey.
4. **Sample size.** At least 30 answers per closed price → 210 for B, the same for A, plus ~80 for drop-off ≈ 500 total. † Work back from your expected response rate: at 5%, a list of 10,000 yields ~500. Recruit existing or prospective customers first; panels are a fallback. Say "one quick question" and offer a reward (coupon, prize draw). Keep it open 2–3 weeks.
5. **Test before launch.** Preview both surveys, open the split link in incognito several times, confirm prices randomize, delete test answers.
6. **Clean.** Drop $0 and absurd answers from Survey A (the script drops anything above 5× the median; adjust with `--max-multiple`). †
7. **Correct.** `corrected = open_answer − mean(open) + mean(closed)`, where `mean(closed)` is the area under the yes-share curve across the ladder (the script uses the trapezoid rule, assuming 100% would buy at $0 and 0% one step above the top price). Each corrected value is an estimate of that respondent's willingness to pay; the distribution, not any single value, is what you use.
8. **Demand and profit.** For each ladder price: share of corrected values ≥ price; buyers = share × audience; profit = (price − unit cost) × buyers. Pick the maximum; the script also lists near-ties (within 10% of the best profit), refuses to recommend a price when none is profitable at the given cost or when any price (or the open survey) has fewer than 30 answers, and stops on any closed-survey answer that isn't a recognisable yes/no. *Audience* is the number of people who will actually see the offer (expected launch reach), not the size of the survey list. *Unit cost* is the marginal cost per sale; for digital products that is near $0 plus any per-sale fees, and the profit-maximizing price then equals the revenue-maximizing one. If two adjacent prices are within ~10% on profit, treat them as a tie and let strategy (step 9) or a live A/B test decide.
9. **Override by strategy.** A premium brand may choose a higher price at lower profit to attract future buyers; a reach-first launch may go lower. Say which you chose and why. †

Run `scripts/wtp_calc.py --open open.csv --closed closed.csv --cost 5 --audience 8000` (`open.csv`: column `wtp`; `closed.csv`: columns `price,answer`). Example output on synthetic data (230 open answers, 30 per price on a 7-step ladder):

```
Open survey: 230 answers, 3 dropped (<=0 or >5x median)
Closed survey: 210 answers across 7 prices
       30.00  yes  96.7%  n=30
       60.00  yes  86.7%  n=30
       90.00  yes  40.0%  n=30
      120.00  yes  23.3%  n=30
      150.00  yes  13.3%  n=30
      180.00  yes   3.3%  n=30
      210.00  yes   0.0%  n=30

mean_open   = 125.59
mean_closed = 94.00
correction  = mean_closed - mean_open = -31.59  (applied to every open answer)
corrected mean WTP = 94.00

     price  share willing     buyers       profit
     30.00          83.7%       6696       167401
     60.00          63.0%       5040       277181
     90.00          42.3%       3383       287577
    120.00          26.9%       2150       247225
    150.00          17.2%       1374       199295
    180.00          11.9%        952       166520
    210.00           7.5%        599       122819

Profit-maximizing price: 90.00  (profit 287577 at cost 5, audience 8000)
Near-ties (within 10% of the best profit): 60.00 — treat as tied; decide by strategy or a live A/B test.
Override for strategy: a premium brand may pick a higher price; a reach-first launch a lower one.
```

## Output

| Item | Value |
|---|---|
| Price ladder | 7 prices, equal steps, range stated |
| Survey A / B wording | exact text |
| Sample plan | ≥30 per price; target ~500; channels; incentive; duration |
| Correction constants | mean_open, mean_closed, shift |
| Demand & profit table | per ladder price |
| Recommended price | with the strategy override, if any |

## Common mistakes

- Averaging "what would you pay?" answers — overstated; needs the closed-survey correction.
- Reaching for Van Westendorp — gives a range, not the profit-maximizing point.
- Uneven ladders ($20, $50, $100) — steps must be equal for the correction to hold.
- Too many prices for the sample (10 prices with 276 answers) — under 30 per price; use 7 and run longer.
- Thin product description — respondents price what they imagine; show images or a demo.
- Picking the top-profit price blindly when the brand is premium — state the strategic override.

## Quick reference

| Step | Rule |
|---|---|
| Ladder | 7 prices, equal steps, generous range |
| Split | 50/50 random to open vs closed; random price within closed |
| Sample | ≥30 per closed price, ~500 total |
| Clean | Drop 0 and outliers (>5× median) |
| Correct | open − mean(open) + mean(closed) |
| Choose | max (price − cost) × audience × share ≥ price, then apply strategy |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 2 — file references are metadata, not package contents.

View saved version →

---
name: pricing
description: Use when setting or reviewing the full price of a product, plan, or add-on for an online store — e.g. "should this be $199, $200, or $197.50", "what should the add-on cost", "should we do a flat rate or pay per use", "where should the price go on the page", "how should I price this for a Japanese audience".
---

# Pricing

Choose the digits, structure, and placement of a full price so it sends the signal you want. The last digits of a price tell people whether to think about quality or about cost, complex prices get read as expensive, and where the price sits on the page changes how big it feels.

## When to use

- Deciding whether a price should be rounded ($200), just-below ($199), or precise ($203)
- Pricing add-ons, upgrades, and per-unit components
- Choosing flat-rate vs pay-per-use
- Placing and coloring prices on a product page
- Setting an opening price for a negotiation or auction

## When not to use

- Working out *how much* people will pay → the `price-research` skill
- Wording a sale or coupon → the `discounts` skill

## Inputs

Collect, or state your assumption:

- Product, candidate price, currency
- What the buyer is optimizing for (quality/risk vs price); brand positioning (premium vs value)
- Whether this is a promotion price or the everyday price
- Whether the price will be negotiated, and whether the counterparty is an expert
- Market (Japan/China change the lucky ending)
- Audience gender mix, if known

> **Classify the product first.** *Hedonic* = bought for enjoyment, pleasure, or the experience (chocolate, fashion, decor, travel, spa, perfume). *Utilitarian* = bought to do a job, judged on function and value (detergent, appliances, insect repellent, tools, accounting software). If it's mixed, decide by what the buyer is optimizing for at the moment of purchase.

## Procedure

1. **Default to rounded** ($80, $200, $35) when the decision runs on feeling or quality: premium or luxury goods, hedonic products, most services, a premium brand image, or when you want an effortless choice. Note the boundary: the roundedness study found *non-rounded* prices did better when buyers evaluate analytically (typical utilitarian purchases) — for those, rounded earns its place only as a low-price-image avoidance (a 99-ending signals bargain) or for simplicity, not from that study. The risky-purchase examples (medical, safety) are a practitioner extension. † Also when the price may be bid *up* (a hot auction, tight housing market). When in doubt, round.
2. **Use just-below** ($199, $4.99) only when the left digit actually drops relative to the price you'd otherwise charge ($2.99 instead of $3.00 works; $3.49 instead of $3.50 doesn't) and one of these holds: it's a promotion price; buyers are price-focused; you want a low-price image; the item is cheap and low-risk. Lower left digits (1, 4) feel like bigger savings than high ones (7).
   - **Four figures or more** (common in ¥, ₹, ₺): repeat a low digit in the last three places instead — ¥2,111 beats ¥1,999. †
   - **Japan and China:** an 8 ending is the convention (prevalence evidence, not a measured lift), not 9.
   - **Upgrades:** if you want people to step up to a dearer option, keep both prices on the same left digit ($40.50 → $48.50, not $39.99 → $48.50).
3. **Use precise** ($2,865; $53,832/month) when you need to signal expertise and calculation in a big decision (B2B quotes, renovations) or when the buyer will negotiate *down*: precise anchors pull counteroffers into smaller steps, so you get fewer but higher offers. With expert buyers, stay moderately precise ($23,750, not $23,721.85) or justify the number with a cost breakdown. Evidence for precise as a "slightly above rounded" everyday price is weaker than for just-below.
4. **Prefer a flat rate** that covers what most customers want ($50/month, all-in; unlimited revisions), priced *above* the pay-per-use equivalent — a $6,000 fixed fee beat $50/h × 100 h. Keep pay-per-use only as a low-commitment entry tier.
5. **Keep components simple** when you must price separately: rounded amounts ($0.10/min, not $0.07), the same ending across components, no service charge tacked on, options easy to compare side by side. Add-ons follow the same rule ($15, not $12.35).
6. **Place the price below the product** image, not above; the same price reads lower.
7. **Color:** mostly male audience and a routine, low-scrutiny purchase → every price in red; the effect disappears for expensive or important purchases. Mixed red and black backfires. Audience unknown or mixed → black.
8. **Hand off:** sale prices → `discounts`; validating the level → `price-research`.

## Output

| Item | Value |
|---|---|
| Recommended price | $200 |
| Digit style | Rounded — premium hedonic product, quality is the decision driver |
| Add-ons / components | Monogramming $15 (rounded, same ending as core price) |
| Structure | Flat rate / pay-per-use / n.a. |
| Placement & color | Price below the product image; black (mixed audience) |
| Alternatives rejected | $199 (just-below reads as price-led; save for promotions), $197.50 (precision signals nothing here) |

## Common mistakes

- Picking $195 or $225 "for confidence" with no rule — the choice is rounded *because* quality is the driver; state the reason.
- Using $9.99 everywhere by habit — just-below cues price-focus and undermines a premium image; use it for promotions and price-led categories.
- $1,999 in a four-figure currency — the just-below effect is overridden; use ¥2,111-style low repeating digits.
- Setting the upgrade tier across a left-digit boundary ($39.99 vs $48.50) when you want people to upgrade.
- Underpricing the flat rate to match pay-per-use — people will pay a premium for the cap on spend.
- Fiddly component prices ($4.61/GB plus a $6 fee) — people assume you're the expensive option without calculating.

## Quick reference

| Situation | Digits |
|---|---|
| Premium, hedonic, risky, service, unsure | Rounded |
| Promotion, price-led, cheap and low-risk | Just-below (left digit must drop) |
| Four-figure prices | Low repeating last digits (2,111) |
| Japan / China | 8 ending |
| Want upgrades | Same left digit on both tiers |
| B2B quote, negotiated down | Precise; moderate with experts |
| Bid up (auction) | Rounded |
| Usage-based service | Flat rate, priced above pay-per-use |
| Components / add-ons | Rounded, same endings, no service fee |
| Page layout | Price below the product |
| Mostly male audience, routine purchase | All prices red |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: product-copy
description: Use when writing or rewriting a product description, product title, benefit list, feature bullets, or on-page product messaging for an online store — e.g. "write a description for this product", "rewrite my product copy", "how many benefits should I list", "should I say 500g or 25 servings", "name this color", "should we mention it's lab-tested".
---

# Product copy

Write product descriptions that people find persuasive, easy to grasp, and worth paying more for. The rules change how copy is *structured* and *measured* — number of claims, units, names, register — which move purchase intent and willingness to pay more than adjectives do.

## When to use

- Writing or rewriting a description, title, bullets, or hero message
- Choosing which features to mention, how many, in what units and register

## When not to use

- Auditing the whole page → `product-page`; photo/video direction → `product-visuals`
- Discount or promotion wording → `discounts` / `promotions`

## Inputs

Collect, or state your assumption:

- Product facts: materials, quantity, specs, ingredients or components (any branded?)
- The full list of candidate benefits
- Audience and how technical they are
- Where the copy sits and the length limit

> **Classify the product first.** *Hedonic* = bought for enjoyment, pleasure, or the experience (chocolate, fashion, decor, travel, spa, perfume). *Utilitarian* = bought to do a job, judged on function and value (detergent, appliances, insect repellent, tools, accounting software). If it's mixed, decide by what the buyer is optimizing for at the moment of purchase.

## Procedure

1. **Pick three claims.** Rank the benefits; keep the top three (three claims were 10.4% more persuasive than four; skepticism starts at the fourth). Move the rest to a spec table.
2. **Best features only.** Don't pad with merely good features — the impression sinks to the weakest claim.
3. **Units people use.** Uses or pieces (30 washes, 8 cans), not standardized measures (750 g, 96 oz). Raises product evaluations and willingness to pay.
4. **Smaller units for precision.** Drop to the next unit down: 180 seconds, not 3 minutes; 180 minutes, not 3 hours; 32 oz, not a quarter gallon. Reads more precise and credible; strongest for time.
5. **Evocative color names.** "Midday sky blue", not "light blue".
6. **Feature branded components.** A branded ingredient or part (Gore-Tex, a licensed enzyme blend) goes up front; demand up to 40% higher. †
7. **Register by product type.** Utilitarian → science and effectiveness language works. Hedonic → avoid it ("lab-formulated" on a treat: up to 30% less likely to be chosen). †
8. **Plain language.** Unknown terms make the product seem worse and pricier; define or drop. †
9. **Describe touch** where feel matters (textiles, skincare) — the top online clothing pain point.
10. **Rational up, emotional down** in vertical layouts.
11. **Frame as an experience** (coffee with friends, not bean origin) when you want reviews. †

## Output

1. Headline and body within the length limit.
2. Rules table:

| Rule | Where |
|---|---|
| Three claims | stain removal, cold water, gentle on skin |
| Intuitive units | "30 washes" for "750 g" |
| Branded ingredient | licensed enzyme blend, first sentence |

3. What was cut and where it can go.

## Common mistakes

- Working every supplied benefit in — six benefits read as a pitch; pick three.
- Leading with pack weight or volume — lead with uses.
- Keeping the supplier's color name.
- "Lab-tested" on a hedonic product.

## Quick reference

| Element | Rule |
|---|---|
| Claims | Exactly three, best only |
| Quantity | Uses/pieces, not grams/ounces |
| Precision | Smaller unit |
| Colors | Evocative names |
| Components | Name the brand |
| Register | Science only if utilitarian |
| Vocabulary | Plain |
| Texture | Describe if feel matters |
| Layout | Rational up, emotional down |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: product-page
description: Use when reviewing, auditing, or designing a product page, landing page, or offer page for an online store — e.g. "audit my product page", "look at this landing page and tell me what's wrong", "why isn't this page converting", "how should I show variants / sold-out options / the price", "should I show how many people viewed this", or a URL/screenshot of a page that sells one thing.
---

# Product page

Audit a product page against peer-reviewed findings and return prioritized changes with expected effects. Layout, counts, and choice structure make a product feel wanted and worth the price before the copy does.

## When to use

- A user shares a product page (URL, screenshot, or description) and asks what to change
- Designing a new product page template, or deciding how to show variants, stock, social proof, CTAs, or price

## When not to use

- A cart, checkout, or payment form → `checkout-ux`
- Writing the description → `product-copy`; briefing photos/video → `product-visuals`
- Discount framing → `discounts`; setting the price → `pricing`

## Inputs

Collect, or state your assumption:

- The page: URL (fetch it), screenshot, or element-by-element description
- Product and price; whether it is naturally bought in multiple units (consumables, cans, candles)
- Number of options and how many are sold out
- View and purchase counts and the views-to-purchase rate, if tracked
- Company size

> **Classify the product first.** *Hedonic* = bought for enjoyment, pleasure, or the experience (chocolate, fashion, decor, travel, spa, perfume). *Utilitarian* = bought to do a job, judged on function and value (detergent, appliances, insect repellent, tools, accounting software). If it's mixed, decide by what the buyer is optimizing for at the moment of purchase.

## Procedure

1. **Social proof numbers.** Show a view or purchase count (purchase intent up to ~58% higher). Views while the numbers are small; once views run into the thousands (e.g. 15,600), switch to the purchase count, because large view counts are hard to process. If views→purchases ≥ 5%, show both. Change what the badge counts; don't remove it.
2. **Sold-out options.** Visible and marked sold out (sales up to 31% higher — measured among shoppers who hadn't settled on a specific option). Keep the sold-out share at 10–30%; above that, hide some or restock.
3. **Quantity CTAs.** Multi-unit product → Buy 1 / Buy 2 / Buy 5 buttons instead of "Add to cart" (+14% conversion average, sales up to +28%). Start with three; test quantities.
4. **Choice partitioning.** More than a handful of options → steps of 2–5 similar options, technical before aesthetic, big decisions first. If one attribute alone has more than 5 values (8 colors), group them by similarity (warm / cool / neutral) and let people pick the group first. Never one grid of every combination.
5. **Add-on pricing.** Rounded prices with the same ending as the base ($22 → $32), or a flat rate.
6. **Price position.** Below the product image (reads ~9% lower; store sales +35.2%).
7. **Copy (depth in `product-copy`).** Three claims max; best features only; intuitive units; plain language; science wording only if utilitarian. †
8. **Visuals (depth in `product-visuals`).** Hedonic → video; models look away; first-person hand touching the product; a shine; graspable product toward the right; products spaced apart; complementary props.
9. **Trust.** Small business → say so (unless high-tech). Extraordinary warranty when the category is known but the brand isn't.

Rank by expected effect × distance from the rule. Cart and payment-step issues are out of scope here; Depth: `checkout-ux`.

## Output

| # | Issue | Change | Measured effect (study context) | Source |
|---|---|---|---|---|
| 1 | Single "Add to cart" on a multi-unit product | Buy 1 / 2 / 5 buttons | +14% conversion | Duke & Amir 2023 |

Close with what the page already does right.

## Common mistakes

- Deleting a large view count because it "looks fake" — switch it to purchases.
- Greying out sold-out options only for email capture — the point is the quality signal; watch the 30% ceiling.
- Splitting a variant grid but leaving 8+ options in one step — 2–5 per step.
- Ranking price position last as cosmetic — it moves perceived price and sales.

## Quick reference

| Check | Rule |
|---|---|
| Social proof | Views while small → purchases once views hit thousands; both if ≥ 5% conversion |
| Sold-out | Show, marked; 10–30% of options |
| CTA | Quantity buttons for multi-unit products |
| Options | Steps of 2–5; big decisions first |
| Add-ons | Rounded, same ending, or flat rate |
| Price | Below the product |
| Claims | Three, best only |
| Media | Video if hedonic; hand, gaze away, shine, spacing |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: product-visuals
description: Use when planning, briefing, reviewing, or selecting product photos, product video, hero images, or ad creative for an online store — e.g. "write a shot list for our photographer", "do we need video for this product", "how should the model pose", "which product image should we use", "brief for the product shoot", "why does this image underperform".
---

# Product visuals

Produce a photo/video brief in which each shot is tied to a finding — most with a measured effect, a few marked † as practitioner tips. Format, gaze, the viewer's sense of touching the product, and composition (shine, spacing, side, props) move sales and willingness to pay more than lighting taste does.

## When to use

- Writing a shot list or creative brief; choosing video vs stills or between candidate images
- Reviewing why a product image or ad creative underperforms

## When not to use

- Page layout, CTAs, variants, price position → `product-page`
- Description text → `product-copy`

## Inputs

Collect, or state your assumption:

- The product; can it be picked up and held? Is it premium?
- One main function (energy drink) or many?
- Worn, and by whom (womenswear vs menswear)?
- Where the visuals run (product page, ads, social); audience reading direction

> **Classify the product first.** *Hedonic* = bought for enjoyment, pleasure, or the experience (chocolate, fashion, decor, travel, spa, perfume). *Utilitarian* = bought to do a job, judged on function and value (detergent, appliances, insect repellent, tools, accounting software). If it's mixed, decide by what the buyer is optimizing for at the moment of purchase.

## Procedure

1. **Format.** Hedonic → video, GIF, or slideshow is required (willingness to pay 79% higher). Utilitarian → stills are enough. Premium hedonic → add a slow-motion action shot (+22% sales) †. Video should feel personal and human, not corporate. †
2. **Gaze.** Hedonic → model looks away from the camera (sun hat: +30% sales). Informational or negative-emotion messages → model looks at the camera.
3. **Touch.** A first-person hand touching, holding, or using the product (+16% purchase likelihood).
4. **Side.** Graspable product toward the right of the frame, angled for a right hand. If it can't be held, add a graspable object beside it (handedness doesn't change with reading direction — don't flip this). Future-oriented products right, vintage left; flip *that* for right-to-left readers.
5. **Shine.** Light or polish for a visible reflection; a digital reflection is fine only if it stays representative of the real finish (a matte product shown glossy misleads buyers). Never brief "no reflections" (+23% attractiveness, +32% purchase likelihood).
6. **Spacing.** Clear empty space between products in multi-product shots and grids (liking +11.4%, sales nearly doubled). In ads, wide margins with related products clustered in the center. †
7. **Multiples.** Product bought for one specific function (an energy drink for energy, a stain remover) → several copies together (+10.7% perceived effectiveness) †. Not a rule for treats or decor.
8. **Props.** Complementary items around the focal product, varied in shape and color; the props' own sales are unaffected.
9. **Context.** Womenswear → a fitting setting, not white background. Menswear → irrelevant.
10. **Saturation.** High to look bigger, low to look smaller. Pretest candidates with a poll where cheap. †

## Output

| # | Shot | Direction | Rule | Effect |
|---|---|---|---|---|
| 1 | Hero video, 8–12 s | Truffle broken in slow motion | Hedonic → video; premium → slow-mo | WTP +79% |
| 2 | Hand shot | First-person hand lifting a truffle, box on the right | Touch; graspable right | +16% |

Then list what *not* to do for this product (e.g. no direct gaze, no matte finish).

## Common mistakes

- Making video "secondary" for a hedonic product — it is the highest-effect item.
- Briefing "avoid reflections" — shine raises attractiveness and purchase intent.
- Centering a graspable product — it belongs toward the right.
- A single can for a single-function product — show several.

## Quick reference

| Element | Rule |
|---|---|
| Format | Hedonic → video; utilitarian → stills |
| Premium hedonic | Slow-motion action shot |
| Gaze | Away for hedonic; at camera for info or negative emotion |
| Hand | First-person hand touching |
| Position | Graspable product right (never flipped); future-right/past-left flips for RTL |
| Finish | Visible shine |
| Layout | Space apart; wide ad margins |
| Single-function | Multiple copies |
| Props | Complementary, varied |
| Womenswear | Contextual background |
| Size cue | High saturation = bigger |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: promotions
description: Use when designing a promotion that isn't just a price cut — pre-order incentives, free gifts, bundles, "buy X get Y", launch offers, sale events, pick-your-price, urgency mechanics — e.g. "how do we get pre-orders", "gift or discount", "what should the Black Friday hook be", "should we cap it at 500 units", "how do we word the free bonus".
---

# Promotions

Pick the promotion mechanic and the framing that move sales without training customers to wait for markdowns. Gifts, bundles, invented occasions, and small token fees often beat a straight discount, and the wording of the same offer changes how lucky it feels.

## When to use

- Launch and pre-order campaigns
- Bundles, free gifts, buy-one-get-one, upgrades
- Seasonal or event promotions and their theme
- Urgency mechanics (deadlines, unit caps, sold-out labels)
- Anything offered "free"
- Loyalty or promotional rewards

## When not to use

- Sizing or wording a plain price reduction → the `discounts` skill
- Setting the everyday price → the `pricing` skill

## Inputs

Collect, or state your assumption:

- Product, price, and any complementary product you also sell
- Goal: pre-orders, volume, new customers, clearing stock, returns reduction
- Launch date (if pre-order) and today's date
- Whether buyers typically take one item or a basket
- Brand strength and whether quality is easy to judge before buying

> **Classify the product first.** *Hedonic* = bought for enjoyment, pleasure, or the experience (chocolate, fashion, decor, travel, spa, perfume). *Utilitarian* = bought to do a job, judged on function and value (detergent, appliances, insect repellent, tools, accounting software). If it's mixed, decide by what the buyer is optimizing for at the moment of purchase.

## Procedure

1. **Pre-orders: gift or discount by launch distance.** Launch 3+ weeks out → give a gift (a complementary product, extra content); it beats a discount by a wide margin. Launch within about a week → gift or discount both work. Keep the pre-order window short for utilitarian products; hedonic products tolerate long windows.
2. **Bundles: make the wanted product the gift.** Pair the product with a complement and frame the *target* product as the gift: "Buy the scale, get the fitness tracker at no cost" — shown to people searching for the tracker. Only for items of comparable value (not "buy a case, get a phone").
3. **Frame the bundle as a gift, not a special price.** "Buy X, get Y at no cost" cut returns by up to half compared with "X and Y for a special price". The gift framing is what matters, not the word.
4. **Wherever the word "free" would appear, write "for $0".** The study tested the numeral: "$0" reminds people of money not spent; "free" reads as a giveaway with no value. "At no cost" is an untested paraphrase — use it only where "$0" can't work grammatically. (The two studies used different wording; combining gift framing with "$0" is the sensible default, but the combination itself wasn't tested.)
5. **Upgrades: charge a token fee.** A larger size or add-on for a token fee (¥0.1 in the study) was rated a better deal than the same upgrade free (39% better; 17.1% more chose the larger size). The tiny price gives people a reference point to compare the benefit against.
6. **Tie the promotion to an invented, fitting occasion.** "National Picnic Day Sale" beat "Annual One Day Sale" by 25% for a picnic retailer. The day must be creative and relevant; Father's Day or a generic sale doesn't count. Examples: founding day, customer anniversary/birthday, World Cotton Day (Oct 7, apparel), World Radio Day (Feb 13, electronics), World Chocolate Day (Jul 7), World Pasta Day (Oct 25), International Coffee Day (Oct 1).
7. **Urgency: deadline, never a unit cap.** A short, justified deadline with a timezone is fine ("ends Sunday 23:59 CET, birthday week"). "Only 500 available" makes people who miss out angry and drives them to competitors. Online, deadlines mostly raise engagement rather than immediate sales — pair them with a real hook.
8. **Sold-out options: keep 10–30% visible.** Mark sold-out variants as sold out instead of hiding them; sales rose up to 31%. Above 30% sold out, hide some or restock — the effect reverses.
9. **Pick-your-price, not pay-what-you-want.** Offering 3–4 fixed price options ($25/$35/$50/$100) raised purchases; most choose the lowest, so set it deliberately. Don't use it for basket purchases (fashion, groceries). Pay-what-you-want (any amount, including zero) reduces sales — only for donations.
10. **Rewards: make them uncertain.** "20% or 80% off your next order" outperforms a fixed reward on repeat behavior, even if fewer sign up at first.

## Output

| Item | Value |
|---|---|
| Mechanic | Pre-order gift: smart scale bundled, tracker framed as the gift |
| Headline | "Order the scale, get the tracker for $0 — launch offer ends 30 Sept 23:59 CET" |
| Why | Launch 9 weeks out → gift beats discount; target product as gift; "$0" not "free"; deadline not unit cap |
| Avoid | 15% off (rational, deferred reward); "only 500 units"; "FREE" |
| Test | Gift vs 15% off on a split of the launch list |

## Common mistakes

- Offering a discount for a launch that is weeks away — a gift wins; discounts are boring deferred savings.
- Dropping "FREE" without replacing it — the replacement is "for $0" / "at no cost".
- Bundling with "extra 10% off the pair" — frame the searched-for product as the free item instead.
- Capping units for urgency — use a deadline; unit caps cause anger and switching.
- Hiding all sold-out variants — keep some visible (≤30%) as a quality signal.
- Making the upgrade free — a token fee makes it feel like a better deal.
- Pay-what-you-want "to build goodwill" — it lowers sales; use 3–4 fixed options.

## Quick reference

| Situation | Do |
|---|---|
| Pre-order, launch 3+ weeks away | Gift, not discount |
| Pre-order, launch within a week | Gift or discount |
| Bundle | Wanted product = the gift; "get Y at no cost", not "special price" |
| Anything free | "For $0" / "at no cost" |
| Size/add-on upgrade | Charge a token amount (1¢) |
| Sale theme | Invented, relevant special day |
| Urgency | Short justified deadline; never a unit cap |
| Sold-out variants | Show 10–30% as sold out |
| Participative pricing | 3–4 fixed options; never pay-what-you-want |
| Loyalty/promo reward | Variable, uncertain amount |

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: raise-prices
description: Use when someone wants to charge more without losing sales — "raise prices", "charge more", "customers say we look cheap", "move upmarket", "will a higher price hold", "justify a higher price", "competitors charge more than us" — for any product or plan sold online.
---

# Raise prices

Turn "we're at $9.99 and want $14" into a ranked backlog of what must change on the product and the page for the higher price to hold. Willingness to pay is moved by how the price is written, the unit, what the buyer sees, and which quality signals are present — not by the number alone. Leave the how-to and the papers in the Depth skills. Do not forecast the user's sales at the new price.

## When to use

- The user wants a higher price and asks how to make it stick
- Customers or the founder call the product "cheap-looking"
- A price above the category norm needs justification

## When not to use

- Working out *what* the price should be → `price-research`
- Wording a markdown → `discounts`; more buyers at the current price → `increase-conversion`

## Inputs

Collect, or state your assumption:

- Product, current price, target price, competitor range, currency
- Company size; whether the product is high-tech
- Branded ingredient or component; real sustainability practice
- How quantity is expressed now, and what one "use" is
- Current media, copy, price position, add-on and plan prices

> **Classify the product first.** *Hedonic* = bought for enjoyment, pleasure, or the experience (chocolate, fashion, decor, travel, spa, perfume). *Utilitarian* = bought to do a job, judged on function and value (detergent, appliances, insect repellent, tools, accounting software). If it's mixed, decide by what the buyer is optimizing for at the moment of purchase.

## Procedure

Walk each check; include applicable levers the user didn't mention; list the rest as considered.

**The number**

1. **Rounded list price.** Just-below only for a promotion price. One move vs stages isn't tested; a single move is the practitioner default. † Depth: `pricing`.
2. **Simple components, same ending.** Add-ons and plans rounded like the main price. Depth: `pricing`.
3. **Flat rate above pay-per-use** where usage varies. Depth: `pricing`, `subscription-pricing`.
4. **Announcing the increase.** If you pre-announce it, compute the percentage against the old price so the honest number is the larger one. Existing subscribers usually need advance notice. Depth: `discounts`.
5. **Validate the level** with `price-research` if the target is a guess.

**The unit and the words**

6. **Concrete units** (applications, washes, days), smaller unit for precision. Depth: `product-copy`.
7. **Branded ingredient in the headline.** † Depth: `product-copy`.
8. **Register by product type.** Science/effectiveness language only if utilitarian. † Depth: `product-copy`.
9. **Three best claims, plain words.** Depth: `product-copy`.

**What the buyer sees**

10. **Media by type.** Hedonic → video; utilitarian → stills. Hand, shine, multiples of a single-function product. Depth: `product-visuals`.
11. **Price below the image.** Depth: `product-page`.
12. **Horizontal layout** among alternatives when quality should win. Depth: `store-audit`.
13. **Sold-out signal** at 10–30% of variants. Depth: `product-page`.

**Quality signals**

14. **Cost breakdown; say you're small** (low-tech); **extraordinary guarantee** if unknown; **real sustainability** stated if true. Depth: `store-audit`.
15. **Rank** by measured effect and distance from the rule; free same-day changes (digits, units, copy, price position) ahead of a shoot or a guarantee program.

## Output

| # | Finding | Change | Measured effect (study context) | Depth |
|---|---|---|---|---|
| 1 | $9.99, quality-led product | $14 list; just-below for promos only | 99-ending signals low price | `pricing` |
| … | | | | |

Then: what already complies. Each effect was measured on its own; don't add them up. Ship the top two or three, measure, then take the next.

## Common mistakes

- A 99-ending on the new price "to soften the jump" — that signals low price, the thing being fixed.
- Keeping fiddly add-on prices as "the real margin numbers" — complex components read as expensive.
- "More photos" — the measured levers are a hand, a shine, and multiples.
- Hiding that the company is small — for a low-tech product it raises perceived quality.
- Inventing a lift for the new price.

## Quick reference

| Lever | Depth |
|---|---|
| Rounded list; simple add-ons; flat rate | `pricing` |
| Units, branded ingredient, three claims, register | `product-copy` |
| Video / hand / shine / multiples | `product-visuals` |
| Price below image; sold-out share | `product-page` |
| Cost breakdown, size, guarantee, sustainability | `store-audit` |
| Validate the level | `price-research` |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: referrals
description: Use when designing or fixing a customer referral program for an online store — the headline offer, who gets what reward, how big the reward is, cash vs gift, and the pre-filled share message — e.g. "design our refer-a-friend program", "should the reward be cash", "how much should we give per referral", "write the referral share message", "referrals aren't converting".
---

# Referrals

Design a referral program that gets more customers to refer and keeps referred customers profitable. The lever is framing: people refer because they think the friend will like it, and a reward aimed at the referrer's self-interest works against that.

## When to use

- The referral offer and headline; reward size and type
- The pre-filled share message
- A program with many referrals but low-value customers

## When not to use

- Reviews, ratings, or complaints → the `reviews` skill
- Loyalty rewards for repeat purchase → the `post-purchase` skill

## Inputs

Collect, or state your assumption:

- Current headline, rewards for referrer and friend, reward type, share message
- First-order margin (to size the reward)
- Goal: more referrals, more conversions, or better-quality customers

## Procedure

1. **Make the friend's benefit the headline.** "Give a friend $20 off", not "Earn $20 per friend". Keep a referrer reward, stated second and smaller. Altruistic framing made people up to 60% more likely to refer at least once; total referrals rose 86%.
2. **Pre-fill the share message with one signal.** Either "I already bought this" (+11% friend conversion) or "you can earn a reward for referring" (+53% onward referrals). Pick by goal. Never both. Separately from the persuasion signal, if the sender earns a reward the message must say so — a brief "(I get a thank-you if you use it)" — because US/UK advertising rules treat an incentivized recommendation without disclosure as deceptive; the studies predate this expectation.
3. **Reward in kind, not cash.** A free gift, upgrade, or other non-monetary benefit over cash or a cash discount; store credit as the fallback. †
4. **Cap the reward size.** Higher rewards raise volume but pull in poor customers: when an online bank raised its reward from €20 to €50, profitability of referred customers fell 48%. Sizing against first-order margin is a conservative practitioner policy, not the study's finding — with strong repeat value a larger reward can still pay; if you use lifetime value, use *incremental* profitability of referred customers. †
5. **Pay out after the friend's return window closes.** (Hygiene, not a study finding.)

## Output

**Program spec**

| Element | Value | Why |
|---|---|---|
| Headline | "Give a friend $20 off their first order" | Altruistic framing: +86% referrals |
| Friend reward | $20 off / free gift | Friend's benefit leads |
| Referrer reward | Free [upgrade/gift], shown second | Non-monetary; secondary |
| Reward size | Below the margin on a first order | High rewards → −48% profitability |
| Payout timing | After return window | Hygiene |

**Share message** (one signal):

> "I bought [product] and think you'd like it — here's $20 off your first order: [link] (I get a small thank-you if you use it)"

or, if the goal is chain referrals:

> "Here's $20 off [product] — and once you've tried it you can give friends $20 off too (I get a small thank-you for sharing this): [link]"

## Common mistakes

- "Earn $50 cash per friend" as the headline — lead with the friend's benefit.
- A share message with both "I bought this" and "you can earn too" — one or the other.
- Raising the reward for volume — referred-customer profitability drops.
- Cash payouts — gift, upgrade, or at minimum store credit.

## Quick reference

| Element | Do |
|---|---|
| Headline | Friend's benefit first, referrer reward second |
| Share message | One of: "I bought this" / "you can earn too" |
| Reward type | Gift or upgrade > credit > cash |
| Reward size | Modest; cap against first-order margin |
| Payout | After return window |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: reviews
description: Use when planning how to get product reviews or feedback, when to ask, what to do about the average rating, or how to respond to reviews and public complaints — e.g. "when should we send the review request", "should we remove this 1-star review", "how do we reply to this complaint", "we need more reviews", "our rating is 4.9".
---

# Reviews

Get more reviews and answer complaints in a way that wins over the people watching. Counterintuitive parts: an imperfect rating sells more than a perfect one, asking too soon reduces reviews, and the best first question is not "rate us".

## When to use

- The review-request email or sequence
- A negative review or a too-perfect average
- Replying to a public complaint
- A product launch that needs a healthy review profile

## When not to use

- Referral mechanics → `referrals`
- Loyalty, returns, packaging → `post-purchase`

## Inputs

Collect, or state your assumption:

- Product and how long it takes to judge (instant: soda; slow: skincare, a course)
- Current ask timing and wording; average rating; reviews the user wants to remove
- Whether reviews get replies; complaint volume, channel, and tone (polite vs rude)

## Procedure

1. **Ask at 9–13 days** after purchase or delivery. A request at 24 hours *reduced* reviews, worst among younger customers. Earlier only for products judged instantly; later for products that take weeks to evaluate.
2. **Lead with a positive open-ended question.** Before any rating: "What did you enjoy most about [product]?" (free text). Responders spent 131% more afterward; even receiving the request helps. Usual questions follow. Automate it.
3. **Never gate the public review by score.** Routing low scores to a private form and high scores to the review site is unsupported by the evidence and treated as deceptive by regulators (US FTC, UK CMA). Everyone gets the same path.
4. **Get reviews in early at launch.** Ask *all* early customers as soon as step 1 allows — the same path for everyone, no picking the happy ones (that is selective solicitation, see step 3). The point is speed: a negative first review left the average 0.29 stars lower a year later on Amazon and Best Buy, so make sure the first review isn't left to chance by waiting.
5. **Leave the imperfection.** 4.0–4.5 stars sells more than 4.5–5.0; perfect looks suspicious. A negative review irrelevant to most buyers helps: an espresso machine with four 5-stars rated 16% better with an irrelevant 1-star added. † Reply; don't remove.
6. **Reply to every review.** When companies started responding, ratings rose 0.12 stars and review volume 12%.
7. **Public complaints: once, fast, warm, then private.** You're replying to everyone reading, not the complainer. Reply fast and only once: empathy, an explanation if you have one, a private channel to continue. Replied-to posts get boosted on X and LinkedIn, so that one reply is what people see. High volume → a separate support account.
8. **Rude complaints: humor about the situation, never the person.** Onlookers engaged more and were 18.3% more likely to buy than with no response. If it isn't funny, use step 7.
9. **Describe the product as an experience** in copy (people sharing coffee, not bean origin) — such products collect more reviews. †

## Output

1. Solicitation plan: trigger, day offset, first question, follow-ups.
2. Rating policy: what to leave, reply wording for an irrelevant negative.
3. Templates: public complaint reply with private handoff; rude-complaint reply with situational humor; positive-review reply.

## Common mistakes

- Asking 1 day after delivery — reduces reviews; 9–13 days.
- Opening with "rate 1–5" — open with what they enjoyed most.
- Score-gating — unsupported and legally risky.
- Removing an irrelevant 1-star — it makes the rest credible.
- Protecting a 4.9 — 4.0–4.5 sells more.
- Replying to a complaint repeatedly in public — once, then private.

## Quick reference

| Situation | Do |
|---|---|
| Request timing | 9–13 days (earlier if judged instantly, later if slow) |
| First question | "What did you enjoy most?" |
| Launch | Ask all early customers promptly; same path for everyone |
| Average rating | 4.0–4.5; leave irrelevant negatives |
| All reviews | Reply to each |
| Public complaint | One fast empathetic reply → private |
| Rude complaint | Humor about the situation |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: store-audit
description: Use when reviewing an online store as a whole — homepage, category pages, channel mix, or trust — e.g. "review my store", "audit my homepage", "Instagram or Google Ads?", "should we say we're a small team", "how should we label recommendations", "carousel or grid", "how do we look more trustworthy", "should we sell on Amazon too". For one product page, use product-page; for cart or checkout, use checkout-ux.
---

# Store audit

Make the site-wide calls that depend on what you sell and what kind of company you are. Most follow from one classification, so classify first.

## When to use

- Where marketing budget goes (social, search, comparison sites, marketplaces)
- Homepage/category layout, carousels, recommendation labels, visual style
- Which trust signals to show
- Whether to list on a marketplace

## When not to use

- A single product page → `product-page`
- A cart, checkout, or payment form → `checkout-ux`
- Shipping, returns, loyalty, retargeting → `post-purchase`
- Prices or discounts → `pricing` / `discounts`

## Inputs

Collect, or state your assumption:

- The product and who buys it (for the classification)
- Company size and whether the product is high-tech
- Current channel mix, homepage layout, recommendation label
- Gross margin; any verifiable sustainability practice; marketplace presence

> **Classify the product first.** *Hedonic* = bought for enjoyment, pleasure, or the experience (chocolate, fashion, decor, travel, spa, perfume). *Utilitarian* = bought to do a job, judged on function and value (detergent, appliances, insect repellent, tools, accounting software). If it's mixed, decide by what the buyer is optimizing for at the moment of purchase.

## Procedure

1. **Channels by product type.** Hedonic → social media and your own product pages; the journey is impulsive and starts on social. Utilitarian → search ads, comparison sites, third-party reviews, deal sites; the buyer compares for value.
2. **Recommendation labels.** Utilitarian → algorithm-based label ("Recommended based on your browsing"). Hedonic → human label ("Our expert's pick"). Keep the module; change the label.
3. **Layout direction.** Horizontal (carousel) makes people weigh quality and the primary feature — 26% more chose the pricier, higher-quality laptop. Vertical (scroll) makes them price-sensitive. Horizontal where quality should win, vertical where you compete on price.
4. **Visual style.** Structured (straight lines, symmetry, geometry) reads reliable; unstructured (rounded, incomplete, unpredictable) reads fun. Match the brand: matched perfume ads got 39% more clicks. †
5. **Cost transparency.** If margin ≤ ~55%, show a per-product cost breakdown (materials, labor, shipping, packaging, development); group beyond ~8 categories. Sales were 22% higher for a $115 wallet with one.
6. **Company size.** Small, low-tech → say so prominently; buyers infer motivated staff and better quality. Small, high-tech → show R&D effort. Large, low-tech → show employee passion.
7. **Warranty.** Offer one far longer than the category norm (10-year hoodie): 34.8% more bought. Best for buyers who know the category but not your brand. For consumables, the equivalent is an unusually strong effectiveness or satisfaction guarantee (untested, same mechanism).
8. **Sustainability.** If real, state it on product and packaging: 6.4% higher revenue across 10 million grocery sales, and people enjoy the product more when they know. Small businesses gain most. †
9. **Marketplaces.** If unknown, list on a major marketplace: 1 in 71 marketplace sales creates a direct customer; acquisition cost averages 24% of revenue vs 30–100% direct. Give a reason to switch, within marketplace rules.

## Output

Table: priority, area, recommendation, rule + number. Then what to leave alone, and why.

## Common mistakes

- Removing an "algorithm" label as fake — for utilitarian products it outperforms; for hedonic, relabel as human.
- Calling carousels weak in general — direction sets whether quality or price wins.
- Choosing search over social on intent alone — the rule is product type; hedonic starts on social.
- Hiding that you're small — for low-tech products it raises perceived quality.
- Hiding costs to protect margin — the effect holds to 55%.

## Quick reference

| Situation | Do |
|---|---|
| Hedonic | Social + product pages; human-labeled recs |
| Utilitarian | Search, comparison, review, deal sites; algorithm-labeled recs |
| Quality should win / price should win | Horizontal / vertical layout |
| Margin ≤ 55% | Cost breakdown, ≤ 8 categories |
| Small, low-tech / small, high-tech | Say you're small / show R&D |
| Unknown brand | Extraordinary warranty; marketplace listing |

† No primary citation on file for this rule — a practitioner tip, not a measured finding.

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.

View saved version →

---
name: subscription-pricing
description: Use when structuring or reviewing prices for a subscription, membership, recurring plan, or member dues — plan tiers, upgrade paths, pay-per-use vs flat rate, extra-seat or overage pricing, membership perks — e.g. "how should I price Basic vs Pro", "flat monthly or per video", "what should extra seats cost", "how do we get more upgrades", "is a paid membership worth it".
---

# Subscription pricing

Structure recurring prices so people choose a plan, pay the flat rate, and step up a tier. This skill is the delta versus `pricing`, `discounts`, and `post-purchase`. One-off digital products → `digital-product-pricing`.

## When to use

- Designing or reviewing plan tiers and their prices
- Choosing flat-rate vs pay-per-use, or pricing seats, storage, minutes
- Adding a paid membership perk, or rewards for members

## When not to use

- One-off digital products → `digital-product-pricing`
- Shared digit and component rules in isolation → `pricing`
- Wording a plan discount → `discounts`; measuring willingness to pay per month → `price-research`

## Inputs

Collect, or state your assumption: tiers, prices, billing period; any per-unit component; whether members optimize for quality or lowest price; which tier you want people to move to; order/usage frequency.

## Procedure

Only the rules that change for recurring plans. Shared catalogs live in the Depth skills.

1. **Lead with a flat rate that covers what most members want**, priced *above* the pay-per-use equivalent. Keep pay-per-use only as a low-commitment entry. The source study is purchasing professionals buying usage-based services; applying it to consumer memberships is a transfer. Depth: `pricing`.
2. **Put the upgrade path on the same left digit.** No just-below on the lower tier if you want Basic → Pro ($40 → $48, not $39.99 → $48). Depth: `pricing`.
3. **Simplify every component.** Rounded per-seat / per-GB prices, the same ending, no admin fee. Depth: `pricing`.
4. **Paid perk needs a minimum.** A membership that unlocks free shipping or member rates raises spend; without a minimum, members split into many small orders. Depth: `post-purchase`.
5. **Rewards: automatic, single tier, threshold.** Optionally variable; combining variable rewards with auto-enrollment is untested. Depth: `post-purchase`.

Shared, unchanged: quality-led → rounded, price-led or post-trial → just-below (subject to step 2); "for $0" on free trials; abandoned-signup reminder at 24–72 h is an assumption transferred from carts. Depth: `pricing`, `promotions`, `post-purchase`.

## Worked example

Membership site, plan: Basic $9.99, Pro $29.99, pay-per-video $2.49, extra seats $4.61 + $6 admin fee, opt-in 3-tier rewards, goal: more Basic→Pro upgrades.

- **$10 → $30** removes the just-below start that discourages the step up. Same-left-digit ($40.50 → $48.50) is the stronger version and needs closer tiers.
- Pay-per-video stays as entry, priced so ~10 videos equals Pro.
- Seats: **$5, no admin fee**.
- Rewards: **auto-enroll, one tier**.

## Output

| Item | Value |
|---|---|
| Tiers | Basic $10/month · Pro $30/month |
| Upgrade path | No .99 on Basic |
| Pay-per-use | Entry only; heavy users cheaper on Pro |
| Components | $5 per extra seat, no admin fee |
| Perk / rewards | Minimum order on the perk; auto-enrolled single tier |
| Why | one line per choice |

## Common mistakes

- Adding a mid-tier to "shrink the jump" — the jump is the .99 boundary; fix the digits first.
- Pricing the flat rate at the pay-per-use equivalent "to be fair" — people pay more for the cap.
- A perk with no minimum — members fragment usage to exploit it.
- Opt-in tiered rewards — automatic, single tier, threshold reward.

## Quick reference

| Situation | Do | Depth |
|---|---|---|
| Plan structure | Flat rate above pay-per-use; pay-per-use as entry | `pricing` |
| Want upgrades | Same left digit; no .99 on the lower tier | `pricing` |
| Seats / storage | Rounded, same endings, no fee | `pricing` |
| Membership perk | Minimum order/usage | `post-purchase` |
| Member rewards | Auto-enroll, single tier | `post-purchase` |

Evidence and effect sizes: see [reference.md](reference.md).

Referenced files: 1 — file references are metadata, not package contents.