skills/insight-mining/references/insight-mining.md

# Insight Mining: The Foundation of a Strong Idea

An insight is the foundation of every strong creative idea. Without an insight, an idea becomes decoration. With an insight, it becomes a statement that's impossible to ignore. This document collects seven practical techniques for finding insights before ideation begins.

---

## 1. Mark Pollard's Four Points

Pollard's framework builds a bridge from problem to strategy through four sequential steps:

**Problem — Insight — Advantage — Strategy**

- **Problem**: the real human or business problem, not a symptom. "Sales are falling" is a symptom. "People don't understand why they need this product" is a problem.
- **Insight**: a hidden truth that flips the view of the problem. Not a fact, not an observation, but specifically a truth that everyone feels but no one has articulated aloud.
- **Advantage**: what exactly the brand uniquely offers in the context of this insight. Not a list of features, but one specific strength.
- **Strategy**: one directive sentence that tells the creative team what to do. Not "how," but "what."

**Example**: brief for life insurance.
- Problem: people avoid thinking about death, so they postpone buying insurance.
- Insight: it's not about the fear of death — it's about what you'll leave behind.
- Advantage: the brand makes legacy planning simple and human.
- Strategy: reframe insurance from a topic of fear into a topic of love.

Each step validates the previous one. If the insight doesn't stem from the problem — go back to the problem. If the strategy doesn't rest on the advantage — the advantage is weak.

---

## 2. JTBD (Jobs To Be Done) for Advertising

People don't buy products — they "hire" them for a job. The task of advertising is to understand exactly which job.

Three layers of the job:
- **Functional**: what the product does physically. "Satisfies hunger."
- **Social**: how the product affects how others perceive the person. "Shows that I care about health."
- **Emotional**: what the person feels in the process. "Gives a sense of control over life."

The real competitor is not another brand in the category, but everything that performs the same job.

**Example**: Clayton Christensen's milkshake story. People "hired" the milkshake in the morning for the job of "make the boring commute to work bearable." The milkshake's competitors were podcasts, bananas, bagels. Not other beverages.

For advertising, the key question is: "what job does the audience hire this communication for?" A person doesn't hire a sneaker ad to learn about sole specifications. They hire it to feel like someone who gets up at five in the morning and runs.

Practice: take the product from the brief and write five "jobs" it's hired for. For each job, write three non-obvious competitors. The most interesting job with the most unexpected competitor — that's the direction for an insight.

---

## 3. Tension Spotting

Tension is a gap between how things should be and how they actually are. Three types:

**Cultural tension**: what society declares vs what people actually do.
- Example: "everyone talks about conscious consumption, but orders fast fashion at two in the morning."

**Category tension**: what the category promises vs what it actually delivers.
- Example: "banks promise partnership and care, but treat you like a number in a queue."

**Human tension**: what a person wants vs what stands in the way.
- Example: "wants to eat healthy, but hates diet food" — Halo Top was born from this tension.

How to find tensions:
1. List 5 conventions in the category. What all brands say, show, promise.
2. For each convention, ask: "but what actually happens?"
3. The gap between convention and reality = tension.
4. The best tensions are simultaneously uncomfortable and recognizable. The person hears them and winces, but nods.

If the tension doesn't cause discomfort — it's too mild. If it doesn't trigger recognition — it's fabricated.

---

## 4. HMW (How Might We)

HMW is a question formulation that opens up space for ideas. The quality of the question determines the quality of the answers.

Rules for a good HMW:
- Too broad: "How might we change the world?" — no direction, the team scatters.
- Too narrow: "How might we change the button color?" — no space for thinking.
- Right level: "How might we make [X] feel like [unexpected Y]?"

For each brief, generate three HMWs at different levels:
- **Broad**: culture or category level. "How might we change people's attitude toward insurance?"
- **Medium**: brand and audience level. Usually the most productive. "How might we make choosing insurance an act of love, not fear?"
- **Narrow**: product or specific format level. "How might we turn an insurance calculator into a conversation about dreams?"

Examples of good and bad HMWs:
- Bad: "How might we increase sales?" — this is a business objective, not a creative question.
- Good: "How might we make choosing insurance feel like writing a letter to a loved one?"
- Bad: "How might we make the world better?" — too abstract, nothing to grab onto.
- Good: "How might we turn the boring commute to work into the best part of the day?"

The medium level usually works best because it provides direction while leaving room for freedom.

---

## 5. Abstraction Laddering

Moving up the ladder ("why?") expands the solution space. Moving down ("how exactly?") narrows to specifics.

**Example**:
- "Need a chair" — starting point.
- Up (why?): "need comfort."
- Further up (why?): "need to feel at home."
- Down (how exactly?): "a cozy corner."
- Further down (how exactly?): "a reading nook with warm light."

The result is radically different from the starting "need a chair," though it solves the same problem.

**Rule**: the optimal rung is one level above the obvious. Too high — you get generic ("people need happiness"). Too low — tactics without meaning ("15% discount").

**For advertising**: take the formulation from the brief, go up two steps, come down one. That's usually the spot where a strong idea lives. The brief says "tell about the new feature" — go up: "why this feature?" — "to save time" — "why save time?" — "to spend it on what matters" — come down: "show what the person does with the freed-up time." That's already territory for an idea.

---

## 6. Assumption Mapping

Every brief comes with a set of assumptions that everyone takes as given. The task is to surface them and verify.

Process:
1. List all assumptions about the task, audience, product, competitors. Minimum 10.
2. Place each on a 2x2 matrix: importance (high/low) and confidence (high/low).
3. The quadrant "high importance + low confidence" is the priority zone. These assumptions need to be tested or challenged first.

**Example**: assumption "our audience is men 25-35." But what if the buyer is their wife? Or their mother? The entire communication changes. Tone, channels, message, visual — everything.

Another example: assumption "people choose between us and competitor X." But what if they choose between us and "doing nothing"? Then the main enemy is not a competitor, but inertia. And advertising must fight not against another brand, but against inaction.

The strongest insights often hide in assumptions that no one questions.

---

## 7. Insight Quality Test

Final checklist. A strong insight passes all six checks:

1. **True?** This is a recognized truth, not a hypothesis. People recognize themselves in it.
2. **Fresh?** Not "people like convenience" or "everyone wants to save." Those are observations, not insights.
3. **Contains tension?** There's an internal conflict, contradiction, friction within it.
4. **Triggers the reaction "yes, exactly, but I've never put it that way"?** This is the main marker. If the person shrugs — the insight is weak.
5. **Can the creative team build from it?** An insight that doesn't lead to ideas is a dead end, not an insight.
6. **Specific to this audience and category?** Not a universal banality, but a precise hit on specific people in a specific context.

If the insight doesn't pass two or more checks — dig deeper. Don't cling to the first one found, look for the next.

---

## Insight Format

The final insight is written in the format:

**"[Audience] wants [X], but [Y stands in the way], because [Z]"**

Examples:
- "Young parents want to secure their children's future, but avoid the topic of insurance, because it's associated with death, not care."
- "People aged 30-40 want to eat healthy, but relapse into junk food, because healthy food is positioned as punishment, not pleasure."
- "Small businesses want to grow, but don't take out loans, because banking language creates a feeling of a trap, not an opportunity."

This format forces the inclusion of all three elements: desire, barrier, and reason for the barrier. Without the reason, the insight remains a superficial observation.

---

*Author: Serge Shima ([t.me/aimastersme](https://t.me/aimastersme) · [sergeshima.com](https://sergeshima.com) · [aimasters.me](https://aimasters.me)) · License: CC BY 4.0 — attribution required · Source: [smixs/creative-director-skill](https://github.com/smixs/creative-director-skill)*