Market Research and Competitive Analysis

Don't research the market. Research the hole in it. A market size tells you how big the room is; a hole tells you where to stand.

The concept

Market research as most people do it produces a number, "the wellness app market is $5.8B growing at 16% CAGR", and that number changes no decision you will ever make. It doesn't tell you what to build, what to charge, what to say on your landing page, or why anyone would switch.

Useful market research produces a sentence:

"Every [category] product is [shape A] or [shape B]. Nobody offers [your shape]. Users say '[verbatim complaint]'. We are the only [your shape] for [specific customer]."

That sentence is your product strategy, your positioning, your store description, and your investor pitch. Getting to it is the whole job.

Why anyone switches

Your default outcome is that nobody switches. Users already have a solution; switching costs energy; you're unproven. The only reliable reason someone moves is that their current tool fails them in a way they can articulate.

So your job is to find that articulation, ideally in their own words, in public, and build the thing they're describing.

Product shapes, not features

Every category splits into a small number of shapes. Competing on features within a shape is a losing game; occupying an empty shape is defensible.

A generic map that works in most categories:

  REFERENCE          TRACKER            LIBRARY
  look it up,        log it, streaks,   browse/collect,
  feed, search       dashboards         save for later

  COURSE             COMMUNITY          SERVICE
  sequenced,         people, forums,    done-for-you,
  progression        accountability     human in the loop

  TOOL               PLATFORM           COACH
  do one job well    others build on it  tells you what to do next

Most categories are saturated in two or three of these and thin in the rest. The thin box is your candidate.

Concrete examples of the same insight in other categories:

Caution: an empty shape is not automatically an opportunity. Sometimes it's empty because it doesn't work. The check: find any category where your shape wins. If "course" works for language (Duolingo) and weight (Noom), the shape is viable, it just hasn't been applied here.

Evidence quality

Every claim you collect has a different evidential strength, and once they're all bullet points in your notes they look identical. Tag at the moment you write:

• [VERIFIED, source], you can point to where it came from
• [ESTIMATE], modelled, third-party, or directional

Third-party revenue estimates (Sensor Tower, Appfigures, Adapty) are modelled and can be off by 2×. Use them for order of magnitude. "$250k MRR" means "somewhere between $100k and $500k", which is still enough to prove a shape works.

📐 Best practice

Map shapes before features. List the 5-7 shapes in your category, place every competitor, find the thin box.

Read 1★ and 3★ reviews, never 5★. Five-star reviews tell you what an app does well, useless, you're not trying to match it. Three-star reviews are the goldmine: written by people who want to like the product and are explaining exactly what's missing. That's your feature list, written by your future customers, for free.

Hunt for the complaint that repeats. One person's opinion is noise. The same complaint in three places, App Store reviews, a Reddit thread, a blog post, is a market signal. Quote it verbatim; the exact words become your copy.

Take the price anchor from the market. You are choosing a position on a scale your customers have already been trained on. Collect 3-5 competitor prices before you invent one.

Import proven mechanics across categories. Streaks from Duolingo, quiz-funnel onboarding from Noom, annual-default pricing from everywhere. Proven elsewhere, unproven here, that's an arbitrage.

⭐ Become your competitor's customer, with the screen recorder on. Pay for the subscription. Complete the whole onboarding. Record it. Then transcribe every screen into a document. Reading about a competitor teaches you nothing; $15 to watch a working funnel is the best-value research spend available.

Check whether a missing product is an opportunity or a graveyard. If the biggest brand in your space has no app, ask why not. Sometimes they don't want to. Sometimes they tried and the economics didn't work. Search for a dead app in the store's history before you celebrate.

💀 Common mistakes

Researching the market instead of the hole. You end up with a TAM slide and no product decisions.

Mistaking a feature gap for a shape gap. "They don't have dark mode" is not an opportunity, it ships in a sprint. If a competitor can close your gap in two weeks, it isn't a gap.

Trusting a single revenue estimate. Triangulate across three sources, present as a range, assume the low end.

Only studying the top 10. The top 10 are entrenched. The interesting signal is #20-50, small apps growing fast show you where the category is moving.

Blending verified and estimated claims. By the time you're making decisions you won't remember which was which, and you'll build on a number that was never real.

Reading about competitors instead of using them. The most common and most expensive research mistake.

Letting research go stale. Pricing and top-10 rankings change monthly. Diarise a 20-minute refresh.

Confusing "no competitors" with "great opportunity." It usually means no market. Competitors are proof that people pay.

The professional workflow

 1. LIST THE SHAPES in your category (5-7). Place every competitor.
    → find the thin box

 2. FOR EACH SIGNIFICANT COMPETITOR, record:
      what it fundamentally IS · downloads · revenue estimate ·
      pricing · ⭐ THE HOLE (a structural absence, not a missing feature)

 3. MINE REVIEWS for the repeating complaint. Verbatim quote.
      App Store "Most Critical" · G2/Capterra 2★ · Reddit
      "why I stopped using X" · "X alternative"

 4. COLLECT THE PRICE ANCHOR (3-5 competitors, monthly + annual)

 5. ⭐ SUBSCRIBE TO THE CLOSEST COMPETITOR. Record the full
    onboarding and paywall. Transcribe every screen.

 6. LIST 3-5 PROVEN MECHANICS from other categories you can import,
    each with its evidence

 7. WRITE THE ONE SENTENCE. If you can't say it in one breath,
    keep researching.

 8. TAG EVERY CLAIM [VERIFIED — source] or [ESTIMATE]

Tools, websites & costs

Total: $20-40. Budget the competitor subscription; it's the highest-value line.

The shape-map prompt

I'm building in [CATEGORY]. Map the landscape by PRODUCT SHAPE, not features.

1. Every meaningful competitor: downloads, revenue estimate, pricing, and a
   one-line description of what it fundamentally IS.
2. Classify each into a shape: reference / tracker / library / course /
   community / service / tool / platform / coach.
3. Which shapes are crowded? Which are thin or empty?
4. For each competitor, state THE HOLE — what it structurally does not do.
5. Find the "universal complaint": the criticism that repeats across reviews,
   Reddit and blogs. Quote it VERBATIM with a source.
6. The price anchor: what has this category trained users to expect?
7. For any empty shape: name a DIFFERENT category where that shape wins,
   as evidence the shape is viable at all.

Tag every figure [VERIFIED — source] or [ESTIMATE]. Never blend them.

Alternatives & trade-offs

TAM/SAM/SOM modelling, necessary for raising money, useless for building. It produces a number that changes no decision. Do it when an investor asks.

Surveys, people lie to be nice, and they're bad at predicting their own behaviour. Useful for segmenting an audience you already have; poor for validating demand (Chapter 05).

Analyst reports (Gartner, Forrester), expensive, lagging, aimed at enterprise buyers. Occasionally useful in B2B for the vocabulary buyers use.

Just building and seeing, legitimate when the build is genuinely a weekend and the category is unknowable. Indefensible when the build is three months.

Depth vs breadth: studying one competitor exhaustively usually beats skimming twenty. The exhaustive study gives you their funnel, their pricing psychology, and their gaps. Skimming gives you a table.

Checklist

• I have a shape map with every competitor placed
• I can name the hole in each significant competitor, structural, not a missing feature
• I have the universal complaint as a verbatim quote with a source
• I have a price anchor from 3+ competitors
• I have 3-5 proven mechanics from other categories, with evidence
• ⭐ I have subscribed to and screen-recorded the closest competitor's full funnel
• For any empty shape, I can name a category where that shape does win
• Every claim is tagged [VERIFIED] or [ESTIMATE]
• I can say the one sentence in a single breath
• I have diarised a 20-minute refresh for next month

📓 Case Study: the table that decided a product

Project: SOLIS, entering the Stoicism / self-improvement app category.

The research produced a four-row table. Note that the right-hand column is not "features they lack", it's what each competitor structurally is not:

Read down that column: every entry is a different absence. That's what a healthy opportunity looks like, the category has split into shapes and one is empty.

The sentence that made it a decision came from mining reviews and blog posts, and it's worth more than the entire market-size figure:

"Most Stoic apps show you a quote, ask you to journal, and charge $50/year… they skip the hard part… more flavor than substance."

Four things fall out of one sentence: the price anchor ($50/year), what users currently get (a quote and a prompt), why they're unhappy ("they skip the hard part"), and what they want (substance, structure, actual work).

The resulting strategy: the only structured, gamified, story-driven 90-day course in the category, with visible proof of change. Every subsequent product decision was checked against that sentence.

Imported mechanics. Rather than inventing engagement design, the research collected proven mechanics from other categories: streaks with a freeze (Duolingo published +14% D7 with streak wagers and −40% churn among their best users), quiz→plan→paywall onboarding (Noom/Duolingo, Duolingo's goal-first onboarding reportedly moved free→paid from ~4% to ~9%), annual-default pricing with a savings badge. None of these were novel; none had been applied in this category.

⚠️ Deviation 1: evidence quality wasn't tagged, and it cost real planning time.

One collected finding, "longer trials convert ~2× better (17-32 days ≈46% vs 3-7 days ≈27%)", sat in the plan as fact and shaped thinking about trial length. When the research was later re-run with adversarial verification, that claim was refuted and couldn't be traced to a reliable source.

Best practice is to tag at write-time. The lesson isn't subtle: findings of wildly different evidential quality look identical once they're bullet points, and by the time you're deciding you won't remember which was which.

⚠️ Deviation 2: the single most valuable research artifact arrived on day 22 of 24.

Best practice says record a competitor's full funnel early. SOLIS did it, brilliantly, but almost at the end. A competitor doing ~$20k MRR was screen-recorded through its entire onboarding (5m36s, ~40 screens) and transcribed screen-by-screen into a teardown document.

That document became the most actionable artifact in the project. Twelve reusable mechanics were extracted from it, and the onboarding was rebuilt around them (Chapter 10). But because it landed on day 22, the funnel had already been built once and then substantially rebuilt.

A $15 subscription and a six-minute recording, deferred by three weeks, cost a rebuild. Do it in week one.

Lessons

1. Market size answers no question you actually have. Shape maps answer all of them.
2. The universal complaint is your product spec, written by your customers, for free. Find it verbatim.
3. Tag evidence quality at write-time, or you'll plan around a claim that was never true.
4. Screen-record your competitor's funnel in week one. Reading about it teaches you nothing.
5. Price is already anchored. You're choosing a position, not inventing a number.
6. Import mechanics across categories. Proven elsewhere, unproven here, is an arbitrage.
7. An empty shape needs a viability check, find a category where that shape wins.
8. Research decays. Pricing and rankings move monthly.

Next: Chapter 04: Talking to Users →