Competitor Teardown
A competitor analysis that ends in a table is wasted work. It must end in a decision: the wedge
we take, the price we set, or the conclusion that we don't enter.
Uses research/public-web-research-at-scale to gather and thinking/evidence-grading to grade.
Feeds business/offer-design, business/unit-economics, and design/web-page-build.
Reframe first: competitors are not who you think
Before analyzing named vendors, identify the actual alternatives, in this order of likelihood:
- Doing nothing — the most common and most-underestimated competitor. The status quo has zero
switching cost and infinite political safety.
- A spreadsheet / whiteboard / WhatsApp group — how the job is really done in SMB markets.
- A person — a part-time admin, a bookkeeper, an agency.
- An adjacent tool being misused — the CRM used as a scheduler.
- Named direct competitors — usually the smallest share of the decision.
If the real competitor is "nothing," the offer's job is to overcome inertia, not to beat features.
That completely changes the landing page, the pricing, and the sales motion. Getting this wrong is
the most expensive error in this skill.
What to collect per competitor (all grade-A: their own published material)
| Field |
Where |
Why it matters |
| Positioning one-liner |
Homepage <h1> |
Their claimed wedge, in their words |
| Target customer |
Homepage, case studies, "who it's for" |
Reveals the segment they've conceded |
| Pricing & packaging |
Pricing page (screenshot/quote verbatim + date) |
Sets the market's price anchor |
| Feature surface |
Product/features pages, changelog, docs |
Depth vs breadth |
| Proof |
Logos, case studies, review counts |
How much trust they've banked |
| Onboarding friction |
Signup flow — self-serve vs "book a demo" |
Demo-gate = high ACV, slow, weak self-serve |
| GTM channel |
Blog cadence, ads, SEO footprint, directories, job posts |
Where they spend, where they don't |
| Company shape |
Registry filings, headcount signals, funding, job posts |
Can they respond to us? How fast? |
| Weakness signals |
Reviews (1–3 star specifically), support forums, churn complaints, stale changelog |
Where the wedge is |
| Last shipped |
Changelog/blog/release notes date |
Stale >12mo = a sleeping incumbent = opportunity |
"No pricing published" is itself a finding: it means enterprise/negotiated, which means slow
sales, which means a self-serve wedge is open underneath them.
Highest-signal sources, in order
- Their pricing page — grade A, quote it verbatim with the fetch date.
- Their 1–3 star public reviews — the single richest source. Sort by lowest, read 20, and
tally the complaints. Recurring complaints are the product spec for your wedge. Individual
reviews are grade C; a tally of 20 is a real pattern.
- Their job postings — where they're investing next, and their stack. A company hiring 5
enterprise AEs is moving upmarket and abandoning the small end.
- Their changelog — velocity and direction, or its absence.
- Their docs/API — real capability, unfiltered by marketing.
- Forums/subreddits where their users complain — unvarnished, and where switching intent shows up.
Step 2 — the grid
Rows = the 3–6 buying criteria the customer actually uses (from real discovery, not invented).
Columns = doing-nothing, the spreadsheet, and 2–4 named competitors. Cells = evidence + grade, not
opinion.
Two rules that keep this honest:
- Score them generously. A teardown where we win every row is a fantasy that will be corrected
by the market at full price. Find the rows where they genuinely beat us and write them down.
- The criteria must come from customers. A grid built on criteria we chose because we win them
is marketing, not research.
Step 3 — find the wedge
A wedge is a segment + criterion where the incumbent is structurally unable to compete — not
merely currently worse. Structural means their business model, not their backlog.
| Wedge type |
Signal |
Why it's structural |
| Segment abandonment |
They moved upmarket (enterprise hires, "contact sales", min seats) |
Their cost structure can't serve small accounts profitably |
| Price umbrella |
High price, high margin, no low tier |
A cheap tier cannibalizes their core revenue — they won't |
| Onboarding gate |
Demo-only, long implementation |
Self-serve requires rebuilding the org, not a feature |
| Neglect |
Stale changelog, unanswered support |
Attention is elsewhere; won't return |
| Complaint cluster |
Same 1-star complaint for 2+ years |
If it were cheap to fix, it'd be fixed. It's architectural |
| Geography / vertical gap |
No local presence, no vertical language |
Requires local ops or domain depth they lack |
| Integration gap |
Missing the tool your segment actually uses |
Roadmap reflects their ICP, not yours |
Feature parity is not a wedge. "Same but nicer" loses to distribution every time. If the only
answer is "we'd be better," there is no wedge — say so. "Don't enter" is a legitimate and valuable
output of this skill, and cheaper than discovering it after building.
Step 4 — the decision (mandatory output)
TEARDOWN: <market> date: <>
REAL ALTERNATIVES (ranked by what customers actually do)
1. ... 2. ... 3. ...
PRICE ANCHOR
<verbatim from pricing pages, with urls + fetch dates. The number a prospect
already has in their head.>
GRID
| Criterion (customer-sourced) | Nothing | Spreadsheet | Comp A | Comp B | Us |
THE WEDGE
Segment: <narrow>
Criterion: <the one thing>
Why structural: <their model prevents responding>
Evidence: <grades + urls>
PRICING IMPLICATION -> business/unit-economics
POSITIONING LINE -> design/web-page-build
KILL CRITERIA -> thinking/decision-quality
"We abandon this if <observable> by <date>."
WHAT I COULD NOT VERIFY
<gaps, failed fetches, JS-only pages, unpublished pricing>
Failure modes
| Symptom |
Cause |
Fix |
| Grid where we win everything |
Criteria chosen to flatter us |
Source criteria from customers; score rivals generously |
| Only named SaaS competitors listed |
Missed status quo / spreadsheet |
Redo step 1 — "nothing" is usually the leader |
| Feature-list comparison |
Treated it as a product exercise |
Ask what the buyer decides on; mostly not features |
| Conclusion "we'd do it better" |
No wedge found, unwilling to say so |
Say "no wedge." That's the finding |
| Pricing from a review site |
Not from source |
Fetch their pricing page; quote verbatim + date |
| Analysis rots in 3 months |
No re-check trigger |
Set a re-run date and a trigger (competitor raises, launches, or reprices) |
Boundaries
- Public evidence only: their published pages, public registries, public reviews, official APIs.
- No competitive intelligence by deception — no fake trial signups under invented identities, no
posing as a customer to extract pricing, no contacting their staff under pretext, no NDA'd or
leaked material. It is a small market; this gets found out, and it converts a competitor into an
enemy with a story.
- A trial signup under our real identity to evaluate a product is fine and normal.
- No login-walled scraping of review platforms; read public pages only.
- Quote pricing verbatim with a date — paraphrased pricing is how a business builds a model on a
number that was never real.
1---2name: competitor-teardown3description: Analyze a competitor from public evidence into a decision — positioning, pricing, ICP, go-to-market, weaknesses, and the wedge available to us. Use before entering a market, before pricing, and before writing a landing page. Triggers on "competitors", "who else does this", "how do they price", "differentiate", "market landscape", "are we too late", "competitive analysis".4---56# Competitor Teardown78A competitor analysis that ends in a table is wasted work. It must end in **a decision**: the wedge9we take, the price we set, or the conclusion that we don't enter.1011Uses `research/public-web-research-at-scale` to gather and `thinking/evidence-grading` to grade.12Feeds `business/offer-design`, `business/unit-economics`, and `design/web-page-build`.1314## Reframe first: competitors are not who you think1516Before analyzing named vendors, identify the **actual** alternatives, in this order of likelihood:17181. **Doing nothing** — the most common and most-underestimated competitor. The status quo has zero19 switching cost and infinite political safety.202. **A spreadsheet / whiteboard / WhatsApp group** — how the job is *really* done in SMB markets.213. **A person** — a part-time admin, a bookkeeper, an agency.224. **An adjacent tool being misused** — the CRM used as a scheduler.235. **Named direct competitors** — usually the smallest share of the decision.2425**If the real competitor is "nothing," the offer's job is to overcome inertia, not to beat features.**26That completely changes the landing page, the pricing, and the sales motion. Getting this wrong is27the most expensive error in this skill.2829## What to collect per competitor (all grade-A: their own published material)3031| Field | Where | Why it matters |32|---|---|---|33| Positioning one-liner | Homepage `<h1>` | Their claimed wedge, in their words |34| Target customer | Homepage, case studies, "who it's for" | Reveals the segment they've *conceded* |35| **Pricing & packaging** | Pricing page (screenshot/quote verbatim + date) | Sets the market's price anchor |36| Feature surface | Product/features pages, changelog, docs | Depth vs breadth |37| Proof | Logos, case studies, review counts | How much trust they've banked |38| Onboarding friction | Signup flow — self-serve vs "book a demo" | Demo-gate = high ACV, slow, weak self-serve |39| GTM channel | Blog cadence, ads, SEO footprint, directories, job posts | Where they spend, where they don't |40| Company shape | Registry filings, headcount signals, funding, job posts | Can they respond to us? How fast? |41| Weakness signals | Reviews (1–3 star specifically), support forums, churn complaints, stale changelog | Where the wedge is |42| Last shipped | Changelog/blog/release notes date | Stale >12mo = a sleeping incumbent = opportunity |4344**"No pricing published" is itself a finding**: it means enterprise/negotiated, which means slow45sales, which means a self-serve wedge is open underneath them.4647### Highest-signal sources, in order48491. **Their pricing page** — grade A, quote it verbatim with the fetch date.502. **Their 1–3 star public reviews** — the single richest source. Sort by lowest, read 20, and51 **tally the complaints.** Recurring complaints are the product spec for your wedge. Individual52 reviews are grade C; a tally of 20 is a real pattern.533. **Their job postings** — where they're investing next, and their stack. A company hiring 554 enterprise AEs is moving upmarket and abandoning the small end.554. **Their changelog** — velocity and direction, or its absence.565. **Their docs/API** — real capability, unfiltered by marketing.576. **Forums/subreddits where their users complain** — unvarnished, and where switching intent shows up.5859## Step 2 — the grid6061Rows = the 3–6 **buying criteria the customer actually uses** (from real discovery, not invented).62Columns = doing-nothing, the spreadsheet, and 2–4 named competitors. Cells = evidence + grade, not63opinion.6465Two rules that keep this honest:66- **Score them generously.** A teardown where we win every row is a fantasy that will be corrected67 by the market at full price. Find the rows where they genuinely beat us and write them down.68- **The criteria must come from customers.** A grid built on criteria we chose because we win them69 is marketing, not research.7071## Step 3 — find the wedge7273A wedge is a **segment + criterion where the incumbent is structurally unable to compete** — not74merely currently worse. Structural means their business model, not their backlog.7576| Wedge type | Signal | Why it's structural |77|---|---|---|78| **Segment abandonment** | They moved upmarket (enterprise hires, "contact sales", min seats) | Their cost structure can't serve small accounts profitably |79| **Price umbrella** | High price, high margin, no low tier | A cheap tier cannibalizes their core revenue — they won't |80| **Onboarding gate** | Demo-only, long implementation | Self-serve requires rebuilding the org, not a feature |81| **Neglect** | Stale changelog, unanswered support | Attention is elsewhere; won't return |82| **Complaint cluster** | Same 1-star complaint for 2+ years | If it were cheap to fix, it'd be fixed. It's architectural |83| **Geography / vertical gap** | No local presence, no vertical language | Requires local ops or domain depth they lack |84| **Integration gap** | Missing the tool your segment actually uses | Roadmap reflects their ICP, not yours |8586**Feature parity is not a wedge.** "Same but nicer" loses to distribution every time. If the only87answer is "we'd be better," there is no wedge — say so. **"Don't enter" is a legitimate and valuable88output of this skill**, and cheaper than discovering it after building.8990## Step 4 — the decision (mandatory output)9192```93TEARDOWN: <market> date: <>9495REAL ALTERNATIVES (ranked by what customers actually do)961. ... 2. ... 3. ...9798PRICE ANCHOR99<verbatim from pricing pages, with urls + fetch dates. The number a prospect100 already has in their head.>101102GRID103| Criterion (customer-sourced) | Nothing | Spreadsheet | Comp A | Comp B | Us |104105THE WEDGE106Segment: <narrow>107Criterion: <the one thing>108Why structural: <their model prevents responding>109Evidence: <grades + urls>110111PRICING IMPLICATION -> business/unit-economics112POSITIONING LINE -> design/web-page-build113KILL CRITERIA -> thinking/decision-quality114 "We abandon this if <observable> by <date>."115116WHAT I COULD NOT VERIFY117<gaps, failed fetches, JS-only pages, unpublished pricing>118```119120## Failure modes121122| Symptom | Cause | Fix |123|---|---|---|124| Grid where we win everything | Criteria chosen to flatter us | Source criteria from customers; score rivals generously |125| Only named SaaS competitors listed | Missed status quo / spreadsheet | Redo step 1 — "nothing" is usually the leader |126| Feature-list comparison | Treated it as a product exercise | Ask what the *buyer* decides on; mostly not features |127| Conclusion "we'd do it better" | No wedge found, unwilling to say so | Say "no wedge." That's the finding |128| Pricing from a review site | Not from source | Fetch their pricing page; quote verbatim + date |129| Analysis rots in 3 months | No re-check trigger | Set a re-run date and a trigger (competitor raises, launches, or reprices) |130131## Boundaries132133- **Public evidence only**: their published pages, public registries, public reviews, official APIs.134- **No competitive intelligence by deception** — no fake trial signups under invented identities, no135 posing as a customer to extract pricing, no contacting their staff under pretext, no NDA'd or136 leaked material. It is a small market; this gets found out, and it converts a competitor into an137 enemy with a story.138- A trial signup under our **real** identity to evaluate a product is fine and normal.139- No login-walled scraping of review platforms; read public pages only.140- Quote pricing **verbatim with a date** — paraphrased pricing is how a business builds a model on a141 number that was never real.
cspoerer62