How to build a competitive analysis framework

What a competitive analysis framework actually is

A competitive analysis framework is a repeatable way to answer three questions: who do we actually compete against, how are we genuinely different, and which deals do we win or lose against each. The word that matters is repeatable. A one-time deck goes stale the week after it ships; a framework is a process you re-run.

The output is two artifacts, and only two count:

• Battlecards — one page per competitor, held by a rep on a live call.
• A landscape 2×2 — the map that shows where you sit relative to everyone else.

Everything else — the research, the monitoring, the spreadsheets — exists to produce those two. If your analysis cannot be used in the ten seconds before a prospect asks "how are you different from X," it has failed its only real test.

Why most competitive analysis is theatre

Most competitive decks are self-portraits dressed as maps. They pick the axes where the company wins, plot the logo comfortably in the top-right, and get presented once at an offsite before dying in a shared drive.

The failure has two causes. First, the axes are chosen to flatter rather than to reflect what buyers weigh. Second, the artifact is built for an audience — leadership — that is not in the room when the deal is won or lost. The rep on the call has nothing usable.

A framework fixes both by starting from evidence and ending at the point of use. It maps the competitors buyers actually mention, on the axes buyers actually weigh, into an artifact the person facing the buyer can hold.

There is a second way it goes wrong: treating the competitor as the thing to optimize against. Copying a rival's marketing assumes they know what they're doing — they may be burning money you'd only be amplifying. Building features to "keep up" hands your roadmap to someone else's product team. And leaning on analyst reports imports a consensus that forms after the market has already moved. Competitive analysis is ancillary intelligence, not a strategy generator — the customer is always the stronger signal.

How to pick the real competitor set

The reliable competitor list comes from your deals, not from your fears. Pull recent won and lost opportunities and read what the buyer said they were comparing you against.

That list almost always includes three kinds of rival teams forget:

Competitor type	Example	Why teams miss it
The status quo	"We'll keep doing it in spreadsheets"	It has no sales team, so it feels invisible
The in-house build	"Our engineers could build this"	It is not a company, so it is not on the radar
The adjacent tool	A product bought for something else	It solves the pain as a side effect

The status quo and the in-house build are frequently the two you lose to most, and neither appears on a traditional competitor grid. If buyers name them in deals, they belong in your analysis.

Use the customer–problem–solution test to scope the set

Reading your deals tells you who shows up; a simple test tells you how to categorize them and where the non-obvious threats hide. Every competitor overlaps you on some combination of three things — the customer they serve, the problem they solve, and the solution they use to solve it:

Overlap	Relationship	Example (for a ride-hailing app)
Customer + problem + solution	Direct	Lyft
Customer + solution, different problem	Adjacent — could pivot in	DoorDash
Problem + solution, different customer	Flanking — could expand across	a kids-transport service
Customer + problem, different solution	Alternative	a taxi dispatcher — or just walking

The dangerous entries are the last three. A player already selling to your customer, or already using your solution for a different problem, has the lowest cost to enter your space — customer access and shared technology are exactly what make an expansion cheap. Keep the set to the three-to-five that genuinely threaten you, and deliberately mix a market leader with a smaller, younger player: position is relative, and you can only tell whether you are gaining or falling behind by calibrating against both.

How to build it from win/loss evidence

The trustworthy version of every competitive claim comes from win/loss analysis — a review of why specific deals were won or lost, in the buyer's own words. Assumption-based competitive claims are how reps end up confidently wrong on calls.

For each competitor, collect the evidence and reduce it to three honest lines:

1. Their real strength — the thing they genuinely do better. Naming it builds your credibility; pretending it away destroys it.
2. Their real weakness — the gap your best customers cite when they leave the competitor for you.
3. The one true sentence — a claim a rep can say out loud that survives fact-checking.

That third line is the discipline. A battlecard full of exaggerations gets a rep caught, and a caught rep loses the deal on trust, not features. The single strongest true sentence beats five shaky ones.

What to gather on each competitor — and where

Win/loss tells you why deals move; a fuller profile tells you what you're up against. For each competitor in the set, reverse-engineer four things:

• The company shape — headcount, funding, investors, acquisitions, customer count. Not trivia: at the competitor's current age, their numbers are the benchmark for judging your own growth, and an acquisition is a telegraphed move — a web-analytics company buying a mobile-monitoring one is announcing where it's headed.
• Their go-to-market — who they target, how they position, and the exact words their homepage and press releases repeat. The "About" boilerplate is the message they most want a prospect to internalize, and almost no one reads it as closely as you should.
• Product and pricing — core features, what's genuinely unique, the minimum price and the main price driver (seats, usage, volume), and how hard they are to leave.
• The customer experience — mined from reviews (G2, Capterra, TrustRadius) and, best of all, your own churn and loss interviews. Reviews also leak trigger events — reviewers often say what changed in their world the week they went looking.

One underused source deserves a callout: Glassdoor. Anonymous employees describe a competitor's real culture and internal strain more honestly than any press release — and organizational weakness is a competitive fact.

How to choose the axes of your 2×2

The landscape 2×2 plots every player on two dimensions. Its entire value depends on choosing the right two — the dimensions your buyers weigh when they decide, not the ones where you happen to win.

Find the axes by listening to the trade-off buyers describe. In most markets it sounds like a tension: "we wanted something powerful, but everything powerful was hard to set up." That sentence hands you both axes — power against ease — and the quadrant nobody occupies is your opening.

Rules for an honest 2×2:

• Axes must be things buyers trade off, not features you list.
• Plot competitors by evidence, not by wishful placement.
• You do not have to be top-right. Owning a defensible corner buyers care about beats crowding into a corner nobody weighs.

A 2×2 built on axes buyers ignore is a flattering picture. A 2×2 built on the axes buyers weigh is a map you can navigate.

Score the threats so you know who to worry about first

Not every competitor deserves equal attention, and a battlecard for a player nobody actually loses to is wasted effort. Rank the set by threat, judged on five dimensions:

• Customer overlap — how much of their ideal customer is also yours.
• Growth velocity — headcount, funding, and logos added over the last year.
• Moat strength — how defensible their position is (below).
• Product adjacency — how little work it would take them to ship into your wedge.
• Distribution power — how much customer access and channel reach they already have.

Weight them to taste — customer overlap and growth usually matter most — but the point is the ranking, not false precision. Sort the result into three tiers: active threats that earn a battlecard and close monitoring, watch-list players you revisit each quarter, and peripheral names you note and ignore until something changes. This is a working rubric, not a law of nature; its only job is to force the question "who first?" to have an answer.

Read the moat, not just the feature list

Features get copied in a quarter; moats do not — which makes a competitor's defensibility more strategically important than their current feature set. Assess each rival, and your own concept, on the durable barriers — Tom Tunguz's three, plus two more:

• Data network effect — does the product get smarter the more it's used?
• Network effect — does it get more valuable as more people join?
• Ecosystem — do third parties build on top of it?
• Switching costs — deep integrations, migrated data, and retraining make a customer expensive to win and expensive to lose. Slack shows two barriers at once: network effects from word-of-mouth and an ecosystem from its developer platform.
• Brand / category ownership — is the competitor the name buyers say when they describe the category itself?

A rival with none of these is exposed no matter how polished the product; a rival with two or more you attack on the flank, not head-on. Run the same scan on your own concept and it tells you whether your strategic question is "can we win" or the harder "can we defend it once we do."

Compete differently against different business models

How a competitor sells shapes how you beat them. Joel York's three SaaS models each demand a different counter:

• Self-service — low price, low friction, the bottom of the market. Switching is trivial, so you rarely win on features; you win on relationship or brand, and you guard your acquisition-to-lifetime-value math, because they are guarding theirs.
• Transactional — mid-market, demos and trials expected, buyers who want some hand-holding, where most SaaS lives. The contest is the sales and success experience across the customer's lifetime, not a single feature bake-off.
• Enterprise — high price, long cycles, high touch. You compete on the depth of the relationship and the safety of the choice — while watching the self-service tier below you, because that is where disruption enters a category: from underneath.

Misread the model and you fight on the wrong ground. Trying to out-feature a self-service tool, or out-price an enterprise incumbent, loses to a competitor who simply understood which game they were in.

Find the whitespace — where the analysis becomes a bet

The competitive set tells you where everyone already is; the opening is where they aren't. Take the eight-to-twelve jobs your buyers are actually trying to get done — pulled from reviews, interviews, and sales calls — and score each on two things: how much customers care about it, and how well existing players already serve it. Plot them, and four regions appear, but only one matters: high importance, low coverage — the job buyers feel and no one solves well. That is your wedge.

The discipline is resisting the other three corners. A job everyone already serves well is table stakes — necessary to compete, useless as a difference. A job nobody cares about is a distraction no matter how empty the field. Build where customers care and rivals are weak — and, the strongest signal of all, where that whitespace lines up with the empty quadrant on your landscape map. When two independent views point at the same gap, the bet is as confirmed as competitive work gets.

How to write a battlecard reps actually use

A battlecard is a one-page, per-competitor answer sheet for a live conversation. Structure each one so a rep can scan it mid-call:

• Their pitch — how they position against you, in one line.
• Your counter — the true difference, stated plainly.
• The trap to set — the question that exposes their weakness without you naming it. "Ask how long their setup took" lands harder than "they're slow to set up."
• The objection to expect — the pushback this competitor's champions raise, and the answer.

Keep it to a page. A battlecard that runs long is a battlecard that stays closed. And ship it into live calls before you trust it — the objection handling that reads well in a doc often collapses the first time a real prospect pushes back. Sit in on three calls and revise what did not survive.

Keeping it alive

Competitive intel decays. A competitor drops a price, ships a feature, or a new entrant starts appearing in deals, and every battlecard citing the old reality now makes your reps sound out of date. A rep who quotes a price the rival cut six months ago loses credibility instantly.

Refresh lightly every quarter, and immediately on any material move. Name the moves that force a refresh so the doc can't quietly rot: a competitor raises a round or announces an acquisition, hires a VP of Product who signals a pivot, ships a major release, or starts appearing in three or more of your deals. And if you lose two deals to the same rival in ninety days, the trigger is not a light refresh — it is a battlecard rewrite and a hard look at whether your positioning still holds.

This is exactly the maintenance load that makes competitive analysis a natural fit for tooling — monitoring is mechanical and constant, while the judgment about which change matters stays yours.

The point is not the competitor

It is worth ending where this discipline can go wrong. The purpose of all of it is not to obsess over rivals — a company that steers by its competitors builds a slightly different version of someone else's product. The purpose is to know exactly where you sit so you can find the ground you can own. Focus on the customer, not the competition; done well, competitive analysis is just how you learn to out-serve customers — theirs and yours — in the specific ways your rivals structurally can't.