Competitor analysis is useful the day it changes what you do: a pricing page you restructure, a segment you concede, a weakness you attack. Anything longer than a page per competitor that doesn't end in a decision is research theater. The working framework is small — a five-competactor cap, a positioning snapshot, and a lost-deal log — refreshed quarterly, because competitors' moves matter less than why you lose deals to them. Most founders over-invest in tracking rivals' announcements and under-invest in the one data source that actually explains outcomes: the reasons prospects chose someone else, captured within a week of the loss.
This is a strategy guide based on public information and first-party sales data, not a market report.
How do you pick which competitors to track?
Cap the list at five, chosen by where you actually lose deals, not by who is famous. The useful split is direct substitutes (same job, same buyer), adjacent substitutes (same job, different form — including spreadsheets and doing nothing, which are the real market leaders in most early categories), and the incumbent you'll eventually displace. If your CRM or notes can't name the competitors you lost to in the last quarter, that log is the first deliverable of this whole exercise — before any framework. Per Bureau of Labor Statistics data on firm competition and survival, most young companies lose to status-quo alternatives rather than to named rivals, a fact worth checking against your own lost-deal log before strategizing against a visible enemy.
What goes on the one-page snapshot?
For each tracked competitor, five rows: promise (their one-line positioning, quoted from their homepage, not paraphrased by you), buyer (who signs — the actual persona their pricing implies), pricing shape (not just the number: per-seat, usage, flat, free tier), proof (the two claims they lead with — certifications, logos, benchmarks), and weakness your buyers confirm (from the lost-deal log and win notes, not your own speculation). The last row is the discipline: a weakness you haven't heard from a customer is an assumption, and it's labeled as such. Update the page quarterly in an afternoon; if it takes longer, you're writing an encyclopedia.
How does the lost-deal log work?
Within a week of every loss, record four fields: competitor chosen, stated reason, competing price if known, and the buyer's words on why — even one sentence. Ten entries in, patterns appear that no framework supplies: maybe you lose on implementation time, not features; maybe the losses cluster when the economic buyer is finance rather than operations. Win notes matter equally — why prospects chose you predicts positioning better than why others lost. This log is also the honest input for the only two-by-two worth drawing: plot competitors on the two axes your buyers actually voice (commonly "time to value" and "depth of control"), and check that you occupy a corner a real segment wants, rather than the crowded middle.
What decisions come out of it?
- Steal: a pricing structure, an onboarding step, a proof point rivals use that your buyers echo — adopt, don't invent.
- Attack: a confirmed weakness competitors can't fix quickly (usually their architecture or their business model, not their features).
- Concede: segments or use cases where an incumbent is structurally better; withdrawing focus is a strategy, not a defeat.
- Differentiate: the axis where you deliberately won't match — your positioning sentence lives here.
Each quarter, the snapshot review should end with at most three of these calls. A competitor analysis that produces no changes produced nothing.
What are the classic failure modes?
Feature-table anxiety — the spreadsheet comparing forty checkboxes, which treats product surface as strategy and always concludes "add more features." Announcement-tracking — reacting to every rival's launch as if press releases were losses. And the newest variant, model-answer envy: quoting a competitor's marketing claims about AI capabilities as established fact instead of, per basic editorial discipline, attributed claims with dates. The antidote to all three is the same: your buyers' words outrank rivals' words, and rivals' words outrank your guesses.
How often should you actually do this?
Quarterly for the snapshot and the decision review; continuously for the lost-deal log (it's five minutes per loss); never as a distraction from selling. Before a fundraise or a category shift, run one deeper pass — a strategic review of the top two rivals' likely responses to your next move. Then close the document and go run the plays it named. The point of knowing the field is to stop looking at it every day.
For more context, read Second-Mover Advantage: How Late Entrants Win the Markets Pioneers Opened.
For more context, read business model innovation.
For more context, read Strategy Is Mostly Saying No: A Founder's Focus System.