Is that competitor actually doing well?
A competitor with a beautiful site, a busy feed and confident copy may be thriving, or may be three months from quietly closing. You usually cannot tell from the things they control. This is how to read the signals they do not control.
The question matters because it decides different things. If the incumbent is genuinely strong, entering next to them means finding an edge they cannot easily cover. If they merely look strong, the market may be wide open and you have been talked out of it by a design system.
The trap is that everything a company presents about itself is marketing. Polish is cheap now. A landing page, a brand, a steady content feed and a confident "trusted by teams everywhere" line can all be produced in a week by one person with no customers. None of it is evidence. You want the signals that are expensive to fake — the ones that cost real money, leave dated public records, or would embarrass them if they were false.
The signals worth reading
Hiring
Job postings are the strongest public signal most small companies emit, because a role costs real money and nobody advertises one they cannot fund. Read them for three things: whether they are hiring at all, what functions they are hiring into, and how long the posts sit open. A company staffing up support and delivery is being pulled by demand. One hiring only sales is buying growth it is not getting organically. A single reposted role that has been open for months usually means either a stalled budget or a hard-to-fill spec, and it is worth knowing which.
Shipping cadence
Changelogs, release notes, app-store version histories, public repositories and dated documentation updates all carry timestamps that are awkward to fabricate. What you are reading is not the features — it is the rhythm. Regular, unglamorous releases mean somebody is being paid to maintain the thing. A gap of many months, followed by a burst of marketing activity, is a familiar pattern: the team went quiet and then tried to relaunch. Compare their most recent substantive change to their most recent promotional post. When the marketing is much fresher than the product, the product may be coasting.
Pricing moves
Pricing pages are archived and dated by public web archives, so their history is readable. Prices creeping up, minimums appearing, the cheapest tier quietly disappearing — those are the moves of a company with more demand than capacity. Prices falling, discounts becoming permanent, a free tier appearing late, annual-only deals pushed hard — those are the moves of a company buying cash flow or filling a funnel. Neither is proof on its own. Both are directional, and both are far more honest than the copy on the same page.
Distribution
Ask where their customers would actually come from, then check whether that channel shows any trace of them. If they sell through search, do they rank for anything a buyer would type, or only for their own name? If they sell through a marketplace or app store, is there a review history accumulating over time? If they sell through partners or integrations, are they listed in those partners' public directories? A company that appears everywhere in its own channels and nowhere in anyone else's has not yet solved distribution, whatever the site suggests.
Third-party traces
The most reliable evidence is the evidence they did not publish. Customers describing them in public. Reviews accumulating steadily rather than in suspicious clusters. Being mentioned in a comparison written by somebody with no stake. Job candidates, ex-employees, integration partners and journalists all leave traces. Volume matters less than independence and recency: five unprompted mentions this quarter tell you more than two hundred from three years ago.
Signals that lie
Some of the most quoted numbers are the least informative, and knowing which saves you from confident wrong conclusions.
- Follower counts and engagement. Purchasable, inheritable from a founder's personal audience, and frequently uncorrelated with revenue. A large following with no comments from anyone who sounds like a customer is a vanity asset.
- Third-party traffic estimates. These are modelled from panels, not measured. They are useful for coarse relative comparison — is this site an order of magnitude bigger than that one — and misleading the moment you treat the specific figure as fact.
- Funding announcements. Money raised is not money earned. A raise says investors believed a story at a point in time; it says nothing about whether customers did. It does reliably tell you the company has runway to spend against you for a while, which is a real strategic fact of a different kind.
- Logo walls. "Trusted by" often means one pilot, a trial, or a single user at a large company. Treat a logo as a claim, not a customer, unless there is a named case study behind it.
- Awards and badges. Many are paid placements or category-of-one wins. Check who issues them before assigning any weight.
Triangulate before you conclude
No single signal is load-bearing. The method is to gather several independent ones and see whether they agree — and to take disagreement seriously rather than picking the signal you liked.
A coherent picture of health looks like: consistent shipping, hiring into delivery, prices holding or rising, fresh third-party mentions, reviews arriving steadily. A coherent picture of trouble looks like: nothing shipped in a long while, no open roles or repeatedly reposted ones, discounting that has become permanent, third-party mentions all more than a year old, and a marketing feed that is livelier than the product.
The most common real result is neither: a company that is fine. Modest, stable, serving a slice of the market adequately, not growing fast, not dying. That is a perfectly good answer, and it usually means the opening against them is a specific unserved segment rather than the whole category.
Two disciplines make the difference between research and vibes. Date everything — a signal without a date cannot be distinguished from history. And write down what would change your mind before you go looking, because it is remarkably easy to assemble evidence for a conclusion you had already reached about a competitor you already admired or dismissed.
What you can honestly conclude
Public signals support statements about direction and about whether a company is being actively invested in. They do not support statements about revenue, profitability, or customer count. If you find yourself writing a specific number you did not get from the company or a filing, you have crossed from research into invention — and a number with a false precision attached is more dangerous than no number at all, because you will plan against it.
The honest output is a sentence like: the evidence is consistent with an actively maintained business that has not solved distribution beyond its own channels, and there is no public evidence of growth in the past year. That is genuinely decision-useful. "They are doing about $2M" is not, unless they said so.
The short version
- Ignore everything the company controls the presentation of. Polish is cheap; traction is not.
- Read the expensive signals: hiring, shipping cadence, pricing history, distribution footprint, independent mentions.
- Discount followers, traffic estimates, funding, logo walls and paid badges.
- Require several independent signals to agree before you believe a story, and date every one of them.
- Decide what would change your mind before you look.
- Conclude about direction and investment, never about revenue. Do not invent a number.
Want this done for a named competitor?
Fable's $5 Competitor Teardown gathers exactly these public signals on one competitor and reports what the evidence supports — what they promise, who they serve, where they look exposed — with sources and dates, and with the gaps stated as gaps. The $9 Market Scan runs the same read across a niche when you need the shape of the field rather than one rival.