You are about to depend on a company you do not know. Perhaps as a vendor holding your data, a partner carrying your brand, an investment, or an employer. The stakes differ; the questions barely do.
What follows is a structure for answering them from public evidence, which for most decisions is all you will get.
1. What is the business model, actually?
Not what they sell — how money reaches them, from whom, and how reliably.
The questions that matter: Who pays? Is revenue recurring or transactional? Is there a free tier, and if so what converts? Are they selling to many small customers or a few large ones? Concentration is a risk that rarely appears on a website but often shows in case studies and careers pages.
Method is covered in reading a business model from public signals.
2. Are they solvent enough to still be here?
For private companies you will not get financials. You can get proxies:
- Hiring trajectory. Steady posting suggests funded operations; a hiring freeze after a growth period is informative; a sudden burst in one function signals a bet.
- Funding history and timing. Not the amount so much as the gap. A company eighteen months past its last raise in a market where eighteen months is the norm is at a decision point.
- Filing status. In most jurisdictions, whether accounts and returns are filed on time is public and free. Late filings are a genuine signal.
- Product velocity. Changelogs, release notes and public roadmaps show whether the engine is running.
Each of these is weak alone. Together they usually converge.
3. Is the product any good, according to people who use it?
Their marketing will not tell you. Users will, in review sites, community threads, and support forums.
Read for patterns, not verdicts. Twelve reviews independently mentioning the same onboarding problem is a finding; one furious review is noise. And read the middling reviews rather than the extremes — three-star reviews are consistently the most informative, because they come from people who both use the product and see its limits. See what customers say vs what they rate.
4. Who runs it?
Leadership tells you about direction and stability. Look for tenure, background, and turnover.
The single most informative signal is departures at the executive level in a short window, particularly in finance or engineering. It is not proof of anything, and it warrants a closer look at everything else.
Also check whether the people on the About page still work there. A leadership page listing people who left a year ago says something about attention to detail, at minimum.
5. What is their public trust posture?
If they will hold your data or sit in your delivery path, this stops being optional. Do they publish a security page? Is the privacy policy specific and dated? Are subprocessors listed? Is there any evidence they have ever answered a security questionnaire?
What you are assessing is whether they have thought about it, not whether they have every certification. See questionnaire readiness and what buyers check.
6. Where do they sit competitively?
A company's position determines its future behaviour. A category leader will behave differently from a challenger under pressure — on pricing, on support, on which customers get attention.
Practical question: if this company had a bad year, what would they do, and how would it affect you? A challenger might discount aggressively and cut support. A leader might raise prices. Both are foreseeable from position.
7. What is the bear case?
Deliberately construct the argument against. Most research is unconsciously confirmatory — you have half-decided already and are gathering support.
Force the opposite: what would have to be true for this to go badly, and is any of it true now? Then do the same for the bull case, so you are not merely pessimistic. Covered in arguing both sides properly.
The order to work in
Time-boxed, roughly:
- Fifteen minutes: website, About, pricing, careers, changelog. Establishes whether the company is alive and what it claims.
- Thirty minutes: reviews and community threads. Establishes whether the claims hold.
- Thirty minutes: registry filings, funding records, leadership backgrounds. Establishes durability.
- Fifteen minutes: write the bear case. Establishes what you have been avoiding.
Ninety minutes gets you most of the available signal. Beyond that returns diminish sharply until you can talk to people directly.
What public research cannot tell you
Worth stating plainly. You will not learn actual revenue, churn, runway, internal culture, or whether the founders are still speaking to each other. Public research bounds your uncertainty; it does not eliminate it, and a report claiming otherwise is inventing.
Our Company Deep-Dive works through this structure on a single company — business model, financial signals, product, customer voice, brand, leadership, and a bull and bear synthesis — with sources cited and estimates labelled as estimates rather than presented as fact.