How to benchmark your startup idea against YC companies

For founders applying to YC, or using YC as a reference set for what a fundable idea looks like. This will not predict whether you get in.

8 min read · Updated

YC is useful as a reference set for a boring reason: it is the largest public corpus of early-stage companies with consistent, structured descriptions. More than 6,000 of them, each with a batch, an industry, a location, a stage and a one-line description of what they do.

That makes it the closest thing founders have to a comparison group. Not because YC companies are always right, but because “is my idea unusual?” is only answerable against a population.

Step 1 — find your actual comparables

Most founders look up the wrong companies. They search their industry, find the two most famous YC alumni in it, and conclude either that the space is taken or that they are the next one. Neither is informative.

Your real comparables are companies solving an adjacent problem for a similar customer at a similar stage — usually not famous, and usually from recent batches. Filter by industry and batch together, then read one-liners rather than logos.

What to look for in each comparable

  • The wedge. What narrow thing did they start with? YC companies are rarely funded on the broad vision.
  • The customer. Same buyer as you, or a different one wearing similar clothes?
  • What they left out. The one-liner tells you what they chose not to do, which is often where your opening is.
  • Batch recency. Three companies in your space across the last four batches means the thesis is live. Three across ten years means it keeps failing.

Step 2 — read the crowding honestly

Finding companies doing something similar is the outcome founders fear most and should fear least. It proves the problem is fundable — the hardest thing to establish. What matters is whether their existence closes your opening or defines it.

The genuinely bad signal is different: several companies in your exact space across many years, none of which grew. That is a pattern YC partners recognise on sight, and it usually means the problem is real but the willingness to pay is not.

Step 3 — turn it into a number that means something

“Promising” is not a finding. A percentile is, because it forces a reference set. Truewick scores an idea across ten dimensions and reports where that score sits against the cohort — so instead of a vague sense that the idea is decent, you get a position and, more usefully, a weakest dimension.

What a benchmarked result looks like

Illustrative sample — not a real customer result

71/100

A compliance-evidence tool that continuously collects SOC 2 artefacts from a startup's cloud accounts and flags drift before an audit.

Better than 81% of B2B SaaS ideas scored on Truewick

  • 9/10Problem validityAudit-prep pain is heavily documented; the cost of failure is concrete and dated.
  • 8/10Market sizeBottom-up: every funded startup selling to enterprise eventually needs this.
  • 8/10Business modelAnnual per-company pricing is established by incumbents.
  • 7/10TimingLive, but the window is narrowing as incumbents expand downmarket.
  • 4/10CompetitionWeakest dimension. Several well-funded YC alumni already here; the wedge must be narrower than 'compliance'.
  • 5/10Unique valueContinuous drift detection is a real differentiator, but it is a feature others can add.

The instruction in that scorecard is unambiguous: the idea is fundable in shape but the competitive answer is not yet sharp enough. The next two weeks go on narrowing the wedge, not on building. An investor will ask the competition question first, and “we do it continuously” is not yet an answer.

Using the YC directory directly

You can browse the full directory here by batch and by industry — useful for finding comparables by hand, and for getting a feel for how narrowly successful applications describe themselves. The one-liners are the most instructive part: almost all of them name a specific customer and a specific job, and almost none of them describe a platform.

Three things YC one-liners consistently do

  • Name the customer explicitly. Not “businesses” — “independent physiotherapy clinics”.
  • Describe one job, not a category. “Chases no-shows over SMS”, not “practice management”.
  • Avoid the word platform. A platform is what you become, not what you launch.

Frequently asked questions

Can a tool predict whether I will get into Y Combinator?
No, and be sceptical of anything that claims to. Admission depends heavily on the founding team, the interview, and traction — none of which can be read from an idea description. What you can measure is whether your idea resembles what YC funds and which dimension of it is weakest relative to companies that were accepted.
How do I find YC companies similar to my idea?
Filter by industry and batch together rather than searching your industry alone, then read one-liners instead of scanning logos. Your useful comparables are usually recent and not famous: companies solving an adjacent problem for a similar customer at a similar stage. Fame is a poor proxy for relevance.
Is it bad if YC has already funded companies in my space?
Usually it is a good sign — it proves the problem is fundable, which is the hardest thing to establish. The genuinely bad pattern is several companies in your exact space over many years, none of which grew. That normally means the problem is real but nobody will pay enough to solve it.
What does a percentile against YC companies actually mean?
It positions your scored idea against a cohort, so 'promising' becomes a number with a reference set. It is a relative measure of how the idea reads on ten dimensions, not a prediction of outcome. The more useful output is the weakest dimension, because that is the thing to go and fix.