What Investors Look At in Your Product Before Series A

The deck gets you the partner meeting. The cohort chart decides the round, and the cohort chart is made by the product.

At Series A, investors judge the product mostly through its data and its customers: retention by cohort and segment, usage that drives expansion, and two or three customer calls before the term sheet. The demo matters less than whether the product can prove those numbers on request.

If your raise is inside the next 12 months, the product is already part of the pitch. Here is the checklist, with what each item asks of the product itself. The longer version of the argument is on our product design before Series A page.

The window you are in

Carta’s State of Private Markets for Q1 2026 puts the median time from seed to Series A at two years. The median company raising a Series A that quarter was 2.8 years old.

The median Series A closed at $63M post-money. Split by type: $55M for non-AI companies and $65M for applied AI. Series A pre-money valuations fell 7.1% year on year.

So the round is two years out from seed on average, and companies without an AI story are priced lower for the same stage. Twelve months before the raise, the evidence has to exist already. It cannot be built in the data-room month.

The checklist

1. A KPI sheet with two years of history

Point Nine’s Christoph Janz asks for one or two pages of monthly KPIs covering about two years (October 2024). His warning is worth taping to the wall: big changes or sudden improvements look suspicious.

What it asks of the product: event tracking that has not changed definitions every quarter. If you re-instrument activation in month 11, your best chart is the one investors trust least.

2. Cohorts by segment, with excellent retention in at least one

Janz wants revenue cohorts that do not mix monthly and annual plans, excellent retention in at least one segment, and a credible path to 100% net dollar retention.

What it asks of the product: you know which segment loves it, and the product is built for that segment first.

3. Expansion that comes from real usage

CRV’s criteria for B2B and AI companies (March 2026) ask for clear proof that expansion is driven by usage, and for usage data showing the product is embedded in daily workflows.

What it asks of the product: a pricing and packaging model that grows when usage grows, and screens that make the next level of use obvious.

4. Unit economics you know cold

The same CRV page treats a CAC payback of 12 to 18 months as acceptable, and expects founders to know these numbers without opening a spreadsheet.

5. Customers who say the right thing on a call

SaaStr’s Jason Lemkin describes 2 to 3 customer calls before the term sheet, to confirm customers love the product, and as many as possible after it (February 2025).

What it asks of the product: the product shows up in diligence as what your customers say about it. If onboarding still needs a call with your CTO, they will mention it.

6. An activation event that is defined and stable

Investors read retention. Retention starts at activation. If you cannot say which action defines an activated account, and how many reach it within a fixed window, the cohort chart has no starting line. Our guide on measuring activation and time to first value walks through the setup.

7. For AI products: retention after the tourists leave

a16z recommends measuring AI retention from month 3, because early users who are only curious churn in the first quarter. The ratio of month 12 to month 3 retention is their leading signal for net dollar retention above 100% (September 2025).

CRV adds the wrapper test: would the product still exist without the AI component? It also asks for pricing aligned with inference costs, and for compute costs documented in the data room.

Bessemer’s State of AI (August 2025) names context and memory as the likely new moats. That is a product design decision, not only a model decision.

8. A data room with a demo or sandbox

Data-room guidance from Peony (October 2026) lists a demo video or interactive sandbox, a technical architecture overview, a 12-month roadmap and cohort data.

What it asks of the product: a sandbox with realistic demo data, not a staging server with “test test test” in every field.

Do investors click through the product?

Honest answer: some do, and nobody has measured how many. I found no survey of how often Series A investors use the product themselves.

What the sources do show is that the product reaches investors through two channels. One is the data. The other is your customers.

AI changed the first impression. A prototype tool can now make a product look like Series A in an afternoon. A cohort chart cannot be generated that way, and experienced investors know which of the two to look at.

Signals that make a product read as a prototype

This list is my view from 20+ years designing for startups and scale-ups, not a VC survey.

Two modules that look like they were built by two different companies
Empty states that show nothing and suggest nothing
Onboarding that only works with a founder on the call
No roles, permissions or admin view in a product sold to teams
No audit trail in a FinTech or HealthTech product
Settings pages that expose internal names for things

Each one is cheap to fix and expensive to explain in a partner meeting.

The board slide version

The deck says Series A. If the activation rate says seed, that gap is not a design opinion. It is a line on the next board slide.

The Clarity Sprint exists for this window: two weeks inside the product, your metric against its category benchmark, and every fix ranked with a conservative ARR figure. €5,500 to €8,500. You keep the report whether or not you continue.

If an investor asked for activation by segment tomorrow, could you export it within an hour?

Let’s check before they do. Book a free 30-min call, or read more about product design before Series A.

Sources

What Investors Check in Your Product Before a Series A