Startup metrics investors want at pre-seed and seed

Most early decks have one chart going up and to the right. Investors have seen thousands of them, so their first question is rarely how big the number is. It’s what exactly you’re counting.

This guide covers the startup metrics investors want at pre-seed and seed, model by model, with benchmarks taken only from people who published them: Y Combinator, Andreessen Horowitz (a16z), Bessemer Venture Partners and a survey of growth leaders. You’ll also see how to present the numbers so they survive the second meeting, the one where somebody opens your spreadsheet.

Which startup metrics do investors want at pre-seed and seed?

Fewer than you’d think. At pre-seed, the first outside money, there may be no revenue at all, and that’s normal. What investors look for is traction, measurable proof that someone wants what you’re building, plus a sign that you know which number matters for your model.

Paul Graham, co-founder of Y Combinator, put it plainly in 2012: “The best thing to measure the growth rate of is revenue. The next best, for startups that aren’t charging initially, is active users.” YC tracks growth week by week, partly because Demo Day comes fast and partly because young startups need frequent feedback from users.

His benchmark is still quoted in accelerators everywhere: “A good growth rate during YC is 5-7% a week. If you can hit 10% a week you’re doing exceptionally well. If you can only manage 1%, it’s a sign you haven’t yet figured out what you’re doing.” Compounded, 5% a week means growing 12.6 times in a year.

Those figures describe three intense months inside an accelerator. They aren’t a law. A startup selling software to Italian hospitals is unlikely to grow 7% a week in its first year, and any serious investor knows it. What nobody accepts is a number without a definition.

The metrics investors ask for, model by model

The right metric follows your revenue model. A short version follows, with invented Italian examples that make each metric concrete.

Business model	Core metrics	What it looks like (invented example)	Common trap
SaaS (subscription software)	MRR, monthly growth, churn, net revenue retention	Milan booking software for dental clinics: €9,000 MRR from 91 clinics at €99	Counting set-up fees or annual prepayments as MRR
Marketplace	GMV, take rate, liquidity (match rate)	Naples platform for private chefs: €40,000 of bookings a month at a 15% commission	Presenting GMV as if it were revenue
Consumer app	Active users, DAU/MAU, retention cohorts	Bologna language app: 12,000 monthly and 2,400 daily active users	Showing downloads or total sign-ups
B2B hardware or services	Paid pilots, qualified pipeline, pilot-to-contract conversion, letters of intent	Turin sensors for industrial ovens: two paid pilots at €8,000 each with manufacturers near Brescia	Free pilots presented as customers

SaaS: MRR, growth, churn, NRR

MRR (monthly recurring revenue) is the monthly income from subscriptions, and ARR is its annual version. a16z’s rule is strict: no one-time fees and no professional services in recurring revenue. Its warning is specific too. People often multiply one month’s all-in bookings by 12, and in doing so count set-up fees, hardware and consulting as if they’d repeat.

Churn is the share of customers, or of MRR, you lose in a month. Say the Milan startup starts March with 80 clinics and 3 cancel: unit churn is 3 ÷ 80, or 3.75%. Now suppose two other clinics upgrade from €99 to €199. Net MRR churn falls to about 1.2%, because €200 of upgrades offsets most of the €297 lost.

Net revenue retention (NRR) takes the customers you had twelve months ago and compares what that same group pays today with what it paid then. Above 100% means existing customers spend more over time, even after some of them leave.

Marketplaces: GMV, take rate, liquidity

GMV (gross merchandise value) is the total value of transactions on your platform. Revenue is only the slice you keep, your take rate. The Naples marketplace that handles €40,000 of chef bookings at a 15% commission has €6,000 of revenue. Not €40,000.

Liquidity is how easily buyers and sellers find each other, which a16z measures with a match rate. If 320 of 500 booking requests find a chef within 24 hours, that’s 64%. a16z also expects newer user cohorts to retain better than older ones, since they join a bigger network.

Consumer apps: DAU/MAU and cohorts

For consumer products a16z looks at engagement ratios such as daily to monthly active users. The Bologna app with 2,400 daily and 12,000 monthly users has a DAU/MAU of 20%: the average active user opens it about six days a month. Write down what counts as “active”, though. a16z points out that companies have almost unlimited definitions for it.

B2B hardware and services: pilots, pipeline, LOIs

Long sales cycles mean little revenue early on, so investors look at the steps before it. A paid pilot is worth far more than a free one. A letter of intent (LOI), a signed but non-binding statement that a customer plans to buy, helps too. But a16z is clear: letters of intent and verbal agreements are neither revenue nor bookings. Show your pipeline of qualified prospects by stage and how many pilots became contracts.

What counts as good growth and good retention?

For monthly growth, a16z prefers the compounded monthly growth rate (CMGR) to a simple average. The formula: latest month divided by first month, raised to the power of one over the number of months, minus one. If the Milan startup goes from €4,000 of MRR in January to €9,000 in July, its CMGR is about 14.5% a month, a single figure that’s far easier to compare than six bumpy monthly percentages.

Retention is where early numbers tell the truth. For their 2020 benchmark, Lenny Rachitsky and Casey Winters asked 20 experienced growth leaders (from companies such as Slack, Dropbox and Uber) what good and great retention look like, and checked what they said against public data:

Business type	Good / great user retention after 6 months	Good / great net revenue retention after 12 months
Consumer social	about 25% / 45%	not given
Consumer transactional	about 30% / 50%	not given
Consumer SaaS	about 40% / 70%	about 55% / 80%
SMB and mid-market SaaS	about 60% / 80%	about 90% / 110% (land and expand)
Enterprise SaaS	about 70% / 90%	about 110% / 130%

Two cautions. These come from mature companies, so read them as a direction rather than a pass mark. And later-stage numbers are another game entirely: in Bessemer’s 2021 benchmark, built on its cloud portfolio over the previous decade, companies between $1 million and $10 million of ARR grew nearly 200% a year on average, and Bessemer describes strong retention as about 85% gross and 120% net.

How to present your metrics honestly

• Put the definition next to the number. “Active = finished at least one lesson in the last 30 days” takes one line and saves ten minutes of questions.
• Show monthly figures, not cumulative ones. A cumulative chart keeps rising even while the business shrinks.
• Use a cohort table for retention. Make one row per sign-up month and one column per month since sign-up, with each cell showing the share still active. a16z looks for retention that flattens after six or twelve months, and for newer cohorts beating older ones.
• Keep revenue, bookings and LOIs apart. Bookings are signed contracts, revenue is earned as you deliver, and LOIs are neither.
• Split paid from organic acquisition. a16z finds paid CAC more telling than blended CAC, and our guide to unit economics walks through the formulas.
• Make every number match across the deck, the financial model and the bank statements.

Where do the numbers go? Usually on a single traction slide right after the product, as our pitch deck structure shows. Investors read them alongside the team, so pair this page with our guide to what investors look for in a startup. After the round, the same metrics become the backbone of your monthly investor update.

New to all of this? Start from the beginner’s map from zero to a first round, and if you haven’t tested demand yet, validating your startup idea comes before any metric. For the full picture of a round, see our guide to raising capital in Italy.

Your metrics checklist before you send the deck

1. Pick the one metric that best shows demand for your model and write its definition.
2. Show at least six months of monthly data, if you have them.
3. Quote CMGR, not your best month.
4. Build a cohort table, even with only three cohorts.
5. Strip set-up fees and one-off projects out of MRR.
6. Label every pilot as paid or free, and every LOI as an LOI.
7. Check that each figure matches the model and the bank.
8. Have the raw data export on hand. Someone will ask for it during due diligence.