Metrics Investors Care About: What to Show and How to Defend It

Investors care most about a short list of numbers: how fast you grow, whether customers stay and spend more, how much cash you burn to get that growth, and how quickly you earn back what you spend to acquire customers. The rest of your deck supports those four answers. Below we cover each metric, what a skeptical partner reads into it, and where founders get caught out.

Trydialectic

Defend Your Metrics With Confidence

Dialectic stress-tests your pitch against tough investor questions, helping you uncover fragile assumptions and strengthen your answers before the meeting.

Stress-test your pitch

TL;DR

Growth alone no longer carries a round. Investors now weigh retention and capital efficiency together, not top-line speed.
The core set: ARR growth, net revenue retention, gross margin, burn multiple, CAC payback and runway. Know the definition, the trend and the cohort behind each.
The number matters less than your ability to explain it. Every forecast assumption should tie to an operating driver you can name.
Benchmarks differ by source, stage and sector. Treat them as ranges, and have a reason ready for any metric outside them.
Rehearse the hostile follow-up question before the meeting, not during it.

Why the list of metrics got shorter and harder
The metrics, one by one
What investors look for at each stage
Where founders get caught
How to put the metrics in your deck
Our perspective: the number is the opening, not the answer
Stress-test your numbers before the meeting
FAQ
Sources

Why the list of metrics got shorter and harder

Several years ago a steep growth line could cover a lot of sins. That has changed. SVB’s 2025 enterprise software report puts the median Series A growth rate at 69% year over year, down from 171% in 2021. The same report finds that only 13% of enterprise software companies meet the Rule of 40 (growth rate plus profit margin), compared with 31% in 2021.

Series A growth and Rule of 40 rates fell

The practical meaning: the bar for “fast” has moved, and investors are asking what it cost you to get there. Sequoia’s writing on sustainable growth describes product-market fit and positive net growth as the two foundations, and names retention as the best indicator of the first. If you grow by pouring money into a leaky bucket, the metrics will show it.

The metrics, one by one

You do not need to show all of these on one slide. You do need to be able to produce any of them within a minute when asked.

ARR or MRR, and the growth rate behind it

Annual recurring revenue (or monthly, for earlier companies) is the headline. Investors read three things from it: the level, the growth rate, and the shape of the curve. A company that grew 3x last year but flattened over the last two quarters tells a different story than the annual figure suggests, so show monthly or quarterly points.

Define what counts. Recurring revenue excludes one-off services, pilots that have not converted and usage you cannot expect to repeat. If your ARR includes signed contracts that have not started, say so on the slide. A partner who finds the inclusion on their own will discount everything else you say.

Net revenue retention and churn

Net revenue retention (NRR) measures revenue this year from the customers you had a year ago, including upgrades, downgrades and cancellations. Above 100% means your existing base grows without any new sales. It is the cleanest signal of whether customers get lasting value.

Pair it with gross retention, which ignores expansion. High NRR can hide heavy churn if a few large accounts expand quickly. Investors will ask for both, and for retention by cohort, so prepare the cohort table before they request it. If you sell to small businesses or consumers, logo churn by month of signup is often the more honest view.

Gross margin

Gross margin shows how much of each revenue dollar remains after the direct costs of delivering the product: hosting, support, payment processing, and for many AI products, model inference. It decides how much of your growth can ever turn into profit.

AI-native companies get extra scrutiny here. SVB’s H2 2025 report notes that median Series A AI companies burn about $5 for every $1 of new revenue, and the question that follows is whether your unit economics improve with scale. Be ready to show inference cost per customer over time and what changes it: caching, smaller models, pricing tiers or contract terms.

Burn multiple

Burn multiple is net cash burned divided by net new ARR added in the same period. If you burned $3M to add $2M of new ARR, your burn multiple is 1.5. Lower is better, because it tells an investor how much capital you consume per dollar of growth.

This is the number that most directly answers the shift from growth at all costs to efficient growth. Published thresholds differ by source and stage, so do not quote a single magic figure. What holds in every version: a rising multiple needs an explanation, and a falling one is your best friend in the meeting. If your multiple spiked because you hired a sales team ahead of revenue, say that, show when the hires ramp and what the pipeline looks like.

CAC payback and LTV to CAC

Customer acquisition cost (CAC) payback is the number of months of gross profit it takes to recover what you spent to win a customer. LTV to CAC compares the lifetime value of a customer with that cost. Payback is usually the more trusted of the two, since lifetime value depends on churn assumptions that early companies cannot yet prove.

Calculate CAC fully loaded: salaries, commissions, tools, content and agency fees for sales and marketing, not just ad spend. Split it by channel if you can. A blended number that mixes free word-of-mouth customers with paid ones will not survive the first follow-up question.

Runway

Runway is cash divided by net burn, but investors read it as a statement about your judgment. Sequoia’s Extending Your Runway guidance frames it with a car analogy: what matters is not the gas you have now but whether it lasts until you reach your destination. In funding terms, that means the milestone the next round will be judged on.

One practical habit: work out what it would realistically take to reach that milestone, then add a buffer. Raising well before cash gets tight keeps you from negotiating out of desperation.

What investors look for at each stage

The set of numbers you lead with changes as the company matures. This is a general pattern, not a rulebook, and individual funds weigh things differently.

Stage	What you can usually show	What the investor is really testing
Pre-seed	Waitlist, pilots, early usage, customer interviews	Whether the problem is real and you are the team to solve it
Seed	First revenue, early retention, growth month over month	Signs of repeatable demand and a believable first sales channel
Series A	ARR, cohort retention, gross margin, burn multiple, CAC payback	Whether the engine works and money will make it faster
Series B and later	NRR, efficiency, segment-level economics, forecast accuracy	Whether growth is predictable and scalable

If you are earlier than the table suggests, do not fake the later metrics. Show the leading indicators that apply to you, label them as such, and explain what you will measure next. Choosing the right program or funding path for your stage matters here too. If you are weighing structured support first, this comparison of accelerators and incubators for tech founders is a useful read on fit beyond stage alone.

Where founders get caught

Most problems in the room are not bad numbers. They are numbers that do not hold under questioning. These are the patterns we see most.

Hiding the good part inside the total

If one product line is growing much faster than the rest, break it out. In Sequoia’s conversation about MongoDB, Roelof Botha describes how the CEO isolated Atlas metrics and spotlighted them separately with his team and board so its scale was properly appreciated. Blended totals can drown out the part of the business that matters most.

Forecasts with no drivers

Sequoia’s guidance says founders should stop raising on pure story and tie each forecast assumption to a clear operating driver. A line that says revenue triples next year invites the question: from what? Convert it. How many reps, at what ramp time and quota attainment, closing how many deals from which pipeline, at what average contract value?

Then show scenarios. A downside case that you have actually thought through is more persuasive than a single optimistic line, because it proves you know which levers move the outcome.

Pricing that does not match value

SVB’s H2 2025 State of the Markets points to a shift toward a subscription base layered with consumption-based pricing. Investors look for metrics showing your pricing tracks the value customers receive. Be ready to explain why you charge per seat, per usage or per outcome, and what happens to margin when a customer uses the product heavily.

Definitions that shift between slides

A customer is a paying account on one slide and a signed letter of intent on another. Churn is monthly in the deck and annual in the model. Pick definitions, write them in the appendix, and use them everywhere. Inconsistency reads as either carelessness or concealment, and neither helps.

Benchmarks without context

Published benchmarks vary widely depending on the dataset, the year and the segment. A median from a survey of mature companies is not a target for a seed-stage startup. When you cite a comparison, say where it comes from and why it fits your business. When you fall outside a range, say why and what changes it.

How to put the metrics in your deck

Lead with the two or three that make your case. If retention is your strength, put it on the traction slide, not slide 14.
Show trends, not snapshots. Twelve months of data beats a single quarter-end figure.
Define each metric once in a footnote or appendix.
Link every forecast number to a driver and keep the model one click away.
Prepare the appendix with cohort tables, CAC by channel, margin breakdown and the scenario cases.
Mark estimates as estimates. Investors forgive early-stage gaps far more readily than inflated certainty.

Our perspective: the number is the opening, not the answer

We built Dialectic around one observation: founders rarely lose a meeting because a metric was bad. They lose ground because the first hard follow-up exposes an assumption they had never tested. Retention looks strong until someone asks how much of it comes from two accounts. Burn multiple looks fine until someone asks what it becomes when the sales hires you plan are added.

Most feedback founders get is polite. Friends, advisors and general-purpose AI tools tend to encourage. A partner on the other side of the table is paid to look for the crack. So we think the useful preparation is adversarial: find the fragile assumption behind every figure, decide whether you can defend it, and if you cannot, fix the business or the story before the meeting.

A practical test you can run today: for each metric on your traction slide, write the single most hostile question an investor could ask, then answer it out loud in under 30 seconds with evidence. Where you hesitate is where the work is.

— D

Stress-test your numbers before the meeting

If you want that hostile questioning without waiting for a real investor to supply it, you can upload your deck or memo to Dialectic. It scrutinizes the pitch the way a skeptical VC partner would, scores fragility and risk, flags blind spots, and lets you rehearse your defense against hard questions in a boardroom-style session.

Pricing as published: the Tactical Audit is $25 per audit, one-off. Founder Pro is $49 per month. Venture & Studio, with 5 seats, is $149 per month. You can also see how this works on a real document in our take-rate teardown of a board memo.

FAQ

What metrics do investors care about most at the seed stage?

At seed, investors usually care about evidence of repeatable demand: early revenue or strong usage, how many customers come back, and month-over-month growth. Formal ratios like CAC payback are less reliable with little data, so show the underlying inputs honestly. Retention of your first cohorts is often the most persuasive number you have.

What is a good burn multiple?

Lower is better, since it means you spend less cash per dollar of new ARR. Exact thresholds differ between sources and between stages, so check several recent benchmark reports rather than relying on one figure. Whatever your number, show the trend and explain what drove it.

Should I include metrics that look weak?

Yes, if an investor will ask about them, and they will. Show the metric with context: the cause, what you changed and early evidence it is working. Leaving a weak number out and having it discovered in diligence costs far more credibility than presenting it with a plan.

How many metrics should be in a pitch deck?

Feature three to six on the main slides and move the rest to an appendix. The goal is a clear story, such as strong retention, improving efficiency and a credible path to the next milestone, not a dashboard. Keep cohort data and the model ready for follow-up questions.