At pre-seed, the numbers you track should help you understand whether you’re getting closer to a business worth building. Are you solving a problem that a specific customer experiences often enough, and painfully enough, to seek a better option? Does your product help them achieve something they couldn’t easily do before? Will they keep using it, recommend it, or pay for it?
The answers won’t come from one universal KPI. A B2B SaaS company testing its first design partners has different evidence available than a consumer app trying to establish a repeat behavior. Both, however, need to replace assumptions with observed customer behavior.
Start with the uncertainty that could most easily undermine the company. You may need to establish that the problem is actually real, that people will change an existing workflow, that they can reach value in an early product, or that the value is significant enough to justify a budget. The metrics you choose should show whether that uncertainty is shrinking.
This article covers the signals that help pre-seed founders evaluate customer problems, demand, product value, repeat use, willingness to pay, and runway. The measures will look different across companies, but each should give the team evidence it can use in a real decision.
As always, if you have any questions about anything you read, email us at [email protected], and we’ll help you understand.
Start with the uncertainty that could derail the company
Founders often inherit a list of startup KPIs before they have a product motion that makes those KPIs meaningful. They start tracking total signups, monthly active users, traffic, social engagement, or projected lifetime value because those are the numbers they expect investors to ask about.
Some of those figures may become important. Early on, though, a number only deserves attention if it changes a decision. If it won’t affect who you talk to, what you build, how you price, where you spend time, or when you raise, it can wait.
For most pre-seed companies, the first decisions are straightforward, even if the answers aren’t:
- Which customer has the sharpest version of this problem?
- How are they dealing with it today?
- What has to happen for them to see value in the product?
- What behavior would show that the value is real?
- What evidence would justify investing more money and time in this direction?
The measures below help answer those questions. They won’t prove every part of the business at once, but they have potential to show whether the company is getting closer to a product and customer model worth building on.
Track the problem before you track growth
The earliest evidence often comes from customer conversations, but the goal will never be to maximize the number of interviews over learning whether people who share a relevant role, context, or workflow describe a problem with similar urgency.
Track the conversations that give you that evidence:
- Qualified customer conversations: How many conversations have you had with people who actually fit the customer profile you’re trying to serve?
- Problem frequency: Does the problem happen every day, every week, during a high-stakes event, or only occasionally?
- Cost of the status quo: What does the current workaround cost in time, money, missed revenue, risk, or frustration?
- Existing workarounds: Are people relying on spreadsheets, internal processes, agencies, competing tools, or manual work to get the job done?
- Pattern convergence: Are people independently describing the same situation, consequence, and workaround, or are you hearing a collection of unrelated complaints?
A founder building software for operations teams might hear that customers spend hours each week reconciling data across systems. That is a useful observation. It becomes more valuable when the same type of customer can explain what those hours delay, who is responsible for the work, what they have already tried, and why the problem hasn’t been solved internally.
The difference is important. “That sounds useful” is feedback. “We lose a day before every reporting deadline and have already hired someone to manage the spreadsheet” is evidence of a costly, recurring problem.
Look for commitments, not polite interest
Interest is painfully easy to collect. A prospect can praise an idea, join a waitlist, or say they would use the product without changing anything about their current behavior.
Commitment asks more of them. It costs time, access, reputation, money, or all four.
Early demand measures should reflect that distinction:
- Qualified waitlist conversion: Of the people who sign up, how many match your intended customer profile and take a meaningful next step?
- Demo-to-pilot conversion: How many qualified prospects agree to test the product in a real workflow?
- Design-partner commitments: How many customers agree to provide regular feedback, share relevant context or data, introduce colleagues, or evaluate the product over time?
- Time to commitment: How long does it take a qualified prospect to move from an initial conversation to a pilot, evaluation, or other substantive next step?
- Referrals and introductions: Do early prospects introduce you to peers who have the same problem?
- Inbound questions with stakes: Are people asking about pricing, integrations, implementation, security, or availability? These questions don’t guarantee a sale, but they often reveal that a prospect is considering how the product would fit into real work.
A broad waitlist can still be useful for learning. It becomes misleading, however, when it’s presented as proof of demand without knowing who joined or what they did next.
To demonstrate what we’re talking about, let’s compare two early signals. One thousand people submit an email after a social post reaches a broad audience. Ten heads of operations from similar companies book a call, explain the same painful workflow, and agree to test an early version with their teams.
The first number is larger, sure, but the second one gives the founder a clearer direction for product, positioning, and sales.
Define the moment when users receive value
Once people can use the product, total accounts and logins won’t tell you much on their own. You need a definition of activation that reflects the product’s promised outcome.
Activation should reflect the product’s promised outcome. In a B2B workflow tool, that might mean connecting a data source, completing the first core workflow, and sharing the result with a teammate. A marketplace may define activation as a buyer receiving a viable response to a request or a supplier completing a first transaction. For an AI product, activation may occur when a user generates an output they keep, export, share, or use to complete a real task.
The definition should be specific enough that you can recognize success, but it shouldn’t be so elaborate that a user has to complete a full onboarding checklist before they count.
Once you have that definition, track:
- Activation rate: What percentage of new users reach the value event?
- Time to value: How long does it take users to get there?
- Activation by customer segment: Does the customer you intended to serve reach value more often than everyone else?
- Activation blockers: Where do people abandon the process, and what do they say prevented them from continuing?
- Depth of early use: Do activated users complete the core action once, or do they begin using the product in a live workflow?
Activation is where positioning, onboarding, and product design meet. If the intended customer isn’t reaching value, the issue may not be a single feature. The product may be promising the wrong outcome, asking for too much setup, or attracting people who were never likely to benefit from it.
Watch whether value holds up over time
Retention is often treated as a later-stage metric, but pre-seed teams need an early view of whether the product has a place in a customer’s life or workflow. The relevant time frame depends on how often the underlying problem occurs.
A consumer product built around a daily habit should look for returns quickly. A B2B product used for monthly close, quarterly planning, or an infrequent compliance task shouldn’t be judged on daily active users. In each case, the question is the same: when the customer encounters the problem again, do they come back to your product?
Track retention in ways that match that behavior:
- Repeat core actions: How many activated users complete the central task more than once?
- Cohort retention: Of users or accounts that started in the same week or month, how many are still active in a later period?
- Usage cadence: Does activity match the natural rhythm of the problem you’re solving?
- Team adoption: Has use expanded from one champion to colleagues, additional seats, or another workflow?
- Churn and disengagement reasons: When people stop using the product, was the issue value, usability, trust, timing, missing functionality, or a change in their circumstances?
You don’t need perfect retention before you can learn from it. A small cohort can show where the product creates lasting value and where it breaks down. The important thing is to avoid treating a return visit as proof without understanding what brought the user back.
Test willingness to pay before pricing feels finished
We’re happy to tell you pre-seed founders don’t need a polished pricing model to learn whether the product has commercial weight, but they do need to move beyond hypothetical questions as soon as they can.
A paid pilot, deposit, implementation fee, signed letter of intent, or explicit budget conversation can all reveal more than a customer saying they would pay someday. Each comes with limits, and none is a substitute for sustained use. Still, they show whether the problem is important enough for a customer to commit resources.
Track the evidence that emerges during the sales process:
- Paid pilots and early revenue: How many customers have made a financial commitment?
- Pilot-to-paid conversion: When a pilot ends, do customers want to continue?
- Average pilot or contract value: What does the customer consider the problem worth solving?
- Sales-cycle length: How long does it take to move from first conversation to a decision?
- Win and loss reasons: Why did a customer choose the product, keep the status quo, or select an alternative?
- Pricing and procurement friction: Are objections really about price, or do they point to trust, implementation work, budget ownership, security, or timing?
- Expansion signals: Does a successful early user ask to add seats, bring in a second team, or apply the product to another use case?
A small number of paid customers doesn’t prove a repeatable sales motion. It can, however, reveal who buys, what they buy for, what slows the process down, and where the product’s value is strongest. Those are the ingredients of a more credible go-to-market strategy.
Keep runway tied to a milestone
Cash and burn aren’t just finance numbers. They determine how long the team has to turn uncertain signals into stronger evidence.
Track:
- Cash on hand
- Monthly net burn
- Runway in months
- Committed revenue versus projected revenue
- Major planned expenses, including product development, hiring, and customer acquisition
- The milestone the current cash is meant to fund
You can calculate runway with a simple formula:
We love a formula, but we love the connection between the number and the work ahead more. If you have eight months of runway, what should be true by the end of those eight months? You might need to convert five design partners into paid customers, demonstrate repeat use among a narrow customer segment, or prove that a specific acquisition channel consistently produces qualified conversations.
Without that milestone, a runway figure is only a countdown. With one, it becomes a planning constraint that helps the team decide what not to do.
Metrics that can wait
Some numbers may be directionally interesting at pre-seed, but they can create false confidence when they aren’t connected to a real product or customer motion.
Be careful about making too much of:
- Total signups without activation or repeat use
- Website traffic without evidence that the visitors fit the intended customer profile
- Social reach, media mentions, or follower count as proof of demand
- Customer acquisition cost before you have a stable channel and enough conversion data to interpret it
- Lifetime-value estimates based on assumptions rather than observed retention and revenue
- Annualized revenue projections based on a handful of unusual early deals
- A single growth benchmark that ignores product cadence, buying cycle, customer type, and market conditions
These metrics aren’t inherently bad. They become more useful when the company has a clearer customer, a defined value event, enough usage history to observe patterns, and a more repeatable way of finding demand.
Review what changed each week
A weekly review should show what the team learned, where the evidence is weak, and what needs to change. It shouldn’t become a ritual for reciting activity numbers.
A simple review can cover six areas:
The review should reveal movement that changes the plan. Consistent feature requests before a prospect will agree to a pilot may signal a gap in the roadmap. When activated users return but don’t invite a teammate, the product may not yet support the way collaboration works for that customer. A customer segment that closes quickly, while another needs months of education, gives the team a reason to focus its sales effort where urgency already exists.
The goal is better evidence
At pre-seed, the numbers you track won’t remove uncertainty, but mercifully they can show you where it still lives.
Track the evidence that reveals whether a specific customer has a costly problem, whether the product gives them a meaningful outcome, and whether that outcome is strong enough to create repeat use or willingness to pay. Pair those signals with a clear view of runway, so the team knows how much time it has to strengthen them.
The next product choice, customer conversation, and capital decision should rest on clearer evidence than the last. The measures that belong in this article are the ones that make that possible.