KPI discipline before a seed round

Going into a seed round with good KPI discipline does not mean having a dashboard full of charts. It means you can answer three questions without hesitating: which few numbers describe your business, exactly how you calculate them, and what you decided in recent months by looking at them. An investor who asks two follow-up questions about a metric can tell within minutes whether the team runs that number or is run by it. The discipline is built before the round, with simple rules and a steady rhythm, not the week before you open the data room.

In short

• Pick three to five metrics tied to how you create value, not everything your software lets you measure.
• Write a definition for each metric: formula, data source, what is included and what is excluded.
• Review the numbers every week at the same time, with one owner per metric.
• Keep a log of the decisions you made from the data: it is the most convincing proof of discipline.
• Never change definitions silently: if they change, note when and why.

Why discipline matters more than absolute values

At pre-seed and seed, the numbers are almost always small. Investors at this stage know that and do not expect large volumes. What they can assess is the quality of how you work: whether the numbers are consistent with each other, whether the definitions stay stable over time, whether you can tell a signal from a one-off.

A startup with ten customers and a clean table, updated every week for six months, signals reliability. A startup with a hundred users and three different versions of the same number across the deck, the spreadsheet and the CRM signals the opposite, even with the larger volume. KPI discipline is, at bottom, a signal about the team's ability to run a company, before it is a signal about the business itself.

Choosing the few right metrics

Start from the business model, not the tools. Ask which chain of events leads a customer to pay and to stay. For B2B software the chain might be: qualified contact, demo, pilot, contract, active usage, renewal. For a marketplace: listing published, demand, transaction, repeat purchase.

From that chain, choose:

• one value metric, showing whether customers really get what you promise (for example weekly use of the core feature);
• one growth metric, showing whether the base is widening (new paying customers, recurring revenue);
• one efficiency or cash metric, showing what that growth costs you (monthly cash burn, months of runway);
• at most one or two process metrics, tied to the current bottleneck.

If a metric has never changed one of your decisions, it probably does not belong on the main list. It can live on a secondary sheet.

Write the definitions before you look at the numbers

The most common problem is not missing data but ambiguity. "Active customer" might mean someone who logged in, someone who used a feature, or someone who paid the last invoice. Each choice is legitimate, but it has to be fixed in writing.

For each metric, prepare a short card with:

1. name and purpose in one sentence;
2. the exact formula, with numerator and denominator;
3. the source: which system the data comes from (accounting software, CRM, product database, bank);
4. inclusions and exclusions: test users, internal accounts, free-trial customers, refunds;
5. the reference period and when the number is taken;
6. the metric's owner.

These cards become the glossary of your data room. When an investor asks how you calculate a number, you send the card and the conversation moves straight to substance.

The weekly rhythm

A number looked at once a quarter is not managed. Set a short weekly slot, for example thirty minutes on Monday morning, in which:

• each owner updates their metric in the same table;
• the value is compared with the previous week and with the target;
• only changes above a threshold agreed in advance get discussed;
• a decision or action is noted, even when the decision is "do nothing and watch".

Keep the table in one place, with the history visible. A simple tool is fine: a shared spreadsheet with one tab for definitions and one for weekly values. The advantage of a simple system is that you actually use it.

The decision log

This is the least used tool and the one that makes the strongest impression in an investor conversation. Every time a metric leads you to change something, write one line with the date, the metric, what you observed, what you decided and what you expect to see.

After a few months the log tells the startup's story better than any slide: it shows you formed hypotheses, tested them and corrected course. It also shows the mistakes, and that is fine. A team that documents a failed experiment and explains what it learned comes across as more credible than one that only shows rising curves.

Common mistakes to avoid

• Vanity metrics up front: total sign-ups, page views, downloads. They can sit in the deck, but they do not replace value metrics.
• Definitions that change when the number looks bad. If you must change a definition, recalculate the history or flag the break.
• Different numbers in different documents. The deck, the financial model and the data room must all draw on the same table.
• Averages that hide reality. With few customers, one large contract can move everything. Show the distribution too.
• Targets without an owner. Every metric has one person accountable for the number and its quality.

What to do in the next four weeks

1. Map on one page the chain of events that leads a customer to pay and stay.
2. Pick three to five main metrics and write the definition card for each.
3. Rebuild the history of the last twelve weeks, even if partial, and flag the gaps.
4. Put the weekly review in the calendar and name an owner for each metric.
5. Open the decision log and fill it in at every review.
6. Check that the deck and the financial model use the same numbers as the table.
7. If you want an outside view on your metrics before you talk to investors, get in touch.

You will find templates and operating checklists in Resources, and you can compare notes with other founders in the community.

Subscribe to Dispatch, the PATAVIUM newsletter, for practical tools to build your startup every week.