Most founders don’t lack data. They lack a small, trusted set of numbers they can look at on a Monday morning and know, without having to double-check or ask someone to explain it, whether the business is in good shape. What they have instead is a sprawl of dashboards, spreadsheets, and reports that all say something slightly different — which means the real Monday-morning ritual, for a lot of founders, is picking up the phone and asking someone to interpret the numbers for them. That’s not a reporting system. That’s a person standing in for one.
Why more dashboards make this worse, not better
The instinct when reporting feels untrustworthy is usually to add more of it — another dashboard, another weekly export, another line item. This almost always backfires, because it adds more places where the same metric can be defined slightly differently, and more surface area for two numbers that are supposed to match to quietly not match. The fix for untrustworthy reporting is very rarely more reporting. It’s fewer numbers, defined once, calculated the same way every time, by a process nobody has to double-check.
The short list that actually matters
The specific numbers vary by business, but the categories are consistent across almost every founder-led company we’ve worked with. Cash — not a rough sense of the bank balance, but actual position and runway, updated on a defined cadence. Margin — not just revenue, which hides more problems than it reveals, but what’s actually left after the real cost of delivering the work, tracked closely enough to catch erosion before it’s a crisis. Pipeline — a trustworthy read on what’s coming, not an optimistic one; the gap between the two is where a lot of founders get surprised. And three or four leading indicators specific to the business — the metrics that tend to move before cash or margin do, and therefore give you warning instead of an autopsy.
That’s usually five to seven numbers, not fifty. The goal isn’t comprehensiveness. It’s a founder being able to glance at one place on a Monday morning and know, without a phone call, whether anything needs attention this week.
Why this is a design problem, not a tooling problem
Buying a business intelligence tool doesn’t solve this, because the tool isn’t what’s untrustworthy — the definitions underneath it are. If two people on the team calculate “margin” differently, a beautiful dashboard just displays two confidently wrong numbers faster. The actual work is upstream of any software: agreeing, once, on exactly how each number is defined and calculated; identifying where the underlying data actually lives and whether it’s clean enough to trust; and building the process that produces the same number the same way every single time, regardless of who’s running it that week.
Only once that’s settled does the dashboard or report become a genuinely useful layer on top — a way to see the numbers, not the thing that makes them trustworthy in the first place.
What changes once this exists
Founders who get this right describe a specific shift: Monday morning stops being a scramble to figure out where things stand, and starts being fifteen minutes of actually deciding what to do about it. That’s the entire point of instrumentation — not more data, but the confidence to act on the data you have without re-verifying it first.
Building the instrumentation is the second step in how we approach every operations engagement, right after designing the system it measures — see Operations for the full sequence. And because a reporting problem is so often standing in for something else — an undesigned process, an unclear owner — it’s worth reading how we diagnose that first, on How We Work.
Don’t trust your Monday-morning numbers?
Bring us the report you double-check the most. We’ll tell you what we’d redesign.