The KPIs that proved your business worked at Seed don't disappear. They just stop being true, and nobody edits the board deck to say so.
There is a slide that survives almost every board deck redesign: a metric the founder started tracking early, before there was a finance team, and it is still being presented today.
This metric was created before there was a BI tool, or sometimes even before there was a product anyone would pay for. It was the number that proved the thing was working, and everyone in the room recognizes it.
That's the problem.
Seed-stage investors weight a very different set of signals than Series A investors do. At Seed, "founder conviction comes first and demonstrated metrics second": what matters is retention cohorts, engagement depth, organic demand. By Series A, the same conversation changes: metrics stop being directional and start being comparable, with investors benchmarking you against peers, looking for evidence of a repeatable, capital-efficient motion instead of a promising trend line.
Nothing about that shift is controversial. I'd wager every operator who has raised more than one round has felt it. What's less discussed is what happens to the old metric once the bar has moved past it.
It doesn't get retired: it stays on the dashboard while still being updated, framed exactly the way it was when it mattered most, in Seed stage. The business has moved on to a different stage with different risks, but the metric hasn't been told that.
CRV's own framing of this is blunt: stage-mismatched KPIs "destroy startups because they create a false sense of momentum." Not because the number is wrong, but rather because it's still answering to the problems of last year.
The uncomfortable part of KPI drift is that the metric's value can stay flat, or even keep improving, while its meaning quietly degrades.
Take burn multiple. At Seed, a burn multiple above 3x is generally considered high, as spending is expected to be inefficient while you're still finding the model. By Series A, that same number is read completely differently: efficiency has become a primary filter alongside growth, not a secondary consideration. A burn multiple that would have been unremarkable eighteen months ago can now be the single line that stalls a raise.
Gross margin follows the same pattern: SaaS gross margin has a rough Series A benchmark of roughly 70%+, but that threshold means nothing at Seed, where the product economics aren't even settled yet, in most of the cases. Net revenue retention barely registers as a Seed-stage metric, as there isn't enough cohort history to calculate it meaningfully, thus becoming one of the first things a Series A or Series B investor will ask for.
What's easy to miss is that a company doesn't cross these thresholds on a clean date: it drifts across them gradually, while the reporting pack keeps presenting the same handful of metrics, framed the same way, because nobody scheduled a moment to ask whether they still make sense or not.
There's a more precise term for what's happening underneath this: metric drift, known as the gradual divergence of a metric's definition, or its relevance, as it gets copied, reused and re-contextualized across an organization without anyone formally revisiting it.
Kevin Bohan, writing on the topic, puts it plainly: "Few things undermine trust in data faster than two dashboards showing different answers to the same business question." Most of the time, that's discussed as a data engineering problem: one team excludes cancellations from revenue, another doesn't, and the two dashboards diverge. KPI drift is the strategic version of the same failure: the calculation didn't change; the business the calculation was supposed to describe did.
This is also exactly where drift connects back to something we covered in the previous article in this series: report lineage. A well-governed Financial SSOT doesn't just trace a number back to its source transaction. It should also be able to answer when this metric was last validated as still relevant, not only whether it was calculated correctly. Most reporting stacks are built to answer the first question, but almost none are built to answer the second.
Most KPI drift is passive: nobody chose for the metric to stop meaning something. But the same mechanism, pushed further, becomes something else entirely.
WeWork's "Community Adjusted EBITDA" is the reference case. The metric excluded not just interest, taxes, depreciation and amortization, but also marketing, general and administrative expenses, and the cost of opening new locations. The effect was dramatic: a period with roughly a billion dollars in reported losses became, under the adjusted metric, a $233 million "profit." A bond analyst's reaction summed up the market's response: he'd never seen the phrase before in his life.
The company was valued at $47 billion in 2019. It filed for Chapter 11 bankruptcy in November 2023, with shares down to $0.84 — roughly $44.5 million in market value. Not every case of KPI drift ends there, but WeWork is the clearest illustration of the same root failure at its most extreme: a metric's definition drifting so far from the underlying reality that it stopped functioning as information and started functioning as fiction.
Most scale-ups aren't at risk of inventing a metric like Community Adjusted EBITDA. The more common failure is quieter: it doesn't relate to redefining a number to flatter it, but simply never asking whether an old, well-behaved metric is still measuring what the business now needs to know.
We wrote in our first article that poor financial data rarely looks broken. Dashboards load. Reports go out on schedule. The numbers look polished. KPI drift is a specific, particularly stubborn version of that same pattern, since the metric isn't even wrong. It's calculating exactly what it always calculated; however, it's just no longer the thing anyone should be looking at anymore.
Gartner's own guidance on board reporting names the failure directly: a common problem is "overreliance on legacy metrics that once mattered but no longer reflect the strategic context" the company actually operates in. The recommendation isn't to rotate metrics constantly, seeing that a shifting scoreboard prevents boards from building trend judgement; rather, it is to periodically and deliberately revisit whether the fixed set still maps to the business's current risks.
That distinction matters: the goal isn't a board deck that changes every quarter. It's a board deck where someone can say, with confidence, when each metric on it was last checked for relevance and accuracy altogether.
If you're a finance leader wondering whether this is already happening in your own reporting, the exercise is simple, and it pairs directly with the five-number exercise from our previous article. For each metric currently on your board deck, ask:
In most scale-ups, this surfaces at least one metric that has been quietly promoted from "useful signal" to "ritual", tracked out of habit rather than relevance. That's not a failure of anyone's judgement, but it's what happens when a KPI set is built once, during the scramble to raise a round, and never formally revisited against the stage the company has grown into ever since.
Fixing KPI drift isn't about deleting metrics. It's about building the habit of asking whether the reporting layer has kept pace with the business, which is the same governance discipline a Financial SSOT is supposed to provide.
In the next article in this series, we'll look at what that discipline looks like in practice: how to design a reconciliation cadence that catches drift (in both definitions and relevance) before it reaches a board deck, rather than after someone in the room asks a question the metric was never built to answer.
If you suspect some of the metrics steering your board conversations are answering a question your business has already outgrown, that's worth a closer look before it becomes a harder conversation. No pitches, no proposals, just a conversation.
Discuss Your Context ↗