Building Models Nobody Uses: The Analytics-to-Action Gap in PE
The CFO at a PE-backed services company had finally done it. Eighteen months of work, a clean financial model, monthly operating reports going to the right people on the right cadence. He’d built something real.
The CEO thanked him for the reports. Then made the same pricing decision he’d made the previous two years.
The problem wasn’t the model. The model was fine. The problem was that “revenue is down 8% versus plan” doesn’t tell anyone what to do about it. So nobody did anything.
Last week’s post covered a different failure mode: the fear of implementing systems in the first place — the horror stories that make portcos hesitant to upgrade what clearly isn’t working. That’s the paralysis before the decision. This is what happens after you make the decision, build the thing, and ship it to the people who are supposed to use it.
PE-backed companies spend real money on analytics. They hire CFOs, build FP&A teams, stand up dashboards, and implement reporting cadences. The systems exist. The data is (mostly) clean. The models are (mostly) right. And at the end of every monthly reporting cycle, the numbers go out, the board reviews them, and the operators continue doing what they were doing.
This is the most expensive analytics failure in PE. Not the missing data. Not the broken system. The model that works perfectly and changes nothing.
The Three Ways Analytics Fails After You’ve Built It
A CFO with over a decade of experience across PE-backed companies laid out the real challenge clearly: “Not building the models — that’s the easy part. Getting the CEO to actually act on what the numbers are telling us is the hard part.”
Building the model is the easy part.
The frustration behind that observation shows up three different ways inside portcos.
The CEO sees the financials as a report card, not a decision tool. The monthly reporting package lands in the inbox. Leadership notes that margin compressed again and moves on. Nobody asks: what specifically changed in our pricing or cost structure, and what are we changing next month? The model becomes a grade, not a directive. The board gets its data. The company doesn’t change.
End users don’t know what the reporting is asking them to do. As one experienced CFO put it: “A lot of the end users don’t even end up taking the correct actions based on the reporting.” A sales VP sees “pipeline conversion is down” in the monthly deck. Which accounts? Which reps? Which stage? The report shows a number — it doesn’t provide a path to the decision that changes the number. Data literacy is part of this. But it’s mostly a design problem. Reports built for board compliance aren’t built to hand an operator their next move.
Reporting itself becomes the full-time job. The same CFO described his actual first job this way: “Being a goalie — catching as many BS reporting requests as I can to protect my team from bloat.” Inquisitive board members generate a constant stream of data requests. KPIs, trend slices, peer benchmarks, scenario analyses. The analytics team that should be analyzing the business spends its time building decks about it. The signal gets buried in the volume of producing reports about the signal.
These aren’t three separate problems. They’re three symptoms of the same design failure.
Why PE Makes This Worse
Private equity doesn’t create the analytics-to-action gap — but it accelerates it.
The reporting requirements that come with PE ownership are real. Boards demand more frequent updates, more granular KPIs, more documentation of assumptions. This is legitimate. A fund with fifteen to twenty portfolio companies needs consistent, auditable data across all of them.
But here’s the structural paradox: the reporting cadence that satisfies the board runs in the opposite direction from the operational analytics that drives behavior.
Board reporting is designed to flow upward. Portco summarizes the month, packages it cleanly, presents it quarterly. The PE board reviews it and provides strategic guidance — usually financial, occasionally operational. What rarely flows back down is interpreted operational insight: what the numbers mean in terms of the specific decisions the management team should be making this month.
“PE often ignores the operational, cultural, and momentum impacts of personnel changes because those don’t fit into a nice cell on a spreadsheet.” That observation came from a CFO recounting a particularly turbulent period of board engagement. It crystallizes the tension: the board optimizes for what can be measured cleanly. Operations run on what can’t.
The result is a reporting architecture built entirely for compliance and almost not at all for decision-making. The portco produces monthly data because the board requires it. Nobody designed the same data to actually change what happens on Tuesday.
The Fix: Design Around Decisions, Not Data
Tech enablement does not equal analytics capability. You can implement a BI platform, build a clean reporting stack, and automate your monthly close — and still end up with a CFO spending most of their working week producing reports that nobody acts on. The tools are necessary. They’re not sufficient.
Analytics should be designed around decisions, not around data.
Existence doesn’t equal utility. Every metric in your reporting stack should have an owner, a threshold, and a prescribed action. If a KPI goes red and nobody does anything different, the KPI is decoration.
This isn’t abstract. Here’s what it means in practice.
Start with the decision, not the metric. Before building any report, ask one question: what decision does this metric support? If the answer is “we need to track it” — you don’t have an answer. Every metric should map to a specific person who is empowered to act on it, and a specific action they are expected to take when it crosses a threshold. “Revenue per employee falls below $X → the CFO reviews headcount against plan within two weeks” is a decision. “Revenue per employee: $Y” is a decoration.
Build three tiers, not one. Most PE-backed companies have exactly one reporting layer: board-level summary. What they’re missing is the operating layer that actually changes behavior. Consider three tiers:
| Tier | Purpose | Audience | When to act |
|---|---|---|---|
| Flash | Something needs attention now | CEO / COO | Within 24 hours |
| Operating | Are we on track this month? | Department heads | When variance exceeds threshold |
| Strategic | Are we building long-term value? | PE board | Informs next quarter’s priorities |
The flash tier is usually what’s missing. It’s not a monthly deck — it’s a weekly signal that something is off and someone needs to look at it today. The board reporting gets built first because it’s required. The operational early-warning system that actually drives behavior never gets built because nobody demanded it.
Embed analytics in the workflow, not the inbox. Reports that live in email attachments don’t change Tuesday’s standup. Analytics that show up in the tools people use every day — a Slack alert when a metric crosses threshold, a standing agenda item in the weekly ops meeting, a visible dashboard at the point where the decision gets made — those drive behavior. The question is never “did we produce the report?” The question is “what happens when the number goes red?”
Cut report volume, increase report quality. That metaphor reveals something real. If the CFO’s job is protecting the team from bad report requests, the team is fielding a lot of bad report requests. Fewer metrics, higher signal, mandatory commentary — “this is what changed, this is what it means, this is what we’re doing about it” — is more valuable than ten dashboards nobody opens. The 20-page monthly report that technically contains everything but communicates nothing is a volume problem. The solution isn’t more granular data — it’s more intentional decisions about which data earns a place in the operating system.
The Last Mile Is Where Value Gets Lost
The analytics-to-action gap doesn’t fix itself. It lives in the space between the analytics function, which builds the reports, and the operations team, which should be acting on them. Neither owns it.
There’s no villain here. The CFO is doing their job. The COO is doing their job. The gap between them is structurally orphaned — nobody’s accountability, everyone’s problem.
You can read about how cascading your KPIs from board level down to role level creates the structural connection between strategic targets and daily decisions — that’s the operating architecture. But before the cascade can work, the underlying question has to be answered: what decision does this metric support, and who is accountable for making it?
If your analytics function is spending more time building models than your operators are spending acting on them, the analytics function is inverted. The model isn’t the asset. The decision it’s supposed to drive is the asset.
The companies that close the analytics-to-action gap don’t do it by buying better tools or producing more reports. They do it by asking, for every metric in their reporting stack, the one question that makes it real: when this number moves, who does what?
If you can’t answer that cleanly for your top ten KPIs, you have a design problem.
Alex Escoriaza works with PE-backed companies on exactly this problem. If your analytics are rigorous and nothing is changing operationally, let’s talk.