Ask a CFO whether they trust the number in front of them. Most won't say yes outright. Systems have improved. Dashboards have multiplied. Automation has spread through consolidation and reporting. Yet across 29 anonymised, one-to-one interviews with CFOs, Finance Directors and Heads of FP&A in UK mid-market businesses, one behaviour came up again and again: senior people personally reconciling, cross-checking, or quietly rebuilding numbers before they would stand behind them in the boardroom.

This isn't a data volume problem. It's a trust problem. And the same four patterns kept recurring, however much the business had already spent on systems.

Four Patterns Behind the Trust Gap

1. Fragmentation nobody consolidated. Growth by acquisition, multiple ERPs, and legal entities that were never brought onto one data model. Nobody trusts the group number because everyone knows it was stitched together by hand.

2. No single owner of data quality. Data belongs to whoever happens to be able to fix it, usually one person. When that person is on leave or moves on, reporting stalls or breaks quietly.

3. Manual reconciliation is eating the majority of FP&A's time. Estimates across the interviews ranged from 50 to 90 per cent of time spent on pulling, checking, or rebuilding numbers in Excel rather than analysing them.

4. Numbers shaped by what's palatable, not by what the data supports. Forecasts get adjusted in the room to suit a CEO, board, or investor, independent of the underlying drivers.

In Their Own Words

“I don't trust what I'm seeing here is the first default reaction to data. Then you build some trust around it. But there are constant caveats.” Head of Group FP&A, Leisure and Entertainment sector
 

“It's like trying to extract blood from a stone,” Finance Director, mid-market retail and manufacturing business, on getting a usable report out of the core finance system
 

“If it gives you an answer you don't like, you assume the answer is wrong, because you can. You can always blame the data,” Group Finance Director, Aviation and Travel Services sector

The trust gap is starting to show up in a new place, too. Several of the leaders interviewed described AI-generated reports giving different answers depending on who asked the question, or said they were deliberately holding back on AI adoption until the underlying data could be trusted. 

As one Financial Controller in an e-commerce business put it: “Someone will say my report said this, and someone else will say mine said that, and there can be a difference of 6,000 in a month.” 

Automation doesn't fix a trust problem. It reproduces it faster.

How It Shows Up in the Numbers

These patterns aren't abstract. Across the interview set, they showed up as specific, costly numbers.

Missed revenue. At a UK-based professional services firm backed by a private equity investor, the revenue hadn't disappeared. It had simply never been visible in a system anyone fully trusted, so nobody had gone looking for it. The unbilled revenue ran into a few hundred thousand pounds.

Cost surprises. At a medium-sized manufacturing business owned by a PLC, open purchase orders were an ongoing, unresolved issue for years. Then the finance team was landed with over £200,000 in surprise costs nobody was expecting but had to absorb because POs signed off in the past were never closed out or updated once goods, services, or invoices were received.

Systems are rarely the problem. It's the process around them, or the absence of one, that creates the real exposure. The process for surfacing issues like these was either never implemented or never routinely reviewed. Elsewhere in the research, forecast accuracy was described as “60 to 70 per cent on a good day”; one team's forecast error rate more than tripled as headcount grew at half the pace of the business itself. The trust gap doesn't just cost time. It hides money, and it compounds.

What Finance Leaders Do About It

1. Reconciliation moves upstream. Check the inputs at the point of entry. Confidence gets built into the pipeline, not bolted on at the end by the most senior person in the room.

2. One owner per data source, not one owner per report. Trust breaks fastest when nobody owns the underlying source. Name a single accountable owner for each one.

3. The system becomes the default, not the exception. Manual checking becomes rare: reserved for genuinely unusual figures, not applied to every number out of habit.

4. Trust the system record over the anecdote. One Head of Finance in the research described a deliberate policy: if the system shows one interview happened, that is the number used, even when colleagues insist there were five. It puts the discipline back upstream, instead of firefighting downstream.

Why It Bites Harder in PE-Backed, Mid-Market Businesses

The cost of a trust gap compounds faster in a PE-backed business than in a stable, mature one. Multi-entity structures, recent acquisitions, and system migrations, all common in the mid-market PE portfolio, multiply the number of manual handoffs that data survives before it reaches a board pack. It's the same fragmentation pattern seen across the research, just moving faster. Every extra handoff is another point where trust, once lost, must be earned back by someone senior who personally checks the work.

The Test

If you lead FP&A, the test is simple. How many hours do you personally spend each month reconciling numbers that a system is supposed to have already reconciled? More than a handful? The fix usually isn't a bigger system. It's knowing exactly which handoff in the pipeline nobody owns, and giving it an owner before the next board pack is due.

Where's your unowned handoff? That's usually the first conversation worth having.

Note: This article draws on 29 anonymised interviews conducted as part of ABCL's ongoing FP&A research programme.