Why the 100-Day Plan Almost Never Includes the Process That Is Actually Broken

The variance nobody can explain

Month five. The board pack arrives two days late, which is new. Gross margin is down a point and a half against plan, and the explanation on the call is that it is "a timing thing in the cost allocation." The CFO sounds confident. The operating partner writes it down and moves on, because the 100-day plan was delivered on schedule: pricing review done, reporting cadence installed, two senior hires made, a procurement cost-out that already shows up in the numbers.

Month six, the timing thing has not reversed. Month seven, someone finally opens the spreadsheet that does the allocation and discovers it is maintained by one analyst, uses a rate table last touched two years ago, and was quietly patched by hand every month by the controller who left in month three.

Nothing in the 100-day plan was wrong. It was just aimed at a different company, the one visible from the data room.

Who writes the plan, and what they can see

The 100-day plan literature is written by people who arrive with the sponsor, and it shows. Read a dozen versions and the workstreams barely vary: pricing, reporting cadence, the organisation chart, quick procurement wins, a leadership assessment, maybe a working capital programme. These are good workstreams. They are also, without exception, things a deal team can see from outside the building.

That is not an accident. One post-close diagnostic piece puts it plainly: the plan is "largely written before close, by a team that had access to data rooms and management presentations but had never spent significant time inside the business," so it "reflects the deal team's hypothesis about the company, not a tested view of how the company actually runs" (VCI Institute).

The pressure to get the plan right has also gone up. Bain's current framing, summarised as "12 is the new 5," is that deals now need something like 10 to 12 percent annual EBITDA growth to hit target returns, against roughly 5 percent in the cheap-money years (Stanton Chase, citing Bain). Multiple expansion used to cover for a plan that missed. Now the operating plan has to carry the return on its own, which means a process the plan never looked at can sink it.

The broken process is invisible from the data room by construction

This is the pattern we see most often, and it is structural rather than a diligence failure.

A data room shows outputs. Financial statements, KPI packs, customer lists, org charts, system inventories. It shows what the business produces. It does not show how. And the broken process almost always lives in the how, in one of three places:

• A spreadsheet someone maintains. Revenue allocation, commission calculation, landed cost, a rebate accrual. The output is a clean number in the management accounts. The input is a workbook with hardcoded overrides and one owner.
• A reconciliation between two systems that nobody documented. The CRM and the ERP disagree about customer terms, so someone lines them up every month before invoicing. It works, so no one mentions it.
• An approval that routes through a person, not a role. Credit holds, pricing exceptions, and supplier changes all go to the same long-tenured manager, because that is how it has always worked. The system has no rule. The rule is Janet.

None of these appear in the CIM, and not because the seller is hiding them. The seller does not experience them as problems. From inside the business, the spreadsheet is just how margin gets calculated. It has worked for six years. Why would it be on a list of risks?

Growth Shuttle's review of the data problems that survive a quality of earnings report lands on the same ground from the finance side: revenue allocation "rules documented nowhere," a timing assumption "that three people interpret differently," and operations that depend on individuals rather than anything institutional. Their conclusion on the spreadsheet version is the one that matters for a sponsor: the process "cannot scale, cannot be audited, and cannot be handed to a new finance team during integration" (Growth Shuttle).

We made the adjacent argument about diligence in what operational diligence finds that a quality of earnings report cannot. The 100-day plan inherits that blind spot directly. If diligence never asked how the number was produced, the plan built on diligence cannot aim at it.

Why the plan makes it worse, not just misses it

Missing the broken process would be bad enough. The standard plan tends to stress it.

Look at what the usual workstreams do to the back office. A new reporting cadence asks for a monthly board pack faster and in more detail, which means the fragile allocation spreadsheet now runs on a tighter deadline. A pricing review changes price lists, which means the undocumented reconciliation between the CRM and the ERP now has more exceptions. A leadership refresh replaces the person who knew where the bodies were buried. A procurement programme changes suppliers, which puts new records through the approval that only one person understands.

Each workstream is sensible in isolation. Together they put load on exactly the parts of the business the plan cannot see. The process was coping at the old pace. The plan raised the pace.

This is why the failure shows up in month five rather than day ten. The first few months run on the old muscle memory. The variance arrives once the new cadence has run long enough for small errors to compound, and once the person who used to absorb them has moved on or burned out.

What operators know that deal teams do not

The people who can see the broken process are almost never in the room when the plan is written. The controller knows which reconciliation takes three days. The customer service lead knows which orders get re-keyed. The warehouse manager knows which report nobody trusts. None of them are asked, because the 100-day plan is a sponsor document and the first weeks are spent on the leadership team.

That produces two readings of the same company. The deal team reads the business as ready for acceleration because the outputs look clean. The operators read it as barely holding together because they are the ones holding it. Both are accurate descriptions of different layers. The plan gets written from the first layer and executed in the second.

The fix is not to stuff more workstreams into the plan. It is to put a different question into the first thirty days, asked of different people: which of the numbers in the board pack depend on something one person does by hand? That question is cheap to ask and awkward to answer, which is why the answer is worth so much.

The ERP trap

One version of this deserves its own warning. A plan that finds messy systems often reaches for an ERP consolidation as the cure. It sounds decisive, it fits on a slide, and it pushes the problem eighteen months out.

The trouble is that an ERP project migrates the broken process rather than fixing it. The undocumented spreadsheet becomes an undocumented custom report. The person-based approval becomes a workflow configured around one name. We covered the sequencing question in whether a portfolio company should automate before or after the ERP consolidation; the short version is that you cannot configure a system around a process you have not looked at.

Where automation fits, and where it does not

This is where a vendor pitch usually enters, and where we would slow down.

Automating the broken process as found is the most common mistake we see after the first one. A bot that runs the allocation spreadsheet faster still runs the stale rate table. An integration that replaces the manual reconciliation still has to decide which system is right about customer terms, and nobody has decided. Speed applied to an unexamined process mostly produces wrong answers sooner.

Sometimes the right answer is automation. Often it is deleting a step, assigning an owner, or writing down a rule that currently lives in someone's head. The order matters more than the tool. The automation ROI arithmetic only works once you know which of those you are looking at.

A spreadsheet held together by one analyst is a lot like a load-bearing wall nobody put on the drawings. You find out it is structural when you try to move it.

What this means for the next plan

The pattern is consistent enough to plan for. Every mid-market business has two or three processes that produce clean-looking outputs through fragile, person-dependent work. The data room will not show them. The seller will not list them. The standard 100-day workstreams will put load on them. And the cost will show up around month five as a variance that sounds like timing and is not.

Knowing that is the easy part. Finding which processes they are in a specific company, how much of the plan depends on them, and which to fix, delete, or leave alone is work that happens inside the business, with the people who run it, not in a data room.

If you have a 100-day plan on the table, or a month-five variance nobody can explain, talk to us. We will spend the time inside the operation that the deal timeline never allowed, and tell you which process the plan should have been about.