The improvement has to survive the operator

An interim CFO, a consultant, a strong operator, or the founder fixes pricing, inventory, or the close. Margins improve for a few quarters. Then that person leaves, and the numbers start to drift.

The fix was real. It was also stored in one person, and it left with them.

Some of the drift is in the business, as reviews lapse and old exceptions return. Some is only in the numbers, as definitions change and the trend line starts comparing 2 different things.

Either way, a buyer's diligence team will find it. Its advisors rebuild the run rate from the ledger and ask whether each recent gain will persist under new ownership. A step-up in margin with no system behind it tends to get discounted, or priced as a risk.

This is the sponsor's problem as much as management's. The sponsor paid for the improvement, and gets paid for it only if a buyer believes it will last.

Buyers pay full value for gains that keep running after the people who made them leave. The defenses are familiar: an operating cadence, documented systems, and owners below the person at the top. What decides whether they work is when they start.

Gains erode in 4 predictable ways, slowly enough to miss and steadily enough to matter.

Definitions drift. A new controller computes margin a little differently, or fill rate moves from orders to lines. The trend line now compares 2 different numbers, and nobody notices for a quarter. A buyer who finds the change will rebuild the history before trusting any of it.

The logic lives in one head. Pricing rules sit in a spreadsheet one person maintains, along with the reasons behind every exception. The next person inherits the file without the reasons.

The review stops. The weekly meeting that held the line was run by the person who left. The new general manager skips it in a busy month, then the next. Nobody cancels it, and it stops happening.

Exceptions creep back. A large customer asks for its old terms, then another. Discount approvals loosen one account at a time, and nobody watches the total.

The common thread is that the improvement was a set of decisions made by a person. Nothing in the company's systems made those decisions once that person was gone.

The decay is too slow to see in any single month. It shows up in diligence, as a run rate that does not hold.

The first defense is an operating cadence: a fixed set of meetings, each with its numbers, and an owner for each number.

Weekly, the operating review. Bookings, backlog, fill rate, and cash, each reported by the person who owns it, on the same definitions every week. Every action leaves the room with an owner and a date.

Monthly, the close and the reporting pack. One set of numbers, tied to the ledger and owned by the CFO, so the sponsor and the lenders see the same figures management runs the business on.

Quarterly, the board meeting and the plan review. Each initiative in the value creation plan reports its measure against its baseline, including any stop rule that came due. The plan is reset where it needs to be.

One rule holds the cadence together. Every number has exactly one owner, and the owner presents it. A number that belongs to everyone has nobody to defend it.

Each meeting runs from the same pack, in the same order every time: the numbers, the variances, the actions. The weekly review should fit inside an hour. If it runs long, the pack is doing too little of the work.

The numbers also have to arrive on their own. If someone builds the weekly pack by hand, the cadence depends on that person, and the original problem is back.

A cadence turns an improvement into a habit with a calendar. It also outlasts the people in the room.

A number that belongs to everyone has nobody to defend it.

The second defense is a handover pack, started in the first 100 days, when the definitions are being set anyway, and kept current from then on.

KPI definitions. For each number, the formula, the source system, the owner, and how often it refreshes. Fill rate, for example, might be lines shipped complete on the first shipment divided by lines ordered, from the warehouse system, weekly, owned by the head of operations.

System documentation. How data moves from each system to the board pack, who has access, and what to do when a feed breaks. Written in plain words for the next person, it is the part that gets skipped when a build runs late.

Owners. A named owner for every process and number, and a second person who can run it. The owner map also shows a buyer that the management team runs the business, and that is a large part of what a buyer is buying.

History. The cadence's own record: each month's measures, actions, and decisions.

The pack lives in the company's systems and belongs to the company. Its test is simple: a new CFO should be able to run the next board meeting from it.

A buyer asks the same 4 questions about every improvement. Are the KPIs defined the same way over time? Do the reported numbers tie to the ledger? Is the gain in the run rate? Will it keep running after the sale? Each question has an answer the company can build from day one.

Each test fails in a familiar way. A definition changed without a record, or a board pack needs a bridge to reach the ledger. A gain appears only in the last 2 quarters, or a process depends on one person.

Passing looks just as familiar. A year or more of each measure, on one definition, tied to the ledger, and presented by the person who owns it. All of it is cheap to build during the hold and expensive to fake at the end.

The months before a sale are the worst time to start. Management is running the process, the data room is open, and every new number invites a question about why it is new.

So set the definitions in the first 100 days and run the cadence from the first month. Every month after that adds history a buyer can test, and history is what a buyer trusts.

The same rule applies to outside help, firms like ours included. The work is finished when it runs without us.

Build every gain to run without its builder. That is the version a buyer pays for.