Guide
When the Largest Customer Gives Notice: Using the Notice Period Before the Revenue Goes
What a PE-backed company should do after its biggest customer gives notice: what is recoverable, which costs follow the revenue out, what the sponsor needs.
When the largest customer gives notice, the revenue has not left yet. What the company has is a defined period in which the account is still billing, the relationship is still open, and the facts are still gettable, and that period is the only asset the situation hands you. Most of the damage in a concentration event comes from spending that window on disbelief and recovery attempts rather than on the three decisions that actually determine where the company lands: whether the account can be saved, what the real margin hole is, and what the sponsor is told and when.
This is the acute event, not the structural problem. The question of how a company got to the point where one customer could do this, and how to dilute that position over time, is a different piece of work covered in fixing customer concentration. What follows starts the morning the notice lands.
Establish what is leaving, and when, before anything else
The first task is unglamorous and it is almost always skipped. Read the contract and write down what is actually committed.
Notice periods, wind-down obligations and transition-service clauses differ by agreement and often differ by site or business unit within the same customer. A customer that represents a quarter of revenue is rarely a single line. It is usually several contracts, some of them with different end dates, some auto-renewing, some already month to month. Work out which revenue is contractually protected through a date, which is in practice discretionary before that date, and which has already quietly stopped.
Then separate what the customer buys into lines, and record the margin on each. Companies routinely discover at this point that the account’s profitability is nothing like the average, in either direction. A big customer that negotiated hard may be carrying thin margin, in which case the revenue hole is far larger than the profit hole. A big customer on legacy pricing may be the opposite. The response to those two situations is not the same, and you cannot tell which one you are in from the revenue share alone.
Decide whether the account is recoverable, and put a date on deciding
Recovery attempts are worth making and they are worth time-boxing. The question is not whether the company wants the account back. It is whether anything within the company’s control caused the decision.
Three causes behave differently. A decision driven by price or by performance is sometimes reversible, because both are things the company can change and the customer can observe changing. A decision driven by a change at the customer’s end, a new owner, an insourcing program, a site closure, a procurement mandate, is usually not reversible by anything the company does, however well it responds. And a decision that has been made for a while and only communicated now is harder to move than one taken recently, because the customer has already done the internal work of justifying it.
Finding out which of these you are in means asking directly and being willing to hear the answer. The useful question is what would have had to be different, asked of someone senior enough to know. What comes back is either a specific, addressable gap or a general statement of direction. A general statement is a no.
Set the date by which the company stops spending commercial attention on the win-back and starts spending it on replacement. Naming that date in advance is what stops a recovery effort absorbing the whole notice period and leaving nothing done on the other two decisions.
Work out which costs follow the revenue out, and which do not
This is where the forecast usually goes wrong, and it goes wrong in a predictable direction.
Variable delivery cost leaves with the work. Materials, subcontracted capacity, usage-based licensing, anything consumed per unit delivered. Dedicated labor leaves only if the company chooses to let it go, and only at the speed that notice and local employment practice allow. Shared labor, supervision, facilities, systems, insurance and the management layer do not leave at all. They get reallocated across a smaller base, which means the remaining accounts each look worse than they did the month before, through no change in their own performance.
The practical consequence is that the margin hole is usually a larger share of profit than the account was of revenue, and the gap between those two numbers is the first thing a sponsor will want quantified. Produce it early and produce it honestly.
There is a trap on the other side of this. The instinct after a large loss is to cut capacity to match the smaller base, and some of that is correct. But capacity cut now is capacity unavailable to serve replacement revenue later, and rebuilding it costs more than holding it. The distinction worth drawing is between capacity that was specific to the departing customer and capacity that is general. The first can go. The second is the thing that will deliver the replacement, and cutting it converts a revenue problem into a growth problem that outlasts it.
Tell the sponsor early, with the arithmetic attached
The temptation is to wait until there is a plan. That instinct is wrong, and it is wrong for a reason worth stating plainly: sponsors find out. They find out from a monthly pack, a lender covenant conversation, or the customer’s own announcement, and finding out late converts a commercial event into a question about management’s candor. S&P Global’s 2026 work found 71 percent of GPs and 53 percent of LPs naming operational value creation a priority, which is another way of saying the people above this company are paying close attention to exactly this kind of event.
Tell them quickly, and tell them with numbers rather than reassurance. A first update that holds up contains what was lost and over what period, the revenue and the margin effect stated separately, what is contractually protected and until when, the recovery assessment and the date it will be concluded, and what has not yet been worked out. That last item matters. An update that implies everything is understood on day three will be contradicted by the update on day thirty, and the contradiction costs more than the admission would have.
What the update should not contain is a replacement plan invented to make the first conversation easier. A number put on the table to soften the news becomes the number the company is held to.
Be honest about replacement arithmetic
Replacing a large account is not usually a matter of selling harder for a quarter. Work out what it would actually take: how much new revenue, at what margin, closing by when, and what that implies for pipeline at the company’s own historical win rate and sales cycle. Where the shortfall against the plan has to be closed rather than absorbed, the reforecast belongs in the operating rhythm rather than in a one-off exercise, and the method for that sits in diagnosing a repeated forecast miss.
Two things change in the company’s favor, and both are worth counting. Capacity that was committed to the departing customer is now available, which can shorten delivery times or allow work that was previously declined. And the commercial attention that a dominant account absorbs is released, which is frequently more valuable than it looks, because large accounts consume senior time out of proportion to their margin. Claymore’s own case numbers include a 36 percent revenue lift over 24 months on flat marketing spend, which is what it looks like when attention and method change rather than budget.
What does not help is assuming that the replacement will come from the same type of customer at the same size. The concentration that made this event possible is usually the result of pursuing large accounts because they were efficient to win. Rebuilding the same shape rebuilds the same exposure.
What this does to the exit story, and what to do about it
A concentration event inside the hold period will be read by a future buyer, so it is worth managing as an evidence problem and not only as a trading one. Buyers test revenue durability directly, and the way they do it is set out in quality of revenue.
The version of this event that prices well is one where the company can show what happened, when it was known, what was done, and what the base looked like afterwards. The version that prices badly is a gap in the record followed by a recovered revenue line with no explanation. Keep the documentation as you go: the notice, the assessment, the decisions and their dates. It costs nothing at the time and it answers the diligence question years later.
Should a company tell its sponsor before the customer’s notice is final?
Yes, when notice has been given, even if the customer has left room to reconsider. Waiting for certainty means the sponsor learns about a material revenue event from somewhere other than management, which changes the conversation from a commercial problem into a trust problem. Say what is known, mark clearly what is not yet settled, and give a date for the fuller assessment.
How much of the lost margin can be recovered by cutting cost?
Less than most first estimates assume. Variable delivery cost goes with the work, dedicated labor goes only at the speed notice allows and only if the company chooses, and shared overhead does not go at all. That is why the profit effect is usually a bigger share of profit than the account was of revenue. Cutting general capacity to close the gap tends to make the replacement harder and is rarely worth it.
Is it worth trying to win back a customer that has given notice?
It is worth a bounded attempt when the cause is price or performance, because those are things the company can change and the customer can see change. It is rarely worth a long attempt when the decision came from a change at the customer’s end, such as new ownership, an insourcing program or a procurement mandate. Decide which case you are in by asking what would have had to be different, then set the date when the effort stops either way.
How does losing a large customer affect valuation at exit?
Less through the revenue itself than through what it says about durability. A buyer reading a concentration event wants to know whether the base that remains is stable and whether management saw it coming and responded. A well-documented loss with a clear response and a rebuilt base can price acceptably. An undocumented gap in the record invites the assumption that it could happen again without warning.
What should the board see in the first update after notice?
What was lost and over what period, revenue and margin effect stated separately, what is contractually protected and until when, the recovery assessment with a date for concluding it, the cost that will and will not come out, and an explicit list of what has not been worked out yet. No replacement number until there is evidence behind it.
Have a revenue problem the board is asking about? Start a conversation.