Somewhere in every peer-group room is the owner who fired a bad client a year ago, replaced half the revenue, and is quietly furious about it. The complaint sounds like math, but it almost never is. The decision was usually right. The measurement was the wrong number. The owner counted what was lost in revenue and is now trying to refill the bucket from the same line on the spreadsheet — when the line that actually mattered was always one step further down. Two prior conversations in this library are useful companions for the regret-phase owner: net profit is a lie, gross profit is king, and agreement gross profit is the number you are not paying enough attention to. Both make the same case: the revenue line is the wrong place to keep score, and the regret six months after a fire-the-bad-client decision is almost always the receipt for measuring the wrong thing.
What follows is not a defense of every fire-the-bad-client decision. Some of those decisions are bad. But the math underneath the right ones — the ones that look wrong in month six and right in month twelve — is the same math every time, and it is worth saying out loud. A fifteen-thousand-dollar-a-month client running at a thirty-five-percent agreement gross profit is fifty-two hundred dollars of actual money. At the BMK target of sixty-five percent AGP, that is one eight-thousand-dollar client — or, more realistically, two four-thousand-dollar clients. Same gross profit. Less servicing load. Less cultural drag on the team that was paying the tax in the first place. The regret in month six is the owner trying to replace the lost revenue. The right move was always to replace the lost gross profit — and to use the wound as the trigger for the whole-P&L recalibration the business probably needed regardless. By month twelve, the call that felt wrong is recognizably the call that should have been made eighteen months earlier. The hardest part of the right decision is that the right decision and the wrong decision feel identical for the first six months.
An MSP almost never fires a client for being merely unpleasant. The business is too transactional and the cash too useful for that to happen in isolation. What actually moves the decision is the moment two levers are pulled at the same time — the relationship is bad and the profitability is bad. Either lever on its own is endurable. The owner will absorb a great deal of after-hours hostility for a profitable seat, and a great deal of operational drag for a pleasant one. It is only when both conditions arrive together that the decision becomes obvious enough to act on. The owner who fires a client on the strength of personality alone usually regrets it inside ninety days; the owner who fires a client on profitability alone usually finds a way to keep them by raising price. The fire that holds is almost always the one that resolved both levers simultaneously. For the broader case on this point — the conversation that lives one step before the regret-phase one — the prior conversation on why MSPs must fire unprofitable clients is the companion reading.
The regret six months after a justified fire is not a decision problem. It is a measurement problem. The owner is measuring what was lost in revenue and trying to refill that bucket from the wrong line on the spreadsheet. Revenue is not what was lost. Gross profit is what was lost. A fifteen-thousand-dollar-a-month client running at a thirty-five-percent agreement gross profit was fifty-two hundred dollars of actual money. The owner sitting in month six is feeling the absence of fifteen thousand and is unable to land another fifteen-thousand-dollar client to replace it. That feeling is real. It is also entirely the wrong target. The actual target is fifty-two hundred dollars in gross profit — and at a healthy seventy- to sixty-five-percent AGP, that target is two four-thousand-dollar clients, which any MSP that landed a fifteen-thousand-dollar client in the first place can land inside the next two quarters. The math, told honestly, almost always converts the regret into relief. The owner stops trying to land another whale and starts building the steady current of the right-sized accounts the business actually wants on its books.
The BMK Vision benchmark for agreement gross profit is sixty-five percent. That single number reshapes every replacement decision an MSP is going to make in the next twelve months. A fired client running at thirty-five percent AGP loses the business fifty-two hundred dollars of monthly gross profit. To replace that gross profit at the healthy benchmark, the MSP needs eight thousand dollars of new monthly revenue, not fifteen thousand. Two new four-thousand-dollar clients land that gross profit cleanly, consume meaningfully less technician time than the fired account did, and create the rare condition almost no growing MSP enjoys — free capacity on the bench at the moment the new pipeline is being built. The owner who sees the math this way also tends to see the second-order benefit faster: the right technicians free up sooner, the weaker technicians are easier to right-size, and the seat the bad client was justifying becomes the seat the business should never have had in the first place. The math is uncomfortable on first reading, and irrefutable on the second. There is also a deeper accounting point underneath the replacement math — the costs most owners forget to account for before they declare a profit — which is worth its own visit before the recalibration round begins.
Top-line revenue is vanity, and the longer the owner spends running their business off it the more expensive the habit gets. The deeper move is the one that takes longer to teach: net profit also lies, in its own quieter way. A business that runs at a six-percent net profit can do so for years without ever revealing that the gross profit underneath it is collapsing — because the owner is absorbing the gap by not paying themselves, or by deferring infrastructure, or by quietly stretching the patience of an aging client list. The number that does not lie is agreement gross profit, measured per account. AGP per account tells the truth about pricing, scope creep, service-level drift, and the affordability of the next senior hire. It tells the truth about the bad client the owner is about to fire and the good client the owner is about to overspend on. It tells the truth about whether the seat that just opened up should be filled at all. The MSP that measures AGP per account is, almost by definition, the MSP that is not surprised by the question "why am I working this hard and not making any money." The MSP that measures only revenue and net profit is the one who arrives at year-end with a check to write and a confused look on their face.
The financial review is the only place the bad-client conversation usually gets a hearing, and it is also the place where the cost of keeping the bad client is most reliably understated. The tax that does not appear on the financials is the cultural drag the team is quietly paying every week to keep the seat warm. The tickets that come in with hostility. The Sunday-night emails. The negotiation-of-the-bill that consumes a billable hour without producing one. The technician who finishes the day already half-decided that they will be elsewhere by next quarter. None of this shows up in the AGP calculation, and yet all of it is the most expensive cost of carrying the account. The owner who fires the bad client almost always reports the same thing a quarter later: the team felt the relief before the spreadsheet did. The toxic client tax was real and uncountable, and the moment it was removed the company moved faster than the math said it should. This is the part of the fire-the-bad-client decision the regret-phase owner forgets the most, because by month six the team has already metabolized the relief and stopped naming it out loud.
The hardest line in the conversation is the one most owners flinch from. The MSP that runs at six or seven percent net profit is, in practice, an organization where every employee is being paid except the owner. The owner has normalized that condition long enough to forget it is not how the business is supposed to work. The first year of the business produced more cash than the previous job paid, the bar was set at that level, and the bar never moved. A decade later, the owner is still working off the original benchmark and absorbing the gap between the business the company actually is and the business it could be by personally subsidizing the wage line. The fire-the-bad-client decision is the rare moment that opens the door to fixing this — and it is also the moment most owners decline to open it, because the staffing recalibration the fire enables is hard, and the simpler move is to keep the same headcount and try to refill the revenue. The owner who uses the fire as the trigger for the whole-P&L recalibration usually arrives at the back half of the year with a fundamentally different business — fewer accounts, higher AGP per account, a lower fully-loaded payroll line, and, for the first time in years, an owner paycheck that resembles what a healthy MSP at this scale actually produces.
The discipline at the center of all of this is simpler than the math underneath it. The decision was probably right. The regret is probably real. The two feel identical for the first six months, and recognizably different by the twelfth. The owner who fired the bad client and is sitting in month six with half the revenue replaced is in the hardest stretch of the right decision, not the early stretch of the wrong one. The math, when told honestly, almost always confirms it. The replacement target was always gross profit, not revenue. The two four-thousand-dollar accounts can be landed inside the next two quarters by any MSP that landed a fifteen-thousand-dollar account in the first place. The technicians that the bad client was justifying are the right-size move the business was already overdue for. The owner paycheck that has been the silent shock absorber for years is the first line item the recalibration should restore. None of this makes the next six months feel any easier — but it does make them recognizably the right six months. As Josh closes the episode: it is not a bad call. It is a hard call. And the two feel identical for a while, which is the part nobody tells you when they hand you the original advice.
Because the owner is measuring the wrong number. The decision was made on profitability and fit; the regret is being measured against revenue, which was never the right metric. A fifteen-thousand-dollar-a-month client running at a thirty-five-percent agreement gross profit was producing fifty-two hundred dollars of actual money. The owner trying to land another fifteen-thousand-dollar account to "replace" the lost revenue is chasing the wrong target. At a healthy AGP, the same gross profit comes from two four-thousand-dollar clients, which any MSP that landed a fifteen-thousand-dollar client in the first place can land inside the next two quarters.
Sixty-five percent. That single number reshapes the replacement math after a fire-the-bad-client decision. A client running below that target — and many fired clients are well below, sometimes in the thirties — produces less gross profit than the headline revenue suggests. Anchoring every replacement conversation on the sixty-five-percent target reveals the actual size of the hole and tends to convert the regret into a recognizable, completable replacement project.
An MSP almost never fires a client for being merely unpleasant or merely unprofitable. The decision becomes obvious only when both conditions arrive together. Either lever on its own is endurable — owners absorb a great deal of after-hours hostility for a profitable seat, and a great deal of operational drag for a pleasant one. The fire that holds is almost always the one that resolved both levers at the same time. The fire that gets reversed inside ninety days is almost always the one that pulled only one of them.
Net profit lies more quietly than revenue, but it does lie. A business that runs at a six-percent net profit can do so for years without ever revealing that the gross profit underneath it is collapsing — because the owner is absorbing the gap by not paying themselves, by deferring infrastructure, or by stretching the patience of an aging client list. The number that does not lie is agreement gross profit, measured per account. AGP per account tells the truth about pricing, scope creep, service-level drift, and the affordability of the next senior hire.
It is the cultural drag a team quietly pays to keep a difficult account on the books. The hostility-laden tickets, the Sunday-night emails, the negotiation of the bill, the technician who finishes the week half-decided to leave. None of it appears in the AGP calculation, and almost all of it is the actual cost of carrying the account. The team usually feels the relief of a justified fire before the spreadsheet does, and stops naming it out loud inside one quarter — which is why the regret-phase owner has usually forgotten the relief by the time they second-guess the decision.
Often yes, and that is one of the gifts hiding inside a justified fire. The technicians the bad client was justifying are frequently the staffing right-size the business was already overdue for. The owner who uses the fire as the trigger to recalibrate the team — and the P&L — usually finishes the year with fewer accounts, higher AGP per account, a lower fully-loaded payroll line, and, for the first time in years, an owner paycheck that resembles what a healthy MSP at this scale should actually produce.
Gary Boyle is a Partner for Strategy & Business Development at Bering McKinley. With a background spanning network engineering, entrepreneurship, and strategic consulting, Gary brings real-world operator experience to helping MSP owners build stronger, more profitable businesses. He is a recurring co-host on The BMK Vision Podcast's Roundtable format, where he pressure-tests the leadership, financial, and operational decisions that determine whether an MSP grows, holds, or stalls.
Connect with Gary on LinkedIn →
Josh Peterson is the CEO of Bering McKinley and host of The BMK Vision Podcast. Since 2004, Josh has worked with hundreds of MSP owners to build operationally sound, profitable businesses through consulting, peer teams, and direct coaching.
Connect with Josh Peterson on LinkedIn →
If you are in the hardest stretch of the right decision — or quietly suspect the last fire was the wrong call — that is the place where the Vision Operating System does its most useful work. Anchor the replacement math on agreement gross profit, not revenue. Use the fire as the trigger for the whole-P&L recalibration the business probably needed regardless. Restore the owner paycheck first. The decision was probably right. The math, told honestly, almost always confirms it.