The bravest thing a CEO can do is make the business deliberately smaller before trying to make it bigger. Not smaller in people — smaller in customers, products, and activity. I mean walking away from real revenue on purpose, in public, and telling your board you did it on purpose. Every instinct, every incentive, and every dinner-party question — how big is your company? — pushes the other way. But after thirty years of running industrial businesses and more than $3 billion in shareholder value created, I can tell you the pattern has never once failed to hold: the companies that shrank first grew best, and the companies that refused to shrink grew busier, thinner, and eventually backward. The arithmetic explains why, and arithmetic is where we will start.
I have watched this movie from every seat — CEO, chairman, acquirer, and the guy brought in afterward to clean up. The refusal to shrink is the most expensive decision most CEOs never realize they are making, because it does not look like a decision. It looks like keeping customers. Nobody gets fired for keeping customers. They should occasionally be thanked for the opposite.
The Arithmetic Nobody Wants to Do
Run an 80/20 analysis on any mature business and the same picture appears with tedious reliability. The top quartile of customers typically produces 105 to 150 percent of total profit. Read that again: more than all of it. The bottom half produces little or nothing and consumes a third or more of the company’s operating capacity — order lines, engineering exceptions, freight surcharges nobody bills, credit memos, sales calls, management attention. The tail is not merely unprofitable. It is funded by the core, which means your best customers are subsidizing your worst ones, and your people spend their days servicing the subsidy.
Once you accept that math, shrinking stops being radical and starts being obvious. If a quarter of the business generates all the profit, then removing the loss-making tail must, mechanically, raise total profit — while freeing the capacity currently consumed by the tail to serve the core better. Revenue falls. Profit rises. Margin expands. Complexity collapses. There is no strategic cleverness in this; it is subtraction. The cleverness is entirely in the courage, which is why this post is about bravery rather than spreadsheets.
If you doubt your company fits the pattern, run one number tonight: rank your customers by profit at full cost-to-serve — not gross margin, full cost — and find where cumulative profit peaks. At a $350 million equipment business I chaired, the peak arrived at customer number 214 of roughly 3,100. Everything after 214 was, in aggregate, paying us nothing or charging us for the privilege of serving them. The management team had been proudly reporting customer count growth for a decade. They had been growing the part of the curve that slopes down. Nobody had ever drawn the curve.
What Deliberate Shrinking Actually Means
Let me be precise, because the word shrink frightens people into hearing layoffs. Deliberate shrinking targets three tails, none of which is your workforce. The customer tail: accounts that lose money at full cost-to-serve, repriced honestly and allowed to leave if they refuse the honest price. The SKU tail: the forty percent of the catalog that produces two percent of margin while consuming inventory, changeovers, and quality escapes. The activity tail: the exceptions, special processes, custom paperwork, and heroic workarounds that exist only because the first two tails demanded them. You cut the tails. The people who were servicing the tails are redeployed to the core, which has been starved of them for years.
This is the part CEOs consistently get backward. They protect the customer and SKU tails — the actual disease — and cut people, the capacity. That shrinks the muscle and keeps the tumor. An 80/20 shrink does the opposite: it removes work, not workers. In every shrink I have run, headcount reductions were modest to zero, and within a year the complaint from operations flipped from we are drowning to we finally have the capacity to do this right. Layoffs are what companies do instead of focus. Focus is harder and pays better.
The Distributor That Got Smaller and Richer
Here is what it looks like in practice, numbers lightly disguised. An industrial distributor, roughly $150 million in revenue, 11.7 percent EBITDA margins, growth flat for four years despite a salesforce running flat out. The 80/20 quartile analysis showed the classic picture: the top customer quartile produced about 130 percent of profit; the bottom half of the customer list — thousands of accounts — produced a loss and consumed just under 40 percent of order lines, freight events, and branch labor. The catalog told the same story in SKU form. Management had seen versions of this data before. They had simply never been allowed to act on it, because acting on it meant revenue would go down.
We acted on it. Over eighteen months the company repriced the tail to full cost-to-serve, minimum-order-sized the smallest accounts, moved marginal ones to a distributor-served model, and retired about a third of the catalog. Customer count fell 25 percent. Revenue dipped about 8 percent in year one — every dollar of it unprofitable revenue. And EBITDA rose 67 percent, with margins moving from 11.7 to 16.8 percent, because the freed capacity was redeployed onto the core: faster quotes, better fill rates, engineering time for the accounts that paid for it. Year two, revenue passed its old peak — built entirely on profitable business. Smaller first, then bigger. In that order, and only in that order.
Fear One: Revenue Optics
Three fears stop CEOs from doing this, and they deserve honest answers rather than pep talks. The first is revenue optics: the top line will visibly decline, and the top line is the number on the trophy. My answer is to change the trophy before you start. Announce the metrics that will define the program — EBITDA dollars, EBITDA margin, material margin per employee dollar, revenue per SKU — and report the revenue decline as a planned outcome against a published plan, the way you would report capex. A revenue dip that was forecast to the quarter reads as control. The same dip arriving unannounced reads as decay. Optics are managed with forecasts, not with courage speeches.
It also helps to name the number out loud. At the distributor, we told the board on day one: revenue will fall roughly 8 percent next year, here is precisely which revenue, and here is why we are glad to see it go. When it fell 8 percent, credibility went up, not down. The most disarming sentence in business is: this is happening exactly as we said it would.
Fear Two: The Sales Team Revolt
The second fear is the salesforce, and it is legitimate — you are proposing to take accounts, and often commissions, away from the people who fought to win them. Handle it with pay before persuasion. Rebuild the compensation plan so that sellers earn on margin and core-account growth rather than raw revenue, and grandfather the transition so nobody’s family budget takes the hit for your strategy. Then give the team the cost-to-serve data, account by account. Salespeople are not sentimental once they see which accounts have been quietly robbing their best customers to feed their worst. They are competitive people. Point the competition at the right scoreboard.
Expect to lose a few sellers anyway — usually the ones whose books were mostly tail, which is information in itself. At the distributor we lost three of forty. The rest discovered something nobody predicted in the risk register: selling into the core with real capacity behind you is a better job. Quotes went out faster, orders shipped complete, and the sellers stopped spending Friday afternoons apologizing. Attrition in the following year was the lowest on record. The revolt, it turns out, was mostly in the forecast.
Fear Three: The Board
The third fear is the board or the sponsor — the worry that shrinking reads as retreat, or worse, as a CEO who cannot grow. This one is best solved before it exists: never bring a board the word smaller without bringing the arithmetic that makes smaller mean richer. Present the quartile analysis, the cost-to-serve math, and the bridged EBITDA walk from today’s margin to the target. Anchor it in the levers boards already believe in — price, mix, and cost are three of the five levers on any EBITDA bridge, and a shrink-first program pulls all three at once. Framed as arithmetic, shrinking is not a retreat from the growth plan. It is the growth plan, sequenced correctly.
With sponsors, add the exit math. Multiple expansion is dead; nobody underwrites buying at eight and selling at eleven anymore. Exit value now gets built inside the P&L, and a business that walks into a sale process with expanding margins, a clean customer book, and capacity to grow commands a premium precisely because the buyer’s diligence team can see the engine. I have sold shrunk-first businesses and I have sold never-shrunk ones. The former negotiate from strength. The latter negotiate from adjustments.
Why Growth on an Unshrunk Base Collapses
Here is the part that makes sequencing non-negotiable. Growth is a multiplier, and a multiplier does not care what it multiplies. Pour new revenue onto an unshrunk base and every pathology scales with it: the tail customers get siblings, the exception processes get new exceptions, the overloaded branches get more load. The new business inherits the old cost-to-serve, so it arrives at tail economics even when it was won from good customers. Meanwhile the growth spend itself — salespeople, marketing, capacity — lands as employee cost on a ratio that was already below the line. Revenue climbs. Profit does not. Eventually the gap gets noticed, and it gets noticed at the worst possible time, which is whenever someone is trying to sell.
I sat on a board that ran this experiment against my vote, and I described the outcome in detail in my post on earning the right to grow: revenue up nearly 30 percent over three years, EBITDA down 12, CEO gone at month twenty-six. The growth was real. The base was rotten. The multiplier did what multipliers do. Growth built on an unshrunk base is not growth — it is deferred shrinkage, scheduled for the moment you can least afford it, plus interest.
How to Sell Subtraction to a Board
A practical word on the meeting itself, because I have pitched this program to boards a dozen times and been the board hearing it a dozen more. Bring one page. On it: the profit concentration by customer quartile, the cost-to-serve of the bottom half, the planned revenue reduction with a quarter-by-quarter forecast, the EBITDA bridge it funds, and the redeployment plan for the freed capacity. Then make the ask specific: approve the program, approve the metrics that will replace revenue on the dashboard, and approve a two-year horizon so the program cannot be repealed at the first soft quarter. Subtraction dies when it is approved enthusiastically and measured by the old scoreboard.
Anticipate the two questions every board asks. What if a tail customer we fire turns out to matter? Answer: repricing, not firing, is the mechanism — anyone who matters will pay the honest price, and the ones who leave have told you what you were worth to them. What if competitors pick up the revenue we shed? Answer: sincerely hope they do. Loading a competitor with your loss-making accounts is the cheapest attack you will ever launch. I have watched a rival hire trucks to haul away business that was costing us money. I would have paid the freight.
One more piece of board mechanics: put a sunset date on the pain. Commit that by a named quarter, the program’s metrics — margin, EBITDA dollars, core fill rates — will have crossed specific thresholds, and invite the board to judge the program dead if they have not. Deadlines convert skeptics into observers. Open-ended discomfort converts them into opponents, usually around the second soft revenue print, and a shrink program repealed at the midpoint is worse than one never started: you paid the revenue cost and forfeited the margin prize. I would rather lose the vote in month one than win it conditionally and lose the program in month nine.
After the Shrink: Where the Freed Capacity Goes
Shrinking is half a strategy. The other half is the deliberate redeployment of everything the shrink releases, and this is where the growth actually comes from. The freed capacity goes to four places, in order:
- Service the core like it is paying for it — because it is. Fill rates, lead times, quote speed, and engineering attention for the top quartile improve within one quarter, and share of wallet follows within four. Your best customers have been rationing you for years without telling you.
- Price with confidence. A company no longer terrified of losing bad revenue negotiates differently on good revenue. The pricing courage the shrink teaches is worth points of margin on the business you keep.
- Innovate for the quartile, not the catalog. Product development stops spreading itself across every legacy segment and starts compounding where 130 percent of the profit lives.
- Then, and only then, add. New regions, adjacencies, and M&A now land on a clean base with capacity to absorb them — growth that compounds instead of congesting.
At the distributor, the sequence was visible in the numbers: service metrics moved in the first two quarters, share of wallet in the core moved in quarters three through six, pricing gains landed across year one, and the expansion investments began in year two — funded entirely by the margin the shrink had created. No new capital. The growth was paid for by the subtraction that preceded it. That is the whole model, and it fits on an index card.
The Courage Is in the Sequencing
None of this is intellectually difficult. The quartile analysis takes a competent analyst two weeks. The playbook — reprice, rationalize, redeploy, then grow — fits on a page. What makes shrink-first rare is that it demands a CEO spend real political capital making the company look worse by the old scoreboard in order to make it unarguably better by the real one, and hold that position through two or three uncomfortable quarters while the arithmetic does its work. That is a test of nerve, not intelligence. It is also, in my experience, the single highest-return decision available to the CEO of an unfocused business — and almost the only one your competitors will reliably refuse to copy.
If you want to pressure-test the playbook on your own numbers, The 80/20 Institute runs free workshops where we walk operators through the quartile analysis, the cost-to-serve math, and the shrink-first sequence step by step, using the same tools I use in my own companies. Bring your customer file and an open mind. The math will do the arguing. It always does — the only question is whether it argues for you now, on your schedule, or against you later, on someone else’s.
Frequently Asked Questions
What Does It Mean to Shrink the Business First?
Shrinking first means deliberately removing the unprofitable tail of the business — loss-making customers, low-margin SKUs, and the exception-driven activity they generate — before investing in growth. It targets work, not workers: capacity freed from servicing the tail is redeployed to the profitable core, which typically raises profit and margin even as revenue temporarily declines.
Why Does Cutting Customers Increase Profit?
In most mature businesses, the top quartile of customers produces 105 to 150 percent of total profit, meaning the bottom half produces losses that the core subsidizes. Removing or repricing the loss-making tail mechanically raises total profit while freeing the operating capacity — order handling, engineering, freight, management attention — that the tail consumed.
Does Shrinking the Business Mean Layoffs?
No. Deliberate shrinking cuts the customer tail, the SKU tail, and the activity tail — not the workforce. People who were servicing unprofitable complexity are redeployed to the core, which has usually been starved of capacity for years. In practice, well-run shrink programs involve modest to zero headcount reduction and often improve retention.
How Do I Convince a Board to Approve a Revenue Decline?
Bring arithmetic, not adjectives: profit concentration by customer quartile, cost-to-serve of the bottom half, a quarter-by-quarter forecast of the planned revenue reduction, the EBITDA bridge it funds, and the redeployment plan. Ask the board to approve new dashboard metrics and a two-year horizon so the program survives its first soft quarter.
What Happens If a Company Grows Without Shrinking First?
Growth is a multiplier, and it multiplies whatever base it lands on. New revenue added to an unshrunk base inherits the existing cost-to-serve and complexity, so revenue climbs while profit stalls or falls. The typical pattern is several years of top-line growth with declining EBITDA, followed by a forced fix-first program run at crisis prices.
