Every profitability project starts with a conversation in which someone explains why now is not the time. We hear those conversations directly, and we hear them second-hand from the partner firms that sell and implement our method and our software: 41 of them in 19 countries as of April 2026, from one-person consultancies to Big Four practices. That month we sat down to write the objection-handling material those partners had been asking for, and the exercise produced something more useful than a script. It produced a short list. Five objections cover almost every first meeting. Here they are, with what we actually answer.
What are the objections to a profitability project, and how do you answer them?
Five come up in nearly every first conversation: it is not a priority right now; we tried activity-based costing and it was too complex; the model lives in a spreadsheet nobody dares touch; we already know which customers make money; and we allocate overhead by revenue and it works. Each is reasonable and each is answerable with a number rather than a promise, because a first-pass cost-to-serve model on twelve months of transactions takes weeks, not months, and it usually finds the answer to the objection inside the company’s own data.
1. “Profitability modelling is not a priority right now”
This is the most common objection and the most honest, and it usually means one of three things. We are too busy fighting fires to prevent them. We assume we already know which customers are profitable. Or we tried this before and it hurt. All three deserve a straight answer rather than a counter-pitch.
What we answer is that margin erosion does not wait for the agenda. A customer whose orders cost more to serve than they earn keeps ordering, and the discount that made them unprofitable is renewed every month by default, not by decision. The cost of not looking is not zero; it is the same number every quarter, compounding quietly. The other half of the answer is scope. A first pass is not a programme. It is one model, on the ledger and the invoice lines you already have, built to show whether there is anything worth a second conversation. If there is not, the project has cost a few weeks and the objection was right.
2. “We tried ABC and it was too complex”
It was. Classic activity-based costing, as implemented in the 1990s, needed interviews to estimate how every person split their time across activities, and those estimates went stale the week they were collected. Models took months to build, cost a fortune to maintain and were quietly abandoned when the person who owned them moved on. Anyone who says they tried it and it was too complex is describing the method accurately.
What changed is the time equation. Time-driven activity-based costing replaces the interview with two questions per activity: what does a minute of this department cost, and how many minutes does this unit of work take, including the extra minutes when the order is international, or returned, or has forty lines. The answers come from the transaction data and from the people who do the work, not from a survey, and they update when the data updates. The complexity that killed ABC was maintenance, and maintenance is precisely what the time-driven version removed. We wrote up the difference in detail on our TDABC versus ABC page.
3. “The model lives in a spreadsheet, built by someone who left”
This objection is usually offered as a reason not to start, and it is in fact the strongest reason to. The pattern is always the same: a cost model in Excel, built two years ago by someone who has since left, formulas nobody fully understands, three weeks to refresh each quarter, and million-euro pricing decisions still resting on it. Nobody is confident in the numbers and nobody wants to be the person who opens the file.
What we answer is that the spreadsheet is not the model; it is the container, and it is the wrong container for anything that has to be recalculated, audited and explained. The allocation logic inside it, if it is any good, can be lifted out in a matter of weeks and rebuilt where each allocation is traceable from the ledger line to the customer margin and can be rerun monthly in hours. If the logic is not any good, that is worth knowing before the next price list is signed. We have written about that migration in moving from spreadsheets to a costing model and about keeping the result alive in cost model governance.
4. “We already know which customers are profitable”
Most companies know their total revenue and their total cost, and they know gross margin by customer. What they rarely know is what each customer costs to serve after gross margin: the order handling, the deliveries, the returns, the credit terms, the account management, the complaints. Those costs sit in overhead, and overhead is spread by a rule someone chose years ago, so the customer margin everyone quotes is gross margin minus an average.
What we answer is a picture rather than an argument. Rank customers by profit after cost-to-serve and plot the cumulative total. The line rises, peaks and then falls, because the customers at the far end destroy value that the ones at the front created. It is called a whale curve, we have drawn it across more than 150 engagements, and its shape is remarkably consistent. The customers at the tail are not usually the small ones. They are the ones with the most discounts, the most special handling and the most people assigned to them, which is why the intuition that big customers are the best customers so often fails the test. How to build the ranking is on our cost-to-serve calculation page.
5. “We allocate overhead by revenue and it works”
It works in the sense that the numbers add up. It does not work in the sense that matters, because revenue-based allocation charges the most overhead to the customers who bring the most revenue, regardless of how much work they generate. A large customer with clean, predictable orders absorbs a large share of overhead it never caused. A small customer with urgent deliveries, frequent returns and a dedicated account manager absorbs almost none. The rule punishes the customers you most want to keep and hides the ones costing you money.
What we answer is to allocate by activity instead: by delivery stops, order lines, support minutes and handling time, the things that actually consume the hours of the people in the warehouse and on the service desk. This is the core of time-driven costing, and it is the reason a model built on the same ledger can rank customers in a different order. The reordering is not a modelling artefact. It is the subsidy that revenue-based allocation had been hiding, made visible.
None of the five objections is wrong about the past. Each is wrong about what a first pass costs today, and about what it finds.
What we ask in return
One thing: twelve months of transactions and the ledger for the same period. With those, a first cost-to-serve model can be built and read in a few weeks, and the conversation moves from whether to look to what was found. If you would rather score your own starting point first, the Profit Check takes 12 to 15 minutes and covers the seven dimensions where cost models usually fall short.
A thirty-minute call is enough to say whether a first-pass model on your transactions is worth building, and what it would need from you.
Related reading: customer profitability analysis, the cost of unused capacity and a distributor case study.