Do Customers Stay, or Have They Simply Not Left Yet? On the Absence of Switching Costs

When customer attrition reaches the board agenda, the discussion almost invariably begins from the same coordinates: which account left, why it left, and who carried responsibility for it. The reciprocal question — why the remaining accounts have stayed — is rarely put to the same table, and to the extent that staying is treated as a condition requiring no explanation, the actual resilience of the base is left permanently unmeasured. Yet the renewal figure presented in that very deck compresses two entirely different realities into a single number, since a customer who evaluated an alternative and elected to remain and a customer who never evaluated anything at all sit side by side, indistinguishable, in the same percentage. The difference between them remains invisible for as long as contracts continue to renew, which is precisely what makes it the most expensive blind spot a company can carry into its planning cycle.

A second expression of the same pattern surfaces in pricing discussions. When a proposed increase reaches the table, the reflexive objection from the commercial side is usually constructed around a handful of named accounts — this one will not absorb it, that one is already sensitive — and the conversation closes before it can become a structural question about the base as a whole. The objection is itself a data point of considerable quality: the intensity with which a sales organisation resists a price movement constitutes the most honest internal estimate available of how easily customers could leave, and it is almost never recorded anywhere. That a company has not attempted a price increase in several years is frequently interpreted internally as commercial discipline; its functional meaning, more often, is that embeddedness has never once been tested, and that the organisation has protected itself from learning something it would rather not know.

The structure underlying these observations is what the entrepreneurship literature calls switching-cost weakness — a position in which moving to a competitor imposes on the customer almost no monetary, operational or cognitive burden. The mechanism is not single-layered. Switching cost is produced across at least five distinct surfaces, each capable of being weak or strong independently of the others. The first is data: where a customer's history, preferences and configuration accumulate inside the provider's systems, departure means not merely the termination of an agreement but the surrender of institutional memory. The second is workflow, in that a customer whose internal processes have taken shape around the form of the provider's output faces a migration that is itself a project requiring budget and sponsorship. The third is technical integration, the fourth is the learning curve embedded in the user population, and the fifth is the contract calendar together with the mechanics of notice and termination.

This condition does not originate as a defect; at a particular stage it is entirely functional. Early on, low switching cost operates as leverage in the seller's favour, because the same absence of friction that makes departure easy also makes it easy for a prospect to leave an incumbent and adopt a new entrant, thereby lowering the cost of entering the market at all. A frictionless contract structure, a trial period without commitment and a credible promise of data portability shorten the sales cycle and feed growth velocity directly. The difficulty lies not in the shortcut but in the shortcut persisting after the conditions that justified it have changed: once a company crosses from an acquisition-led phase into a retention-led one, that same frictionlessness becomes a door that opens in one direction only. Whatever makes winning easy makes losing easy, and the symmetry turns adverse as the portfolio grows.

The institutional cost registers first not on the balance sheet but in pricing power. In a base with low switching costs, the capacity to pass input cost increases through to price is structurally constrained during an inflationary period; the company either compresses its margin or surrenders a portion of its base, and being forced to choose between the two is by itself evidence of positional weakness. This compression usually appears in the income statement as a quiet erosion of gross margin distributed across several years, and internally the cause is almost always attributed to the supply side — input costs rose, procurement was late — when the actual source of the mechanism sits on the demand side. A second cost accumulates in the cost of sale, since the obligation to replace each departing account converts sales and marketing expenditure from growth spending into standing-still spending. A third emerges in the working capital cycle, because short-tenor agreements carrying live termination risk make prepayment and long-commitment negotiations largely unavailable.

The sharpest cost, however, crystallises in a capital transaction. At the review table, the party assessing revenue recurrence does not examine the renewal rate; it examines the structure sitting behind renewal, and it asks whether there exists a concrete answer to the question of what specifically would break inside the customer's own operation upon departure. Where no such answer exists, the working definition of recurring revenue narrows in practice, and a base that looks stable across cohort analysis year after year is discounted in the forward cash flow projection at a materially higher risk premium. The corresponding effect in the closing structure is rarely a reduction in the headline price; more often it is the distribution of consideration across time — an enlarged earn-out share, key customer agreements converted into conditions precedent, a dedicated heading for customer loss within the representations and warranties, an escrow ratio calibrated upward. On the sell side this is generally experienced as a valuation argument, when its actual subject is the character of the revenue.

What neutralises the condition is not an intensification of customer relationships or an increase in the number of account managers, since both attach durability to individuals rather than to the institution and are therefore lost precisely when the individual leaves. The structural intervention consists of building switching cost deliberately and measurably, and it separates into four components. The first is an architecture in which the customer's data accumulates inside the provider's environment, so that departure carries the cost of losing a record. The second is binding the customer's workflow to a process rather than to a deliverable — not the report being delivered, but the customer's own approval chain having come to rest on the format of that report. The third is graduated commitment in contract architecture, rewarding a move to a multi-year structure through expanded scope rather than through price concession. The fourth is opening second and third service lines, since the difference in departure probability between a single-line account and a three-line account constitutes a stronger bond than any single clause.

Measuring these components requires a discipline of its own, because switching cost is not a directly observable quantity and can be tracked only through indirect indicators. Among the usable ones are the differential in twelve-month attrition between cohorts to which a price increase was applied and cohorts to which it was not; the proportion of renewals concluded after the customer had seen a competing proposal; the number of individuals within the customer organisation who touch the provider's output in a given month; and the average interval between notice of termination and actual departure. This last measure is particularly informative, since as the distance between the decision to leave and the execution of that decision compresses, the space the provider occupies within the customer's operation is demonstrably narrowing — a customer able to exit within a fortnight has, in operational terms, already exited.

BEIREK treats this not as a customer relations problem but as a revenue architecture problem, and the intervention typically proceeds across three records. The first is a mapping of the base, in which each account is scored separately against the five surfaces of switching cost and revenue is stratified according to that scoring, so that in place of a single renewal rate the company holds revenue tranches carrying distinct resilience profiles. The second is a pricing test record: rather than applying price movements to the entire base simultaneously, increases are trialled in selected cohorts on a staged basis, with the response assessed against documentation rather than against the recollection of whoever managed the account. The third is the reconstruction of contract architecture, under which termination mechanics, notice periods, data return obligations and scope-expansion triggers are designed at the moment of signature rather than negotiated at the moment of renewal.

The cadence governing this work is considerably more frequent than an annual review and is deliberately held independent of the renewal calendar, since an assessment conducted at renewal arrives months after the decision was in fact taken and records only the outcome. A quarterly examination of the base instead makes visible, before any loss occurs, which accounts have seen their switching-cost profile weaken — through declining depth of use, a falling number of contact points, or a contraction of scope. In companies with a capital transaction on the horizon, the same exercise allows the company to ask itself the question the counterparty will eventually ask, which alters the negotiating position materially: a weakness discovered at the review table and a weakness the company has documented in its own record and demonstrably worked upon carry entirely different weight, even where they describe the identical fact pattern.

Switching-cost weakness is, in the end, not the invoice for poor management but the natural and deferred invoice for a particular logic of growth; the frictionlessness that made customers easy to win is structurally identical to the frictionlessness that makes them difficult to keep, and where the transition between those two phases is not managed as a deliberate design decision, it resolves itself adversely by default. The question worth putting to the table is therefore not whether customers are satisfied, since satisfaction is a weak predictor of future behaviour once the cost of leaving approaches zero. The question is whether, should the largest account decide today to depart, there exists a concrete description of the workload that decision would trigger inside its own organisation — and, no less importantly, in whose record that description is currently held.