Field-Service Advisory

The installed-base blind spot: why service revenue leaks before sales ever sees it

You may not have a service-sales problem. You may have a problem knowing what you sold, who has it now, and whether it is still running. A seller cannot recover an opportunity the operating system never identified.

By Paula Navarro · Cosmicalley  |  July 2026  |  8 min read

The aftermarket is where the growth is, and industrial manufacturers know it. BCG's services benchmark study reports that for most companies the growth rate of aftermarket services now exceeds that of new equipment sales, and that companies prioritising these services generate a third or more of their total income from them. Spare parts carry gross margins averaging around half of sales. The prize is not in dispute.

The same study says something quieter and more useful: between a quarter and half of the companies surveyed struggle to capture that services potential — and for some of them, BCG names the reason as a lack of information about installed equipment: its condition, location, usage and service history. Not weak selling. Missing knowledge of what is out there.

So the reflex is predictable: hire more service sellers, build a renewals desk, launch campaigns against the installed base. All reasonable. All of it assumes the installed base is a list you can work.

"We’re at X% contract attachment." Of what?

Fill in your own number — it changes nothing, which is the point. Attachment rate is the most quoted metric in service and the least examined, and the figure is the least interesting part of it. The question is: per cent of what? Of every unit ever shipped? Of the units you believe are still running? Of the units whose current owner you could name? Of the units you could actually send an email to tomorrow?

Those are four different denominators, and they produce four different numbers from the same operation. A service organisation reporting attachment against everything it ever shipped is measuring its own manufacturing history, not its commercial performance. One reporting against the assets it can actually identify and contact is measuring something it can manage.

Before you improve the numerator, validate the denominator.

Four populations, not one

The installed base is not a single thing. It is a set of nested populations that get smaller — and more real — at every step.

Everything shipped is a manufacturing fact. What is still active removes the retired, scrapped and quietly replaced. What is known and coverage-resolved removes everything whose owner, location or coverage status you cannot state without guessing — resolved means you can say what coverage the asset carries, including "none", which is often exactly where the opportunity sits. And what is commercially actionable is smaller still: it needs a valid contact and a visible lifecycle trigger — a warranty ending, a contract expiring, an asset reaching the age where a modernisation conversation makes sense.

Only the last population can enter a reliable, targeted commercial motion today. The rest must first be resolved, enriched or timed.

Where the opportunity goes dark

This is our reading of the mechanism rather than a statistic, and it is worth stating plainly because it is where the revenue actually goes. An asset ships, often through a distributor, and the serial number and the end customer are recorded in different systems — or the end customer is never recorded at all. The machine is later moved, resold, or absorbed into another plant. Warranty and contract dates live in a third system, and no longer agree. Service history sits in the field-service platform, unlinked to the commercial record. And the named contact left the company two years ago.

Nothing here is a lost sale. Each one is a lost record. But the effect is identical, because the consequence is the same in every case: no reliable commercial trigger exists, so the asset never enters a sales motion at all. It isn't lost in the pipeline. It never reached the pipeline.

The largest segment is "unknown"

Plot your installed base honestly — asset age against economic potential — and something uncomfortable happens. The assets you can act on appear as dots. The assets whose owner, location or entitlement you cannot establish have no coordinates at all. They cannot be plotted, because plotting requires knowing something about them.

That is the blind spot, and it is not a metaphor. The biggest commercial population in many service organisations is not the uncovered assets everyone argues about in the pipeline review. It is the population whose status is too uncertain to enter any commercial process — invisible to the campaign, invisible to the forecast, invisible to the attachment metric that quietly excludes them by not being able to see them.

Why the leaders attach at the moment of sale

There is a finding in the BCG benchmark that reads like a footnote and is, in fact, the whole argument. The leading companies sign 31% more of their customers to maintenance contracts at the time of purchase than the industry average.

BCG reports the performance difference, not its cause — and better sales discipline is surely part of it. But one likely advantage is timing. At the point of sale, visibility is usually at its peak: the serial number, the customer, the site, the configuration, the commissioning date, the warranty end and a live commercial contact are more likely to be known together than they ever will be again. Not always — a sale through a distributor can be partly dark from day one, which is where the leakage starts earliest. But from the day the asset ships, the record decays: it moves, it changes hands, the contact leaves, the systems drift apart.

That does not prove visibility causes the higher attachment. It does mean the leaders are selling at the one moment when the asset is most visible, while everyone else tries to sell later, into a record that has already degraded — and then calls the shortfall a sales problem. The timing advantage is difficult to ignore.

You don't need a perfect record. You need a decision-grade one.

The wrong conclusion here is "build a flawless digital twin of the installed base before selling anything." A flawless digital-twin programme can easily become a multi-year transformation — and the revenue does not wait. The right conclusion is narrower and much more useful: establish the minimum record each commercial motion actually requires, and stop there.

Different revenue motions need different asset fidelity. Treating installed-base data as one monolithic quality problem is what makes it feel unsolvable. Treating it as "what is the minimum I need to trigger this motion?" makes it a sequence of small, financeable steps. It is also, incidentally, how you decide what to fix first: reconcile the fields that unlock the nearest motion, not the fields that would complete the database.

This is worth saying because installed-base information is not merely an IT hygiene issue — it has a value that can be estimated. Peer-reviewed work in product-service systems by Saccani, Alghisi and Borgman builds a decision-analysis model precisely to estimate the value of installed-base information in a service context. It is a single case study, so it does not tell you what that value is in general — what it demonstrates is narrower and more useful: the value of installed-base information can be modelled, rather than treated as an intangible data-quality benefit.

When visibility will not save you

The thesis has a hard limit, and it should be stated before anyone builds anything. Visibility does not create demand. A perfectly reconciled installed base cannot rescue a weak service value proposition, poor service delivery, uncompetitive pricing, a strong independent-maintainer market, low asset criticality, or a customer base that has decided to maintain its own equipment. If your service offer is not wanted, knowing exactly who is not buying it will not help.

The claim is narrower than "fix the data and the revenue comes." It is this: installed-base visibility does not guarantee service revenue — but its absence guarantees that some opportunities will never be seen, correctly targeted, or measured. You cannot win business you never knew existed, and you cannot manage a number whose denominator you cannot state.

Start with what you cannot see

Before the next campaign, take a sample of the installed base — a product line, a region, a vintage — and try to answer five questions per asset: is it still running, who operates it now, what coverage does it carry, when does that coverage end, and would anyone answer if you called. The proportion you cannot answer is your real starting number. It may prove to be the largest one on the page — and it is often the one management has never measured. It is also the only one whose improvement compounds into every motion you run afterwards.