Your Warranty Claim Needs Evidence, Not Anecdotes

QUICK ANSWER. Every operator knows which machine model causes the most trouble. Almost none can prove it. Warranty claims, procurement decisions and supplier reviews are settled with evidence — fault frequency per model, normalised by actual usage — not with experience. That evidence already exists inside your fleet. It just isn’t assembled.

Ask any service manager which model gives them the most grief. You will get an answer in under two seconds, with feeling.

Ask them to prove it, and the room goes quiet.

That gap — between what an operation knows and what it can demonstrate — is worth real money every time a warranty claim, a procurement round or a supplier review comes up.

Why do warranty conversations stall?

Because the two sides of the table are arguing from different evidence, and only one side’s evidence is written down.

The operator brings accumulated field experience: this model fails, that component always goes, our technicians dread this installation. All of it true. None of it auditable.

The manufacturer brings warranty statistics across a global installed base, where a difficult regional pattern is diluted by tens of thousands of machines behaving normally elsewhere.

“An anecdote cannot be audited. A dataset can be argued with — which is exactly what makes it useful.”

So the conversation resolves the way unevidenced conversations always resolve: in favour of whoever has the numbers.

What actually counts as evidence?

Not “this model is unreliable.” Something a supplier’s engineering team can act on:

• Fault frequency per model, normalised by usage. A machine serving four thousand cups a month and one serving four hundred are not comparable units. Faults per thousand cups is a claim; faults per machine is a complaint.
• Component-level failure patterns. Which part, at what age, under what duty cycle — not “the machine broke.”
• Site and environment context. Water hardness, ambient conditions, installation type. The same model can behave very differently across two customers.
• Repeat-visit rate and time-to-resolution per model. A machine that takes two visits to fix costs you more than one that fails slightly more often and is fixed first time.

Each of those is a number your operation already generates. None of them is a number most operations can currently produce on request.

Where does that evidence already exist?

In three systems that rarely meet:

1. The machines. Error codes, cups served, uptime, duty cycle — the raw record of what actually happened, machine by machine.
2. The service history. What the technician found, which component was replaced, how long it took, whether they had to come back.
3. The commercial record. Machine age, warranty terms, contract, site. The context that tells you whether a fault is your cost or the supplier’s.

Individually, none of the three proves anything. Joined together — and normalised across brands, because almost no operator runs a single-brand estate — they produce a defensible claim. That joining is the hard part, and it is the same problem we described in Why Every Coffee Machine Brand Gives You Different Numbers.

Scale is what makes it worth solving. EVA, the European Vending & Coffee Service Association, sizes the European vending and office coffee service market at roughly 4.5 million installed machines across 24 markets. Fleets that size generate an enormous evidential record every single day, and most of it is currently thrown away as soon as the ticket closes.

Doesn’t this make the manufacturer relationship adversarial?

It does the opposite, and this is the part operators tend to get wrong.

A warranty claim backed by usage-normalised fault data is cheaper for a manufacturer to process than a dispute built on assertion. It arrives pre-argued. It can be routed straight to engineering instead of bouncing between commercial teams for six weeks.

More than that: manufacturers genuinely want this data. Field failure patterns from real mixed-fleet deployments are among the most valuable things an engineering department can receive, and they almost never get them in usable form. An operator who can supply structured evidence stops being a complaining account and becomes a design partner.

The relationship improves precisely because the conversation stops being about whose memory is better.

What changes at the next procurement review?

You stop buying on price and specification sheet, and start buying on demonstrated cost to serve.

The question shifts from what does this machine cost? to what has this model actually cost us to keep running, per cup, across the sites where we’ve installed it? That is the same question as which machines actually make you money, asked one level up — at the model and supplier level rather than the individual machine.

It also compounds with how you run service. A fleet already segmented by usage and risk — see Service the Machine When It Needs It — is a fleet that already knows which machines are working hardest, which is exactly the normalisation your warranty evidence needs. And a fleet that tracks whether a job was fixed first time, as in The Most Expensive Technician Visit Is the Second One, is already collecting the repeat-visit signal.

The full set of implementation actions for building that evidence base, and the management KPIs that keep it honest, is Step 8 of the 9-step operating model in our office coffee service report. The argument is here. The method is in the report.

Key takeaways

• Every operator knows which model causes trouble; almost none can prove it — and unevidenced conversations resolve in favour of whoever has data.
• Evidence means fault frequency normalised by usage, component-level patterns, site context and repeat-visit rate — not “this model is unreliable.”
• The data already exists across three systems that rarely meet: the machines, the service history and the commercial record.
• Structured evidence makes the manufacturer relationship more collaborative, not less — it arrives pre-argued and routes straight to engineering.
• At procurement, the question becomes demonstrated cost to serve per model, not purchase price.

The next time a supplier asks you to justify a claim — what exactly would you send them?
Request the white paper — From Coffee Machines to Profit Machines: a step-by-step model for a more profitable OCS operation.