Measuring SLA breach in fitness operations: lessons from the floor

I remember the exact moment I realised I had no real idea whether we were meeting our SLAs or not.

It was a Monday morning, nine years into my time running operations for a mid-sized leisure operator with seven sites across the Midlands. A regional manager called to say that the free-weights area at our busiest site had been half-closed over the weekend — a snapped cable on a functional trainer, a part that was apparently ordered the previous Wednesday. I pulled up our maintenance system. The ticket was logged. The part was on order. Every box was ticked.

But the machine had been out for five days. Our SLA said urgent faults should be resolved within 48 hours. Nobody had flagged a breach. Nobody had escalated. The ticket just sat there, quietly failing, while members stepped around it.

That was the day I accepted that logging a fault and actually measuring whether you resolved it in time are entirely different things. We were doing the first. We were not doing the second.

Why SLA measurement is harder than SLA writing

Most operators I have spoken to over the years can show you a document with SLA commitments in it. Response times, resolution times, escalation paths — it is all there, usually written during a tender process or as part of a franchise agreement. The document is rarely the problem.

The problem is that measuring whether those commitments were met requires something the document does not provide: a reliable, timestamped record of what actually happened, compared automatically against the threshold that was promised.

In practice, what most teams do is review tickets manually at the end of the month and make a rough judgement call. That approach has several obvious weaknesses:

• Breaches that resolved themselves (equipment fixed on day three of a 48-hour SLA) still look like breaches in the data.
• Tickets that were never properly closed sit in an ambiguous state.
• Disputes about when a fault was first reported — versus when it was first logged — are settled by whoever shouts loudest rather than by a timestamp.
• Seasonal patterns (Christmas closures, bank holidays) are rarely accounted for in breach calculations.
• Nobody looks at near-misses — tickets that resolved with two hours to spare, week after week, across the same piece of kit.

I was guilty of all of these. We had a spreadsheet that a very dedicated ops coordinator updated every Friday. It was accurate about 70 per cent of the time. The other 30 per cent was a source of constant internal argument.

What a breach actually means in operational terms

Before you can measure a breach, you need to agree precisely what one is. That sounds obvious. It took me an embarrassingly long time to work through the detail properly.

Here is what I eventually settled on as a working definition for our sites, which I share not as a template but as an illustration of the thinking you need to do:

1. Breach onset: the moment a fault ticket moves past its target resolution time without a completed close-out action on file.
2. Breach confirmation: a second timestamp, applied automatically, that records when the breach was first identified — distinct from when it was fixed.
3. Breach severity: a classification that distinguishes between a treadmill being out of service for 54 hours against a 48-hour SLA and a treadmill being out for three weeks.
4. Breach attribution: whether the delay was caused by the maintenance team, the parts supplier, a third-party engineer, or factors genuinely outside anyone's control (flooding, for example).
5. Breach resolution: the documented fix time, signed off by the attending engineer and cross-referenced against member-facing signage removal.

Without all five of those elements in a consistent record, your breach data is contested at best and useless at worst. Ours was frequently contested.

The escalation gap that no one talks about

One pattern I have seen at almost every operator I have advised since leaving full-time ops is what I call the escalation gap. This is the period between a ticket breaching its SLA and anyone with authority to do something about it finding out.

In a well-run operation, that gap should be close to zero. In reality, it is often measured in days.

Why? Because escalation depends on someone noticing. And noticing depends on someone checking. And checking depends on someone having time. In a lean ops team covering multiple sites, that chain breaks constantly.

The fix is not to hire more checkers. The fix is to automate the escalation trigger so that when a ticket crosses its resolution threshold without a close-out, the right person is notified immediately — not at the end of the week, not during the monthly review, but within the hour.

I have seen escalation gaps of 72 hours on urgent faults. By the time anyone with authority knew there was a breach, the breach had tripled in length. Members had already had three days of reduced access to kit they pay to use.

How to structure SLA tiers for gym equipment faults

Not every fault is equal. A broken cupholder on a treadmill is not the same as a treadmill with a safety stop fault. Treating them as the same in your SLA framework creates two problems: you under-respond to serious failures and you waste resource chasing minor ones.

A tiered structure I have found workable in a multi-site gym context looks like this:

Tier 1 — safety-critical: any fault that presents a risk of injury or requires the equipment to be taken out of service immediately. Target resolution: 24 hours. Escalation to ops director if unresolved after 12.

Tier 2 — high-impact: equipment that is a primary draw for members (treadmills, rowing machines, cable stations in a busy free-weights area) and is fully out of service during peak hours. Target resolution: 48 hours. Escalation to site manager after 24.

Tier 3 — moderate impact: equipment that is partially functional or out of service during off-peak only. Target resolution: 5 working days.

Tier 4 — low impact: cosmetic faults, minor wear, non-safety consumables. Target resolution: 10 working days, grouped into planned maintenance visits where possible.

The specific numbers matter less than the principle: you need different thresholds for different fault types, and those thresholds need to be hard-coded into your ticketing system so that breach measurement happens automatically at the right threshold for each ticket.

What the data tells you that gut feel does not

Once you have a consistent breach measurement framework in place, the data starts to show you things you would never spot through manual review.

At one of our sites, we had what I thought was a reliable treadmill bank — eight machines, rarely flagged, members seemed fine. When I looked at the breach data properly, I found that two specific machines had generated eleven near-miss breaches in six months. They were always fixed inside the SLA window, but only just. When we investigated, the pattern was a recurring belt-tension issue that the attending engineers had been treating as a one-off each time, rather than flagging as a recurring fault requiring a parts replacement.

A proper breach log — one that captures near-misses as well as confirmed breaches — would have surfaced that pattern at month two. We caught it at month seven.

Similarly, breach data by tier will tell you whether your Tier 1 response is actually working or whether the 24-hour target is aspirational rather than operational. If you are breaching Tier 1 SLAs 30 per cent of the time, that is a structural problem — either in your engineer availability, your parts procurement, or your escalation chain — that no amount of goodwill will solve.

How CRM data connects SLA performance to member behaviour

This is something I came to late in my career, and I wish I had understood it earlier.

SLA breach data, on its own, tells you about your operations. SLA breach data connected to your member CRM tells you about your business.

Specifically, it lets you ask: do members who were active at a site during a period of repeated SLA breaches churn at a higher rate than members who were not?

In almost every case I have looked at, the answer is yes — and the relationship is stronger than most operators expect. A member who visited your facility twice during a week when the cardio kit was repeatedly out of service on Tier 2 and Tier 3 breaches is meaningfully more likely to cancel in the following 30 days than a comparable member who visited during a breach-free week.

You cannot act on that insight if your maintenance data and your membership data live in separate systems with no connection between them. Platforms like Pulse Fitness are built to surface exactly this kind of cross-functional insight — linking equipment status, breach history, and member activity in a single view so that your ops and retention decisions are informed by the same picture.

Building a breach review into your monthly ops rhythm

The last thing I will say is this: measuring SLA breach is not a one-time project. It is a discipline that needs to be built into your regular operating rhythm.

Here is what a monthly breach review should cover, in my view:

• Total breaches by tier in the period, compared to the same period last month and the same month last year.
• Average breach duration by tier — not just whether a breach occurred, but how bad it was.
• Sites ranked by breach frequency, with a flag for any site that has moved more than one position in the ranking month-on-month.
• Attribution split: how many breaches were caused by parts delays, engineer availability, or internal process failures.
• Near-miss count: tickets that resolved within 10 per cent of the SLA threshold, flagged as a leading indicator.
• Actions agreed at the previous review and whether they were completed.

That last point matters more than most. A breach review that produces a list of actions and then forgets them is worse than no review at all — it creates the illusion of accountability without the substance of it.

I got this wrong for years. I have seen operators who are still getting it wrong. The good news is that the framework is not complicated once you have decided to take it seriously. The hard part is the decision, not the implementation.

---

If you want to see how Pulse Fitness handles breach measurement, escalation automation, and CRM-linked equipment data in practice, book a demo at https://www.pulsefitness.ai/demo-request.