Bad data made a good team look lazy. Here’s the fix.
Your service desk is busier than the numbers say. Your engineers fix things all day and log them at the end of it, from memory, if at all. Your ticket counts look thin, so someone upstairs keeps asking why you need so many people. And the honest answer, “they’re doing far more than we record”, has never once survived a budget meeting.
That was the exact position on one of the most demanding trading floors in Europe. This service desk case study is about how it got fixed with a gadget borrowed from a restaurant, and how the same stubborn habit of checking what was actually true turned up millions on an invoice nobody had read properly in years.
I’ve built service desks from an empty room more times than I’d care to count. This one taught me more than most, mainly because the customers were traders, and traders do not do patience.
In this blog, we’ll explore:
- why a trader’s “broken” monitor is rarely about the monitor
- how a restaurant ordering device fixed a bank’s service desk statistics overnight
- why reconciling four asset systems ended this service desk case study with a summons to the board
Built From Nothing, Then Quadrupled
Every service desk case study needs a starting point, and this one starts, oddly, in London. The investment banking arm of a large German bank had a slightly unusual arrangement. Its German-speaking support staff were sitting in London, which was every bit as expensive and pointless as it sounds. The job was to bring the operation home to Germany and build a new supplier contract around it.
It started at around 2,000 users. Then the bank merged with a rival, and the user base grew to roughly 8,000: front office, back office, plus the IT teams running the merger itself. Four times the users. Same starting point of nothing. If you want a service desk case study in scaling under pressure, that’s the headline number.
Traders Do Not Report Faults Politely: A Service Desk Case Study in Speed
The service level agreement was brutal by design, because the business it served was brutal by design. Answer the phone within three rings. Get a second-line engineer to the trader’s desk within fifteen minutes. Resolve the issue within an hour.
Compare that with a furniture retailer, where stability and getting it right first time matter far more than speed. This was the opposite world. A trader whose machine is down isn’t in a holding pattern. They’re losing money every minute, because the market has an unhelpful habit of moving whether or not you can see it.
The trading floor was bigger than a football pitch: around 500 positions across roughly 26 rows, with the second-line engineers sitting in the middle of it all. Here’s what actually happened when a ticket came in, and it’s the scene at the heart of this service desk case study. An engineer set off down the row towards the reported desk. Three desks along, a different trader flagged him down with an unrelated problem. He fixed that one first, because ignoring a trader shouting at you across a trading floor is not a viable career strategy.
Then he carried on to the desk he was originally heading for, where that trader asked, not gently, whether he was ever planning to turn up. By the time either ticket was logged, the timing was fiction. No accurate start time. No accurate duration. One of the two fixes often wasn’t logged at all.
The consequence, and the reason this service desk case study exists at all, wasn’t just messy data. It made a hard-working team look underworked. Reported volumes didn’t match the problems being solved, so the obvious management question arrived right on schedule: why do you need this many people if the numbers say you’re handling a handful of tickets a day?
How much of your team’s real work never makes it into a ticket?
The Idea Borrowed From a Waiter
This is the part of the service desk case study people remember, and the fix came from an evening out. Dinner with my wife, in a restaurant where the waiters took orders on handheld devices: order at the table, and the ticket goes straight to the kitchen with a timestamp. No notepad. No walking back to the till. No delay.
The supplier was tracked down the following week and a batch of the same devices went onto the trading floor. Table number became desk number. The drinks menu became the fault category. An engineer logging an issue at the point of contact, on the spot, produced an accurate start and end time for every single ticket, instead of a rough guess reconstructed from memory forty minutes later.
The traders, being traders, treated the whole thing as comedy rather than progress. The classic call was a “dead monitor”. Nine times out of ten the real fault was the physical dimmer button under the desk that switched every screen off overnight, the sort of thing a new starter hadn’t yet learned existed. One colleague, asked for a photo of the broken screen as a first-contact check, took a picture of a monitor that was simply switched off and emailed it in with a perfectly straight face. That’s the kind of workplace it was. The team eventually printed a line on the coffee mugs: check the button under your desk before you call.
Once the statistics were accurate, the headcount question answered itself. The data finally reflected the real volume of work, rather than an artefact of engineers being too busy fixing things to write them down. Nobody had to argue for the team any more. The numbers did it, which is the quiet lesson of almost every service desk case study worth reading: fix the data before you defend the headcount.
The Invoice Nobody Had Checked in Years
The second half of this service desk case study is less charming and considerably more expensive.
The bank’s IT asset records were scattered across four disconnected places: the purchasing system, a separate system used to build and register each machine, the buildings team’s own records of physical location, and cost-centre ownership that changed every time someone moved teams. None of it reconciled. A machine could be bought in one system, built in another, sitting on a desk according to a third, and charged to a department that no longer existed according to the fourth.
So the team built a reconciliation process across all four and corrected the data. That turned up something very specific. The hardware maintenance supplier was invoicing the bank for machines it no longer had, and billing a fair share of what remained to the wrong cost centres.
Fixing it saved millions, which makes this easily the most expensive service desk case study on this blog. It also earned a summons to the board, to explain in person why one of the bank’s largest suppliers was about to see its maintenance revenue fall sharply. Not a normal Tuesday. And exactly the kind of result that never makes it onto a project retrospective slide, because “we checked an invoice properly” doesn’t sound dramatic enough to mention. It should.
When did anyone last check your maintenance invoices against the kit you actually own?
What This Service Desk Case Study Actually Proves
Neither fix in this service desk case study was clever in the sense of complicated. A restaurant ordering device is not sophisticated technology, and reconciling four systems that should already talk to each other is nobody’s idea of glamour. What both had in common was a refusal to accept “that’s how it’s always been logged” or “that’s what the invoice says” as an answer.
If you only take three things from this service desk case study, make them these:
✓ Log the work where it happens. A ticket written up forty minutes later is a memory, not a record, and memory always undercounts.
✓ Reconcile before you renew. Every supplier contract renewal is a chance to check what you’re paying for against what you actually have.
✓ Let accurate data argue for your headcount. A team defending itself with anecdotes loses to a spreadsheet every time; give them the spreadsheet.
On a trading floor running on adrenaline and one-hour fixes, that habit of checking was worth more than any process diagram, because a diagram only works if someone is willing to hold it up against reality. That’s the thread running through this whole service desk case study. It’s what a properly run service desk does every day, and it’s the heart of a Service Desk Assessment & Improvement: finding where the numbers and the real work have quietly drifted apart. For teams who want a shared vocabulary to keep them together afterwards, the ITIL Certifications path is where most service desks start.
Which number in your service desk report would you least like a board member to check?
To find out more about PDCA Consulting’s expert consulting services either:
- Book 30-minute conversation directly: here
- Get in touch via the contact form here
- Connect with me on LinkedIn
Robert Pinnington
Turning struggling Service Desks & PMOs into high-performing ones