Maybe We Need to Bring Back the Stopwatch

My father-in-law was a time and motion study man at one point in his career. More than 50 years ago, he would literally observe work being done: what was the process, how long did each stage take, how many times was it done, where were the delays – and ultimately, how much resource did you need to do it?

We might think the stopwatch sounds rather dated now.

I’m increasingly wondering whether we should bring it back.

Not literally, perhaps. But the thinking behind it.

Throughout my career I’ve used capacity planning in much the same way. Map the process. Understand the volumes. Work out how long the activities take. Identify the exceptions. Then you know what it really takes to deliver the work.

It can be an absolute pain to do – but I’ve never once regretted doing it. There are other methods – waste walks, for example – but in my opinion nothing gets to the heart of the issue like factual data.

We know how many people we have. Do we know how much work we have?

Capacity planning turns the normal workforce question on its head.

Instead of starting with “We have 40 people – how many could we manage with?”, you start with “This is the work. How much of it is there, why are we doing it, and what does it take to deliver it?”

Only then do you get to the resource.

It sounds obvious, but I’m not convinced we do it consistently anymore – particularly in back-office functions.

Manufacturing, logistics, contact centres and other large-scale operational environments haven’t forgotten it. They routinely understand throughput, capacity, cycle times and bottlenecks.

In HR, Finance and other corporate functions, we can usually tell you our headcount and what it costs.

But ask how many transactions they actually process, how long each takes, what percentage are exceptions and, crucially, why those exceptions exist – and the answer can be surprisingly hard to find.

When you do have that data, it changes the conversation. I had a coffee and catch-up recently with a wonderful ex-colleague, Mark, who was excitedly telling me about a group of services he and I had worked on centralising. The services were later outsourced successfully, but at some point the arrangement no longer fitted the way the outsourced organisation that owned the service wanted to go forward, and the conversation turned to whether the unit should be disbanded and the services handed back to each partner. Mark worked with all the partners to show the efficiency, savings and payback, as well as how well the services could scale, so that they could make an informed decision – and they chose to set the unit up as a separate organisation, wholly owned by the partners. That sounds simple. It wasn’t. But the data made the decision easier, and they are now in year two of continued efficiencies and strong performance – all backed up by exactly the kind of data I’m talking about.

And the exceptions are where it gets interesting

If you have 10,000 transactions that take ten minutes each, the capacity calculation isn’t particularly difficult.

But that’s rarely how work actually happens.

Perhaps most take five minutes. Some take 25.

Why?

Someone entered the wrong information. An approval is missing. A system doesn’t talk to another system. Data has to be re-keyed. A local process is different. Someone has built a workaround to bridge a gap between two technologies. Or perhaps we’re simply doing something because we’ve always done it.

That’s when capacity planning stops being an exercise in calculating headcount and becomes business process redesign.

The interesting question is no longer “How many people do we need to process this?” It becomes “What’s creating this work?”

Fix the cause of the exception and the 25-minute transaction often becomes a five-minute one – or disappears altogether. Only then is it worth asking what’s left to automate, and what still needs a person.

Which brings me to AI

We are understandably excited about what AI can do.

But I’m increasingly concerned that organisations are jumping from “We have 30 people doing this” straight to “How much of it can AI do?”

There is a rather important bit missing in the middle: what exactly are those 30 people doing?

It’s the same blind spot I wrote about in Before You Delete the Role, Do You Understand What You Are Deleting? – just with technology instead of a restructure.

Before deciding whether work belongs with a person, an automated workflow, AI, an outsourced provider or somewhere else entirely, surely we first need to know what that work involves and why it exists.

Process → volume → time → exceptions → demand → value.

Then make the call on where it belongs.

Perhaps we’ve just given the stopwatch better branding

There is an irony here.

We now talk about process mining, task mining, workflow analytics, digital twins and AI-powered process discovery.

These tools are vastly more sophisticated than anything available to my father-in-law.

But underneath all that technology, I suspect he’d still ask some remarkably simple questions:

What’s the process? How often do you do it? How long does it take? Where does it go wrong? And why are you doing it that way?

If we couldn’t answer them, I’m not sure he’d be particularly impressed by our transformation strategy.

So perhaps we don’t actually need to bring back the stopwatch. But we could do with bringing back some of the thinking that went with it.

Because before we redesign organisations, remove roles, introduce another technology or automate work with AI, we ought to be able to answer his questions.

Maybe the stopwatch wasn’t so old-fashioned after all.

Related insight

Don’t Design the Job Yet. Design the Work.

The Shadow System – and How to Avoid It