What it cost to tell the truth

Close hand-drawn sketch over a production coordinator's shoulder as she edits one row on a scheduling screen, with a machine shop visible through the window behind.

I asked a production coordinator to show me what she did when she knew an order would be late.

She opened a job due out on Thursday, changed the date to the following Tuesday, and carried on explaining something else. Then she added, almost as an aside, that otherwise it stays on the overdue list even though the customer already knows.

That last sentence is the whole story that I’m going to share, and at the time it took me a moment to catch it.

The business makes small batches of precision components for robotics and industrial automation customers. About 85 people, modern machines, digital job tracking, and a dashboard in the managing director’s office showing 96% on-time delivery.

He was about to take on another sizeable contract without adding any production capacity, and the dashboard was a big part of why he felt comfortable doing it. Complaints about late deliveries were coming in, but against a figure like 96% they looked like isolated incidents rather than a pattern that would cause a problem.

They had two production coordinators who kept on top of the delivery dates. They scheduled jobs, chased materials and fielded the calls from customers asking where their parts had got to. When a job slipped, one of them would open the order and replace the promised dispatch date with the latest realistic estimate.

Operationally that’s a sensible thing to do. The shop floor needs a schedule that reflects reality, and a plan full of dates that have already passed is no use to anyone.

The difficulty was that the same field fed the delivery report. The dashboard compared the actual dispatch date with whatever promised date was sitting in the system at the time, so a promise that had been met and a promise that had been rewritten looked the same to it.

Sketch of a production morning meeting, one person presenting a flagged report while a coordinator sits back with folded arms.

Now, why were the coordinators moving those dates so readily?

Part of it was the system giving them one field for two entirely different jobs, with nothing to suggest that changing it did anything beyond tidying up the schedule.

The rest came from the morning production meeting. Overdue orders showed up in red on a report reviewed every day, and if an order stayed red, you’d be asked why you hadn’t sorted it out and asked the next morning again. And the morning after that, in front of everyone.

Updating the date cleared the flag. Leaving it there bought you a difficult conversation you’d already had, about a delay you’d already told the customer about, with someone who couldn’t do much to fix it anyway.

So one route took a few seconds. The other cost you a few uncomfortable minutes a day for as long as the job ran late. From where the coordinator was sitting, both produced the same outcome for the customer.

I want to be clear: nobody hid anything from anyone who mattered. The customers had been told. They were managing the meeting.

Close to 400 orders were dispatched over the previous three months, and I checked them twice.

Against the date currently in the system, 384 had gone out on time. 96%, exactly as the dashboard said.

Against the date first confirmed to the customer, only 312 went out on time, or 78%.

So 72 orders became punctual through rescheduling, and those 72 left on average six working days after the date the customer was originally given. Over a week late, on average, and completely invisible to the business.

Close sketch of a printed daily report on a meeting table with several rows flagged, and a hand pointing towards them.

Here’s what I always come back to.

There was nothing wrong with that dashboard as a piece of software. It calculated exactly what it was asked to calculate, using data entered in good faith by two people doing a difficult job well.

A business produced the wrong number by making honest reporting expensive for the two people closest to the facts, then reading their entirely sensible response as evidence of good performance. The measurement problem sat several steps upstream of the measurement.

We separated the two dates. One field captured the original customer commitment, another the current production forecast, and a third recorded genuine changes with a reason rather than silently overwriting the first promise.

The morning meeting changed as well, which mattered more. Coordinators were now expected to flag likely delays early, and managers used those warnings to decide what to expedite and what to renegotiate with the customer. Raising an overdue order stopped being treated as an admission that you’d failed at your job.

At this point, it’s worth saying the operations director introduced that red-flag report herself, presumably to get a grip on late deliveries, which is a reasonable thing to want. When I suggested changing it, no one in the room defended it. She agreed to drop it with little argument.

A report nobody would speak up for had run long enough to reshape how two people did their work every day. The managing director phased in the new contract and approved extra cover at the production bottleneck.

Over the following three months, 91% of dispatched orders met the date the customer had originally been given, up from the reconstructed baseline of 78%.

And the number on the dashboard went down, from 96% to 91%.

That’s the detail I’d take away from this. Delivery performance improved substantially while the headline figure fell because the old measure tracked something else entirely: how well the business kept its revised promises.

Scientists have a habit here that business could use. A published measurement comes with an account of how it was taken, what the instrument was, and how wrong it might be. The uncertainty travels with the number, because a figure without those conditions attached can’t be judged.

Business strips all of that out. The figure arrives on a slide to two decimal places, and everything about who produced it and what it cost them to produce it honestly gets left behind somewhere between the shop floor and the boardroom.

If a number in your business has been reassuringly stable for a long time, it’s worth finding the person who actually enters it and asking them what happens when the news is bad.

What does it cost someone in your business to tell you something is going wrong? And what is your reporting quietly measuring instead?

Have a brilliant week!

Dave Rogers, The Business Explorer

If this struck a chord, my Substack goes further on the same questions. Want help applying any of it to your business? Book a call.

Need a speaker who’ll bring a fresh perspective? Reach me at info@fuelledfitandfiredup.com.

For 30 years, I’ve worked with business owners and leaders who are stuck on the wrong question. My job is to find the right one.

That usually means sitting with the messy “what if” long enough to turn it into clearer thinking and harder questions.