A non-practitioner reads the ERP and concludes that the problem is quality. A practitioner asks to sit in on the 6 a.m. shift handover. Three minutes in, they catch that the outgoing supervisor is describing a process adjustment verbally that has never been written down, and that new hires on the incoming shift are nodding along without repeating it back. The quality variance shows up three shifts later, on the incoming side. The ERP never saw the cause. The handover did.

What the practitioner did differently: went to the floor before opening the data.

A non-practitioner asks where the production schedule lives and gets pointed to the MRP screen. A practitioner asks the scheduler what she opens first when she logs in at 7 a.m. The answer is a spreadsheet on her desktop that she has been maintaining for six years, because the MRP never reflected reality. Every planner on every line uses it. The MRP is a compliance layer. The spreadsheet is the operation.

What the practitioner did differently: asked the person who runs the process, not the person who owns the system.

A non-practitioner tells the customer they need to install new sensors to capture downtime reasons. A practitioner walks the floor and finds that operators are already writing downtime reasons on a clipboard posted next to each machine. The data is fifteen years old. It just was never digitized. Three weeks of clipboard data, re-entered and categorized, answers the original question without a single new sensor.

What the practitioner did differently: treated the absence of a system as evidence that a workaround already existed, not that the data did not.

A non-practitioner looks at a food safety audit and sees green across the board. A practitioner looks at the same audit and notices that every passing line traces back to one person’s signature. They ask who trained the person. They ask how old that person is. They ask who the successor is. The answer is: nobody. The audit passes because one person is passing it. When that person retires in eight months, the audit is at risk, and nobody has documented how.

What the practitioner did differently: read the audit as a human system, not a paperwork system.

If the discovery phase takes longer than the production cycle it is trying to understand, the discovery phase is the problem.

Every platform we deploy was shaped by someone on our team who had to live with its previous generation. If we have not worked inside a version of the problem we are solving, we do not build against it.

The data problems never go away. They get surfaced, named, and made smaller. Anyone who tells you otherwise is selling you the tool, not the outcome.

Time-and-materials without a measured outcome at the end is how advisory work ends up with a report that nobody reads. The commercial model has to force the work to produce something. If we are not willing to stake our fee on the outcome, we are not the right firm.

Leverage models that send unsupervised juniors into operational environments produce work that is expensive to fix later. Our staffing model keeps practitioner-led delivery on every engagement. That limits how fast we can scale. It also keeps the firm worth hiring.

Most operational problems are not technology problems. They are people problems wearing a technology costume. We say so. Even when the customer wanted us to confirm the opposite. Especially then.