Find Where the Work Is Leaking

Most operational problems are described in the wrong unit.

People say the CRM is a mess, or follow-up is inconsistent, or the team is stretched. Those are summaries of a feeling. They are not things you can fix, because they do not point anywhere specific.

The unit that is actually fixable is the handoff — the moment work passes from one person, or one system, to another. Nearly everything that goes wrong in a small operation goes wrong at a handoff, and handoffs are countable.

Here is how to find yours. It takes an afternoon and a notebook. No software, no consultant, and it is worth doing whether or not you ever change anything.

Step 1 — Pick one path, not the whole business

Choose a single journey and follow it end to end. The best one to start with is an inquiry from a new customer, from the moment it arrives to the moment the work is booked or lost — partly because it is short enough to hold in your head, and partly because it is the path where a failure costs you a customer rather than an afternoon. Do not try to map the whole business. Mapping the business is a project that never finishes and produces a diagram nobody looks at.

One path. One afternoon.

Step 2 — Write what actually happens, not what should

Sit with the person who does it and have them narrate a real recent example.

Not the policy. The thing that happened on Tuesday, with the customer whose name they can remember, including the part where they had to go and ask someone.

You are listening for phrases like usually, normally, if I remember, it depends who is in, and I just check. Every one of those marks a step that exists in someone's head rather than in the process. Write each one down verbatim, because the exact phrasing is the finding.

The output is a list of steps in order. Fifteen to thirty is typical, which is usually about three times what anyone expected.

That gap is the first finding, before any analysis. A process nobody could describe in under twenty steps is a process that has never been examined, and it is the reason the automation projects described in most automation fails because nobody designed the workflow first run aground so reliably.

Step 3 — Mark every handoff

Go down the list and put a mark against every point where the work changes hands. Person to person, person to system, system to system.

For each mark, answer three questions in writing:

Who is responsible at this point? A name, not a role and not a team. If two people could each reasonably assume the other has it, write both names and a question mark — that is a finding.

How does the next person know? They are told, they see it, a system notifies them, or they check periodically. If the answer is that they check periodically, note how often, note what happens overnight, and note what happens on the day that person is on a job from eight until four. Periodic checking is the mechanism that fails silently, because from the inside it feels like a process.

What happens if this step does not occur? Does anything anywhere register that it did not happen? This is the important one, and the answer is usually no.

Step 4 — Score each handoff on two axes

For every handoff, rate two things, roughly, on a scale of one to three.

Dependence on memory. Three means it only happens because a person remembers. One means the system carries it whether or not anyone is paying attention.

Cost when it fails. Three means a customer is lost or a job goes wrong. One means someone is mildly inconvenienced later.

Multiply.

Anything scoring six or above is where your problem is. There are usually two or three of them rather than fifteen, which is the most reassuring finding this exercise produces — the sense that everything is broken almost never survives contact with the list.

This is deliberately crude.

The precision does not matter, the ranking does, and a rough ranking done in an afternoon beats a precise one that never gets done — which is the usual fate of the thorough version.

Step 5 — Check the silent failures separately

Go back through your list and mark every step where a failure leaves no trace.

These are different from expensive failures, and they need their own pass, because they are the ones that never come up in a meeting. A missed call with no callback. An inquiry that arrived on someone's personal phone while they were on a job. A quote sent and never followed up because the reminder was in a head that got busy.

Nothing reports these. They do not appear in your numbers as losses; they appear as slightly weaker months, which get explained by something else. The clearest example is the phone — see the calls you never knew about.

Any step that is both high-scoring and silent is where to start. Not because it is the biggest problem — because it is the biggest problem you currently cannot see.

What to do with the result

Resist fixing all of it. The instinct after an exercise like this is to produce a transformation plan, and transformation plans are how this ends up costing a quarter and delivering nothing.

Take the single highest-scoring silent handoff and make it not depend on memory. That usually means one of three things: the work is routed to a named person automatically, a reminder exists outside somebody's head, or a record is created so the failure would at least be visible next month.

Then run the same path again in six weeks and see whether the score moved.

If the exercise produces a list of measurements you want to track rather than a step you want to change, what to measure in an inquiry intake process is the narrower version of the same work.

The reason this works without any technology is that the technology was never the hard part. The hard part is that the process has never been written down, so it cannot be examined, so it gets described in feelings instead.

An afternoon and a notebook converts the feeling into a list. Everything after that is comparatively easy.