Starting with one process not ten, and which one to pick

5 days ago
3 min read

Introduction

The instinct on discovering a capability is to apply it everywhere. A business identifies eight processes that could be improved, starts on several, and finishes none, because each one requires attention that is now divided eight ways. The disruption is felt across the whole organisation and the benefit is felt nowhere.

Doing one thing completely produces a working process, a measured result, a team that has learned how to do this, and an argument for the next one. It is slower in ambition and faster in outcome. The real difficulty is not the discipline of choosing one; it is choosing the right one, because a first project that fails makes the second harder to fund.

1. Starting with one process not ten protects the attention you have

The scarce resource in a small business.

Every project needs someone to own the exceptions, correct the errors and answer questions. That person exists once, not eight times, and dividing them is what causes the failures. This is why the same advice applies to a business of five people and a business of fifty; the constraint scales with the ambition.

2. Pick something painful enough that people want it

Motivation is a criterion.

A process people already complain about will get cooperation, honest feedback and tolerance for early problems. Improving something nobody minds produces indifference regardless of how well it works. Ask the team which task they would remove first, and take the answer seriously.

3. Pick something with data already recorded

The precondition.

Where records are complete and consistent, the project is about the change; where they are not, it is a recording project first. For a first attempt, choose a process where this is not the obstacle.

4. Pick something measurable within a quarter

Fast feedback matters.

A process running weekly gives you twelve observations in three months. One running twice a year gives you nothing useful in that time, and a first project needs to produce a visible answer while attention lasts.

5. Avoid the customer-facing edge for a first attempt

Contain the failure.

Internal processes can absorb an error. A first project on quotes, invoices or customer communication puts the learning curve in front of the people paying you, and mistakes there are expensive.

6. Pick something the owner does not have to run

A test of the design.

If the improvement depends on the owner's continuing attention, it will not survive a busy month. The strongest first projects work without the person who introduced them.

7. Finish it before starting anything else

The discipline itself.

Finished means running in normal operation, documented, owned, and measured against the baseline. Moving on before that produces a collection of half-implemented changes and no gains. Write the definition of finished down before starting, because the temptation to move on arrives at about eighty per cent.

8. Publish the result honestly, including what went wrong

How the second project gets approved.

A credible account of what improved, what did not, and what it actually cost. Overstating the first result makes the second one harder, because the promised benefit never materialised in the accounts.

9. Choose the second one from what you learned

Not from the original list.

The first project will have revealed something about your data, your processes and your people that changes the priority. Rewriting the list at that point is the sign of a business learning rather than executing a plan.

Be careful about the process that is most obviously suited to automation but least important. Technical suitability and business value frequently point at different processes, and value should decide.

Conclusion

Do one process completely, because attention is the constraint and it does not divide.

Choose something people already complain about so you get cooperation, choose something where the data is already recorded, choose something that runs often enough to produce a result within a quarter, keep the first attempt away from customers, prefer a change that does not depend on the owner's attention, finish it properly before starting anything else, publish an honest account including the costs, and pick the second project from what the first one taught you.