DMAIC template: 5 questions that filter out the junk

Every Green Belt training I’ve sat through ends the same way: a slide deck, a worked example you’ll never look at again, and a “DMAIC template” handed out as a PDF or a single-tab Excel sheet. Then you go back to work, open the template, and within an afternoon you realise it’s study notes, not a tracker.

The free DMAIC templates on the first page of Google are the same. They look like a tracker because they’ve got the five phases as headings. They aren’t a tracker because the phases don’t talk to each other, the data doesn’t flow, and the Control phase (the one that actually decides whether your project counts as a win) is two text boxes asking you to “describe the handover plan”.

So before you download another one, here are the five questions I’d ask. Same five I asked when I built mine.

1. Does Define actually link to Improve?

A DMAIC template that works treats DMAIC as a chain, not five separate worksheets. The problem statement you write in Define has to be the same object that gets evaluated in Improve: same wording, same scope, same baseline number. If you have to retype it, it’ll drift. By the time you reach Control, your “improved” state is being compared to a baseline that’s already moved.

In a multi-database template (Notion is the easy place to do this), Define writes one row. Every other phase points at that row. Nothing gets retyped. Nothing drifts.

2. Where do hypotheses live?

This is the question that filters out 80% of templates immediately. Most have a “Root Cause Analysis” box in Analyse and that’s it. But a real DMAIC project generates 10-40 hypotheses across Measure and Analyse, most of which get tested and killed. The template needs somewhere to hold them, with status (open / testing / killed / confirmed), the experiment that tested them, and the data behind the decision.

If your template has no hypothesis register, you’ll either skip the step or run it in a side spreadsheet that disappears when the project ends. The next Green Belt who inherits the work has no trace of what you ruled out.

3. Does Control have artefacts, not text boxes?

Control is where DMAIC projects die. The template asks “describe your control plan” and you write three lines. Six months later the gain has decayed and nobody knows why. The template should make Control hard to skip:

• A Control Plan database (one row per parameter being controlled: measurement method, frequency, owner, response if out-of-spec)
• A Response Plan separate from the Control Plan (what happens when the alarm fires)
• A Handover record (who owns this on day 31, when does it get audited again)

If Control is two text boxes, the template doesn’t believe Control matters. Neither will you, six months in.

4. Is there a worked example you can pull apart?

Every DMAIC template I’ve ever built or rebuilt got better the moment I worked a real project through it. Cells you thought you needed, you don’t. Cells you didn’t think to add, you do. A template shipped with one fully worked example (every phase populated, the numbers tying out, the decisions traceable) is a template that’s been through that loop at least once.

A template that arrives empty is asking you to do the work that should have been done by the person who built it.

5. What happens to your worked example when the template updates?

This is the one nobody asks until they’ve been burned. You spend a week tailoring a template to your project. The author ships v2. Your work is on v1. The “lifetime updates” promise is worthless if the update path involves manually copying your data across.

A Notion-native template solves this: the schema updates, your data stays. An Excel template doesn’t, unless the author ships a migration sheet. Ask before you buy.

What “passing” looks like

A DMAIC template that passes all five is rare on Etsy and rarer on the free template sites. The pattern that works:

• Multi-database (linked), not flat tables
• A hypothesis register sitting between Measure and Analyse
• Control with three separate artefacts (Control Plan, Response Plan, Handover)
• One worked example shipped pre-populated
• A platform (Notion is the obvious one) where structure can change without breaking your data

That’s the spec I built the DMAIC Project Tracker to. 8 linked databases, hypothesis register, separated Control artefacts, a depot pick-rate worked example tracing £196k of waste, and a 6-sheet Excel companion for the auditors who still want a spreadsheet they can print.

If you already have a DMAIC template you like, run it through the five questions. If it passes, keep it. If it fails two or more, the project you’re about to start with it will fail at the same point every project before it has.

What to do next

Two paths depending on where you are:

• You’re about to run your first DMAIC project. The DMAIC Project Tracker on Etsy is £29 and ships with everything above. The worked example is the bit most people use first. Pull it apart, see how the data flows, then start your own.
• You’re running a continuous improvement programme, not a single project. The DMAIC template solves one project at a time. The £97 Lean Toolkit on Gumroad solves the surrounding cadence: wave planning, loss diagnosis, daily management, KPI cadence, and the DMAIC tracker as one tab. If you’ve got more than one CI project on the go at any time, that’s the level to operate at.
• You want the cadence laid out by week, not just the templates. Templates tell you what to fill in. They don’t tell you when. The Playbook is the 90-day implementation cadence that runs the toolkit: what to do which week, in what order, with what owner, what gets standardised at day 30, what handover artefact is due at day 90. Cadence Six tier two: £97 for the first 100 buyers, £197 evergreen after.

Either way, the same rule: a template is only as good as the next person who has to inherit it. Build for that person. Or buy something that already was.

---

Cadence Standard ships practical Notion + Excel templates for ops managers, PMO leads and transformation teams. The free 5S audit tool and the Cadence Six assessment are both good 15-minute on-ramps if you want to see what we’re about before paying for anything.