Most improvement efforts start with good intentions and end quietly. The reason is usually not a lack of ideas; it is that nobody knows where the work starts and where it ends.
DMAIC is a five-step cycle that removes that ambiguity. It is the backbone Six Sigma uses to improve existing processes; but it can also be used on its own, without a Six Sigma programme — which is how we mostly use it.
For DMAIC to work, there has to be a working process in place. If the process doesn’t exist yet — a new plant, a new business unit, a first ERP implementation — the path to follow is DMADV.
Define Measure Analyse Improve Control
→ → → →
│ │
└────────── if deviation appears, cycle again ┘Define
We reduce the problem to a single measurable sentence, such as “Shipment confirmation is delayed by 2 days on average”. “Logistics isn’t working well” is not a problem statement.
In this step we also draw the scope: which process, which site, which period, who is affected. The most common mistake in ERP projects is keeping the scope huge — “the whole purchasing process”. That is where most never-ending efforts are born.
Measure
We turn the current state into numbers. In a company using an ERP this step is much easier than in one that doesn’t: order dates, approval times and error records are already in the system.
We pay attention to two things:
- Is the measurement reliable? If data is entered incompletely or inconsistently, what you measure is data quality, not the process. Then that becomes the first thing to improve.
- Is the baseline clear? To be able to say “we improved” later, you need to know in writing where you started. The route we use to choose what to measure: CTQ.
Analyse
Numbers show the problem, not its cause. In this step we drill down to the root cause; our preferred order is:
Fishbone → Pareto → Five whys
(causes) (priority) (root cause)First we surface the possible causes with the fishbone, weigh which one dominates with Pareto, and drill down to the root of the remaining few causes with the five whys.
Improve
Once the root cause is clear, the solution is often smaller than expected: making a field mandatory, removing an approval step, merging two screens, setting up an alert.
We prefer to try the change in a narrow area first. A solution that works in a single warehouse or a single product group shows its real cost before it spreads across the whole company.
Control
This is the step skipped most often, and the one that pays off most. An improvement that doesn’t stay in place isn’t an improvement.
This is where the ERP’s greatest strength comes in: you can embed the control in the system, not in someone’s memory.
- Mandatory fields and validation rules
- Approval flows and authorisation limits
- Alerts when a threshold is exceeded
- Regular reports and indicator tracking
When the control step isn’t done, the process slips back to its old state within a few months. It is the most common pattern we see in the field: a solution is found and applied, then quietly reverts because nobody is watching.
Why DMAIC fits ERP projects
The measurement already exists. An ERP is a measuring machine without realising it. Most of the data needed for the Measure step is waiting in the system; what is usually missing is the habit of looking at it.
Control can live in the system. In quality management a control plan is a document; in an ERP it is a configuration. Once you put the rule into the system, compliance is no longer optional.
Scope discipline is the weak spot of ERP projects. The “one sentence, clear scope” rule that the Define step enforces is valuable even on its own.
Where do teams stumble?
Three patterns we see in the field:
- Define is skipped and the conversation jumps straight to solutions. Result: the right software solving the wrong problem.
- Measure is done without a baseline. After the improvement, people say “it got better” but nobody can say by how much.
- Control is never done. The project is declared finished, and six months later the same meeting is held again.
Work that avoids these three delivers results even without knowing any statistics.
How we run it
We use DMAIC not as a heavy programme but as a working discipline. We run analysis and improvement work as part of Process Consulting and make the measure and control steps permanent on ErpwareBI. When what is needed is a screen or an automation, custom application development comes in.