The rule that comes before the steps
Digitalizing a bad process produces a bad process that is more expensive and faster. That is why this roadmap has a single emphasis: stabilize and slim down before connecting. A flow with high variation, a different method on every shift and stoppages with no known cause generates confusing data and wrong decisions.
There is also an economic reason. Every sensor, license and integration has an installation and maintenance cost. If the process changes after digitalization, part of that investment has to be redone. Stabilizing first protects the money and shortens the path to the first result.
The five steps follow this logic. The first two prepare the ground. The next three put technology where it actually adds value. The characteristics that define a connected factory are in 7 characteristics of Industry 4.0.
Step 1. Stabilize the process
Stabilizing means making the process behave predictably. Start with the basics: order and cleanliness with 5S, a method that is described and followed in standardized work, and maintenance up to date.
A quick test for step 1: watch the line for one hour and note how many times the operator had to improvise, look for a tool or wait for material. If the number is high, the cause lies in the process, and no system will fix it.
Sign that it is stable: the team can predict the shift's result within a reasonable margin and can explain the most common stoppages. If that is not yet true, the following steps should wait.
Step 2. Slim down the flow
With the process stable, look for waste. Draw the flow with value stream mapping and eliminate obvious waiting, transport and rework. Every step removed is a step that needs no sensor, system or software maintenance.
Illustrative scenario: a part goes through 6 steps, and 2 of them are transport and waiting. By eliminating them, 4 steps remain. The data collection points drop from 6 to 4, a reduction of 2 ÷ 6, or about 33%. Fewer steps also mean fewer sensors, fewer manual entries and fewer points of failure.
The practical rule: if an activity adds no value and can be eliminated, eliminate it. Only automate or digitalize what remains.
Step 3. Choose the indicators and the problem
Define the business problem that digitalization should solve and the few indicators that will prove the improvement. Examples: unplanned stoppages, scrap per shift, order throughput time. A good starting point is the OEE of a bottleneck machine.
Three criteria help in choosing the indicator: it should reflect a real loss, be measured the same way every time and lead to an action the team controls. Indicators that nobody can change merely decorate the dashboard.
Without a clear question, the project becomes data collection with no destination. Write the question and the indicator in one sentence, for example: reduce the minor stoppages on the filling line, measured by stoppage time per shift.
Step 4. Connect in a pilot
Choose one line or cell and install the minimum technology needed to measure the indicator: sensors, a data-entry terminal, a display visible at the workstation. Use the pilot to learn what works, what the operator actually uses and how much it costs to maintain.
Avoid the pilot that never ends. It needs an end date, an owner and a final decision among three paths: close, adjust or scale.
- Define the scope and duration of the pilot before starting.
- Train the people who operate it, and listen to their criticism.
- Compare the indicator before and after, using the same measurement criterion.
- Document what went wrong.
Step 5. Scale with method
If the pilot delivered the expected result, replicate what worked, line by line. Standardize the technical solution, the training and the routine for using the data. Use the PDCA cycle in each new round and include the upkeep of the solution in the budget.
Before each replication, repeat step 1 on the new line. The process that will receive the technology needs stability equivalent to the pilot's, or the result will not repeat.
Scaling is not copying the technology. It is copying the habit: the dashboard is reviewed at every start-of-shift meeting, and a deviation triggers an action with an owner and a deadline.
Summary of the five steps
| Step | Action | Control question |
|---|---|---|
| 1 | Stabilize the process | Is the shift's result predictable? |
| 2 | Slim down the flow | Have we removed the obvious waste? |
| 3 | Choose indicators and the problem | Which loss do we want to reduce and how will we measure it? |
| 4 | Connect in a pilot | Does the operator use the data day to day? |
| 5 | Scale with method | Was the habit replicated, and not just the equipment? |
Next step
Run a one-hour readiness test on a candidate line. Check whether there is a written standard method, whether stoppages are recorded, whether there is an indicator tracked every day and whether the area is organized. For each item not met, open a stabilization action and postpone the technology investment until at least three of the four are in order. When you are ready, the view of Industry 4.0 on the shop floor helps you choose the technology.
Frequently asked questions
Why stabilize before digitalizing?
Because technology amplifies what exists. An unstable process generates noisy data, and decisions based on it make the result worse.
How long does a pilot take?
It depends on the scope, but it should be short enough to learn quickly. Define the duration and the success criteria before starting.
Do I need a large system to start?
No. One sensor, a data-entry terminal and a visible display already let you measure an indicator and prove the value.
Who should take part in the project?
Operations, maintenance, quality, IT and, above all, the operators of the workstation that will be connected.
How do I keep the project from being abandoned?
Tie the data to a routine: a start-of-shift meeting, an action with an owner and a deadline, and periodic review of the results.
Keep studying
Sources
- ROTHER, M.; SHOOK, J. Aprendendo a Enxergar. São Paulo: Lean Institute Brasil.
- Lean Enterprise Institute. Value-stream mapping (glossário Lean). https://www.lean.org/lexicon-terms/value-stream-mapping/
- Lean Enterprise Institute. Standardized work (glossário Lean). https://www.lean.org/lexicon-terms/standardized-work/
- LIKER, J. O Modelo Toyota. Porto Alegre: Bookman.
Want a quick guide to keep on the shop floor?
Download the free Pocket Guide e-book with the lean tools summarized in one page each.
Download the e-book
About the author
Lean Manufacturing & Behavioral Management Specialist
Over 20 years in the automotive and metalworking industries (GM and Dana), Lean Manufacturing practitioner since 2006. SENAI instructor and mentor in Brazil’s Brasil Mais Produtivo program, delivering consulting, training and audits for 50+ companies, combining quality, productivity and people development.
