How is the ROI of an investment in production data calculated?

ROI stands for “return on investment”. The basic formula is: (benefit − cost) ÷ cost. If the calculation gives 0.6, that is an ROI of 60 percent. The measure answers how much benefit each euro invested produces.

It does not answer everything. When the benefit arises and how costs are spread over the years stays invisible. That is why ROI is usually combined with other measures.

The origin lies at DuPont. In 1914, Donaldson Brown, then Assistant Treasurer, developed a formula there that combined earnings, working capital and investment in plant into one measure, and called it “return on investment”. According to the Hagley Museum, DuPont made ROI its main performance measure and required an ROI forecast with every capital request.

The calculation is quick. Purely as an illustration: an investment of EUR 250,000 is expected to bring a benefit of EUR 400,000 over three years. The ROI is (400,000 − 250,000) ÷ 250,000 = 0.6, so 60 percent. The figures are a worked example, not a figure from experience. What matters is that every assumption behind it is named and can be checked.

With production data, the benefit can often be tied to plant measures. The overview lists typical building blocks on both sides of the calculation.

Four points decide whether an ROI calculation is robust.

• Baseline: Without a measured state before the investment, there is nothing to compare with. Improvements can only be proven against a documented basis.
• Attributability: The benefit has to be attributable to the measure. That takes data assigned to the same plants and periods.
• Full costs: Besides the purchase, introduction, training, operation and maintenance count. The guide IEC 60300-3-3 describes life cycle costing.
• Assumptions and period: Observation period, interest rate and forecasts should be documented so others can follow the calculation.

The sequence has proved its worth because it makes assumptions visible early.

1. Measure the baseline

   Capture availability, performance and quality of the affected plants over a representative period.

2. Formulate benefit hypotheses

   Define which measure should improve by how much through the intervention, and record the reasoning.

3. Capture costs over the lifetime

   Include introduction, operation and maintenance, not only the purchase.

4. Verify in a pilot

   Test the hypothesis on one plant or line, compare with the baseline and only then extrapolate.

Four distinctions prevent typical misjudgments.

• Not a guarantee: ROI is only as robust as its assumptions. A forecast remains a forecast.
• Not only cost reduction: Benefit can also come from higher quality, shorter lead times or lower risk, which are harder to quantify.
• Not the same as payback: ROI says nothing about when the costs flow back.
• Not a one-off calculation: Whoever does not re-measure ROI after introduction does not know whether the assumptions held.

The hurdles mostly lie in the data and the organization, rarely in the formula.

• No documented baseline before the investment
• Benefit that arises in a different department than the costs
• Incomplete costs, for example without training or operation
• Measurements whose quality nobody has checked

pronubes provides the data basis on which benefit can be demonstrated at all. The product parts take on different tasks.

• pronubes Edge: Edge captures data from machines and plants at the source and makes it available in a structured way. That creates the baseline without which no improvement can be proven.
• pronubes Zones: Zones maps sites, production areas, systems and data flows in a clear structure. This allows measures to be compared per site and area, and benefit to be attributed.
• pronubes Insights: Insights shows whether data flows run stably in live operation. That supports the statement that the measurements your ROI calculation rests on are reliable.