More data. More tools. More business value? That equation does not balance automatically.
The technical ability to capture data is growing: In a survey published in 2025, 56 percent of manufacturing executives reported they expect the volume of data collected by their companies to at least triple by 2030.
This, however, brings another question increasingly into focus: as data volumes grow, does the ability to make the right decision at the right time grow with them?
This is one of the defining business challenges of industrial digitalization. Companies can monitor processes extensively and still respond too late. They can invest in sensors, platforms, and analytics tools while employees still have to piece together information under time pressure.
Christoph Dorigatti about more data and its limits.
As long as this gap remains, the business value will fall short of its potential. More data creates value only when it leads to better decisions and effective action.
Costs accumulate while we are still seeking clarity
Imagine a pharma production line, which continuously generates temperature readings, cycle times, and quality data. Suddenly a temperature reading rises above its usual range. At the same time, cycle times begin to change. The equipment keeps running. A quality defect has not yet been confirmed.
The shift supervisor has only a few minutes to decide: continue production, adjust the process, or stop the line? To do so, he needs an assessment of quality assurance and, if necessary, maintenance. While the team brings the information together, production continues.
A precautionary stop would cost production time and might require additional inspection and restart work. Intervening too late could risk scrap, rework, and delivery delays. The question is what the best next action can be. And even if the decision is postponed. The financial consequences are already taking shape.
At 100 units per minute, a twenty-minute stop initially means 2,000 fewer units produced. If the line continues running under faulty conditions for ten minutes, another 1,000 units may be affected. The actual cost depends, among other things, on whether rework is possible, lost production can be recovered, and delivery deadlines are at risk.
The shift supervisor must weigh these consequences while meeting quality and safety requirements. That is the operational responsibility of the team. Ensuring that the necessary information for the right decision is available in time and that responsibilities are clearly defined, however, is a management responsibility.
More data needs context to support decisions
The obvious answer seems clear: measure more, connect more, and analyze more. But additional data offers little help if its meaning in the situation at hand remains unclear.
A temperature fluctuation may be expected during equipment startup. During steady operation, the same pattern may indicate a fault. A reliable assessment therefore requires the reading to be considered alongside the product, operating state, and quality information. If employees first have to establish this context across different systems, it takes longer to act – valuable time is lost. Additional data can even increase the effort required to distinguish a meaningful deviation from normal process variation.
The first task of industrial digitalization is therefore to make relevant information available in context. On the production line, this means enabling the team to assess what a deviation means, which consequences it may have, and how urgently a response is needed.
Industrial AI follows the same logic. A prediction may identify a deviation earlier. But its usefulness depends on whether the team can interpret it in the specific operation situation.
The business question guides the choice of solution
Once the information needed for a decision is understood, the next question is what prevents the company from using it effectively. Starting with the choice of a platform, additional sensors, or an AI Model reverses this sequence.
The starting point must be the outcome that needs to improve. On the production line, for example, this means reducing quality losses or limiting avoidable downtime. This then identifies the decision that needs support: when and how to intervene if a deviation occurs.
The appropriate solution depends on the obstacle. Missing measurements may require additional sensors. Inaccessible quality information may require better integration. A delayed alert may require a different notification process. If responsibilities are unclear, the company must define who assesses the deviation, who authorizes an intervention, and who finally acts.
Technology should address the identified obstacle. An alert needs a designated recipient, and an assessment needs a clear path to action. A prediction creates business value when it enables an effective intervention in time.
Expected business impact determines investment priority
Identifying a suitable solution does not yet establish whether it should take priority over other investments. For management, this means comparing opportunities based on their expected business impact. How often does the decision occur? What financial consequences are at stake? How much can the proposed intervention realistically improve the outcome?
Frequent decisions with substantial avoidable losses may offer greater financial leverage than infrequent decisions with limited impact. That benefit must be weighed against implementation, integration, and operating costs, as well as the effort required to change established workflows.
Each investment case should specify the decision to improve, the expected benefit, the costs, and the person accountable for achieving the result. This provides a basis for prioritizing digital initiatives according to their contribution to the business.
Less downtime does not necessarily mean higher returns
A technically successful implementation is not yet a business success. An improved performance metric is not sufficient either.
Consider an illustrative case in which a digital application aims to increase line availability. Availability rises from 90 to 94 percent, but the share of units requiring scrap or rework also rises from 2 to 5 percent. The availability target has been achieved. Whether the business has benefited remains an open question: the value of additional saleable output must be assessed against the extra material, inspection, and rework costs.
Line availability is therefore the target metric; scrap and rework rates are accompanying measures that help reveal whether the improvement creates costs elsewhere. Conversely, an earlier stop may reduce availability while improving the financial result by preventing a larger quality loss
The criteria for success must be established before the investment. These include a documented baseline, an expected improvement, and the related measures needed to assess its overall effect.
Subsequent evaluation must account for changes in product mix, capacity utilization, and the process itself.
Lower scrap and rework costs can strengthen margins. Freed-up capacity can enable additional orders. More reliable deliveries can improve customer relationships. Whether these effects actually materialize must be verified in day-to-day operations.
This evaluation also provides knowledge for the next decision: Which indicators were useful? Which measures worked? Where does uncertainty remain?
The ability to build on this knowledge also depends on digital sovereignty: Can the company continue to access and use its operational data, understand the logic behind decisions supported by digital systems, and adapt those systems when requirements change? These capabilities help the company retain control over how decisions are made and improved. They should therefore be considered alongside expected benefits and costs in every digital investment.
Better decisions are a lasting business capability
Industrial digitalization changes management’s role. Providing technology and delegating implementation to individual departments is not enough. Business objectives, information needs, and operational responsibility must be brought together.
This applies to intervention on the production line as much as to maintenance planning, energy use, and capacity commitments. The same question arises in every case: Can we act in time using the available data, and does that action improve business performance?
Companies that understand their key decisions and the uncertainties associated with them can put operational data to targeted use. They can prioritize investments on a better-informed basis and build a sound foundation for Industrial AI. By then linking decisions to their outcomes, they can identify which information and measures actually make a difference.
This capability becomes more valuable as products, equipment, and requirements change. It helps companies assess new challenges faster, prioritize investments with greater confidence, and adapt their decision-making as the business evolves. That is how industrial digitalization builds lasting value: by helping organizations make better decisions over time.
Every digital investment should therefore start with a clear question: Which business decision are we trying to improve, and how will we know whether the change has made a difference? The answer should guide both the data we collect and how we use it, then be tested in day-to-day operations. Ultimately, better decisions are the promise of industrial digitalization. Better business results are the proof.
A blog article by Christoph Dorigatti.
