How do you efficiently turn a sea of data into actions that improve your business?

Systems often operate independently from each other. When something goes wrong somewhere, it is not always clear what effects occur in other systems. Reports from different systems often don't clarify why something is going wrong, how to fix it and how important it is. In addition, not all feedback is necessarily wrong: an order delivery to the wrong address is not a system error. How do you get a grip on these processes and move from reports to continuous improvement?

Bringing data together

Every company has different processes spread across different systems. Each process often uses different data that is important for it: for example, for the web shop there are extensive product descriptions and reviews, for logistics there are dimensions and weights. Data is never perfect, so things go wrong all the time. Like products that are gathering dust in the warehouse but are not in the web shop, or products that turn out to be too long or too large to be included in the standard delivery.

The first step as an organization to prevent errors is to start working in a data-driven way. Bring data together through a platform and work on data quality in a structured way. For this, you can use a data warehouse or a data lake, but also, for example, Microsoft's more accessible Power BI. The specific solution with which you collect and process data does not matter much. More important is to be clear internally who is responsible for which data. 

The product owner

Who is responsible for the ins and outs of the web shop? And the website? Also, for the CMS? Before our business workshops I will ask these questions in advance, to make sure the correct people are invited. If it is unclear who is the owner of a system - the product owner - then it will also be unclear who can resolve any issues internally.

Therefore, it is important to appoint a product owner for each system. This person not only becomes the owner of the system, the associated processes and all data, but also for solving problems. This improvement will make it immediately clear internally who to contact if something goes wrong in that system. The next step is to ensure the product owner can resolve the reported issues. This requires practical information.

Task-oriented reports

Every system can run reports. The problem is that this creates a mountain of data and reports, and no one knows exactly what to do. Where do customers suffer? The business? What has priority? If you have to fill that in yourself as a product owner, it quickly feels like nothing really makes sense. What is often missing is a link to the operational objectives. You make that link with KPIs: key performance indicators that you use to steer from your role. You give priority to matters that influence them. Data about this from different sources is brought together into one report. In this way, your decision-making becomes not only data-driven but also goal-driven.

By automatically creating tasks in the report and providing them with a certain intelligence, you realize true data activation that can take repetitive work off your hands. Prioritize the issues in your report according to the set goals so that you, as product owner, can make the most impact immediately. For each issue, also add the necessary information, such as order and customer data, fully automatically so that picking up is much faster. Take an incorrect order as an example. What is the history of the order? What happened with previous orders? And what room do you have for giving a discount?

By structurally also improving the information in the reports, the deployment of Generative AI also comes into the picture, as my colleague Korneel explains here.

Communication

Working on separate islands with their own ways of working and goals is often not to the benefit of the customer. When a problem arises, they can be sent from pillar to post. The best solution to this is communication and coordination. Everyone agrees that the client must be helped quickly without too many referrals. How do you ensure that the right person is called in internally so that they can take the right action quickly? By getting together regularly from different departments and discussing the set goals, you gradually arrive at those structural improvements that make everything run better. Otherwise, you keep working past each other.

For example, a customer wanted a widget showing the top ten products on the web shop. The marketing department was thinking of the ten best-selling products, while the logistics department preferred to highlight the ten worst-selling products. The sales department, on the other hand, was looking at the margin. Through coordination, the choice was made to report on poorly selling products with a good margin, so that more attention from marketing is spent on them.

Keep improving step by step

No system is flawless, but every mistake counts. By taking a step-by-step look at what is happening and agreeing on what should happen, you can continuously work on structural data improvements. This is desirable for the customers' experience and for the employees' experience. Work becomes a lot more fun when you get the information and insights you need every morning to make a difference today. In the end, it remains people’s work.