Eijte Foekens – PLUS

Eijte Foekens is Senior Data Scientist at PLUS, a Dutch supermarket chain of the Sperwer Groep which is affiliated with the purchasing organisation Superunie. PLUS currently has 270 locations in the Netherlands. Eijte is a retail analytics expert in the ACE team, the Analytics Center of Excellence. Eijte brings a lot of experience to the table, having a PhD in analytics and working at several supermarket chains in the past in data and Analytics. Eijte: “Within PLUS, analytics is still at the beginning of its journey. Many people within the company don’t know which insights can be generated from all the data we collect. For the ACE team, it means we translate the wishes of the different portfolios to used cases and generate impact on the business. Data-driven decision making means consulting insights obtained from analytics, structurally, to reduce the risk around decisions.”

Main KPIs for growth

Eijte shares how many supermarket chains look at similar KPIs: “Revenue, margin and market share, whereby margin and market share do offer fields of tension. It is a displacement market; everyone wants a bigger piece of the pie. Market share is important, and you can only achieve that if you beat the market. In the last few years, PLUS is characterised by a positive market share development. In contrast to many Superunie members, PLUS gained 1/10 each year and last year it was even 2/10. The trick now is to hold on to that additional growth.” Talking about the ingredients for that growth, Eijte shares how PLUS is a cooperative with entrepreneurs that have their own local interpretation for their shop: “In addition to localisation, we discovered through covid that our e-commerce could be improved substantially. Entrepreneurs who hadn’t done much in the online space suddenly had to and found that it went extremely well. The largest share of that extra growth came from the further professionalisation of the e-commerce strategy, which would not have happened without covid.”

“There are many decisions for which you don’t know how things will turn out, but using insights can help reduce those risks.”

data-driven decisions

According to Eijte, analytics play a very important role: “To me, data-driven decision making means consulting insights obtained from analytics, structurally, to reduce the risk around decisions. Also, data-driven decision making demands an ongoing search for information innovation, i.e. improving existing dashboards with new or better KPIs. There are many decisions for which you never know exactly how things will turn out or how competitors will react, but by using insights in that decision you can exclude many risk factors. Analytics is about reducing risk as much as possible and taking the best decision.”

Eijte shares a few examples: “You have all kinds of buttons you can push in the retail marketing mix, an example is your current store layout. If you leave a store in its current state without doing a remodelling or refresh, the data might reveal that store performances have slowly deteriorated over time. So, a strategy that you can use is to refresh your existing stores from time to time. The interesting part however is to find out how to refresh your stores in a cost-effective way. Here consumer research is necessary to find out how consumers perceive the refresh changes.

The most frequent data-driven decisions can typically be found in price and promotion; you can adjust them today and you will impact tomorrow. Here, econometric models play a crucial role. In the field of assortment, there are many interesting dimensions. Why do stores differ in consumer behaviour? Different types of consumers live around the store, competitive fields wildly vary, but there can also be local brand or product preferences. If your headquarter has a one-size-fits-all approach, you will miss out on all these opportunities. You need to demonstrate these differences and show their potential before a retail organisation can be ready to take it on. Another way to deal with managerial reluctance to add managerial complexity is to conduct store tests. Positive test results will then lead to the rollout of an assortment decision to other stores. Typically, analytics first brings solid evidence to the table before management can act upon it. With data, you can also prioritise decisions that have the most potential. Within the ACE team, we work on the projects we think will have the most impact, and because we are at the beginning of the data and analytics journey, those projects mainly focus on sales, margin, and market share.”

fit for its user

PLUS uses marketing mix modelling to measure their marketing efforts, Eijte: “We measure and evaluate the effectiveness of our above and below the line communication, to see how the communication budget can be allocated in the best possible way. With consumer research, we measure everything from advertising campaigns to creatives to the magazine. Besides communication, we also look internally at the information sent from the head office to the stores. The trick is to keep that compressed: what is ‘need to know’, and what is ‘nice to know’? Do you make an entrepreneur happy if he can compare his store with other nearby stores in a dashboard? Or would it be more relevant to give him a dashboard that shows the most comparable stores within PLUS? Again, you need analytics for that, but you also need to talk to an entrepreneur. Can he benefit from an insight that 10% of the customers in his area buy organic products, compared to an average in the Netherlands of 20%? Is that commercially pointless, or promising? Data should fit the user, and so should the dashboard, so you need to check whether you are making the right comparisons. Experience has shown that a dashboard based on most comparable stores is better: if the dashboard shows that another store achieved 0.5% more turnover from organic products, then the entrepreneur immediately is triggered and wants to know why he is outperformed by his colleague. Franchisees and entrepreneurs are usually very critical, and that provides a very good test case for the information you provide from the head office. You have to have a good story that connects with the user/decision maker.”

staying critical

When it comes to analytics partners, Eijte tells us that balance is key: “Our vision is that you must do some things yourself because it is more cost-efficient, but you also need to be up to date on the newest developments. There are an awful lot of players with all kinds of services, so the trick is to pick the smartest ones out of the bunch. But we also do believe that some analytics should be outsourced. Suppose you are interested in models for the price elasticity of your products. PLUS does have a data lake in the cloud, but it’s not set up to do huge, heavy analytical exercises regularly. An external partner can set this up more cost-efficiently. However, you have to put someone from PLUS on the other side who can be critical and force the partner to keep improving their performance. How do you know that the quality of the models is good? You need to sit down, see the output and ask critical questions. You may conclude that some model specifications need to be changed or adapted, which may take a huge effort, but in the end, it will give you better insights. A non-critical view on a third party’s analytical outcomes is truly risky. Therefore, always take a critical look and use common sense to interpret insights.”