Interview: Minna Vakkilainen

Columbia Road: What was the starting point of your journey towards data centricity at Kesko?

Minna: In Finland, the large grocery chains are in the unique position of having a huge amount of customer loyalty data available. Out of 5.5 million citizens, Kesko has 3.3 million active loyalty card holders — that’s almost 80% of households that are regularly using our loyalty programme to buy groceries! Our data strategy is heavily focused on understanding our customers and their needs, and our data analytics isn’t only for Kesko, it’s also for our ecosystem of independent store owners and entrepreneurs — K-retailers, suppliers, partners and customers. We want to use our data to create synergies that benefit all of these groups.

When starting our roadmap seven years ago, the central premise was to enhance data-driven decision making to improve our food stores’ performance and help customers to maximise the value they got from our loyalty programme. The first phase was to find a better way to give access to and visualise the data in order to support our K-retailers’ everyday decisions. Because each store is different in our strategy, with their own store-specific business ideas, we then focused on presenting the data more holistically and dynamically, combining it to enable K-retailers to use insights about market share, potential, profitability and product categories in a way that would most benefit them — a big part of this was talking to individual K-retailers about their needs and challenges. What we’re doing now is moving towards supporting them to not only see the data but also carry out actions based on it, for example with store-specific product recommendations that help them make more customer-oriented and streamlined assortment decisions and then draw insights from that.

CR: Was this roadmap developed piece by piece or did you have a strategy in mind from the beginning?

Minna: When I started in this role our data assets were siloed — for example, the customer data was managed by an internal company and the loyalty sales data included tax and was not combined with other data sources, which made it less useful and harder to extract real value from. At that point, we had some ideas of what we wanted to do in the future but the immediate action was to tidy up the data and combine it into dashboards that would make it more available and usable. Only once this step was completed could we look into bringing in AI and analytics to support our decision making more holistically and justify this development to the K-retailers, Kesko’s own users and customers.

CR: That must have been a costly investment — how did you push for this internally?

Minna: As the loyalty programme was already in place, we started by looking at how much it was costing us in its current form and how much value our proposed improvements could bring. We had access to all this information about our customers and their preferences, so it made sense to use this valuable insight on the store level, and we also enriched the loyalty data with other data sources and aggregated the data insights on a chain level, which helped to show management that the information had value. They clearly agreed as they made this a core strategic asset for the company.

CR: Based on how you combined things like profitability and store-level data sets, are you able to suggest any best practices for companies that may be struggling as they start out on the same journey?

Minna: It’s still a challenging journey, even for us! My main recommendation is just to be proud and show the data you have as it’s never ready — there will be errors, like mismatched data between the spreadsheets and the dashboard, but making the data available for users to question is the best way to find mistakes. Having visualised the data, such as detailed information for our customers about nutritional values or the origin of their products purchased, our worry when giving the visualisations to them was that the data wasn’t perfect. But if you label it “beta” on the tool then at least it’s out there which is better than nothing. Customers understand that it's under development.

CR: How have you ensured that the planned tools and dashboards will really be taken into use?

Minna: You need to start with the challenge – which for us involves talking to K-retailers and employees to find out what individual challenges they face — then you can look at how data can help on an operational level. The next step is to ensure that you have an iterative process where you're always improving what you're doing through continuous testing and learning — there’s no point where we’ll say, “OK, now it’s ready”.

Service design is crucial in this process as it helps us to turn insights into easy-to-use tools, and the K-retailers have to buy into the process and be kept in the loop throughout; we won’t just create a tool and then push it out to them. Our stores have channels where they discuss these things, and some within those networks have more sway than others. Getting their backing helps to get more K-retailers involved right from the beginning, which allows us to get more data on how effective the tools are.

CR: How can these tools be used to improve store performance?

Minna: The stores can be measured on various KPIs like sales, sales development, profitability and customer satisfaction. We can educate K-retailers and help them address their performance by embracing the data and following our recommendations based on it. For example, if a K-retailer feels that the store isn’t meeting its potential, we can support them to improve their processes by helping them to use their dashboard more. The use of data and dashboards is a new way of working and leading for K-retailers.

CR: How do you cope with the responsibility of operating with this huge amount of customer data? Do you have any examples of that?

Minna: We have ethical principles for creating the AI and using customer data, and we’ve stated clearly that we want to offer value for our customers via data. If it’s not good for them then it’s not good for us and that’s something I communicate heavily in internal discussions, for example, if I’m asked why we’re not prioritising certain profitable products. It can be hard to be on the customers’ side and to find the right balance in these discussions, but it’s so important for us. I’m very happy that we have such strong data ethics and principles in use.

CR: You’ve been utilising AI in various areas to gain more tangible insights and impact from data. Do you have any recommendations for others who are starting their journey with machine learning and AI?

Minna: Stop talking about AI solutions and put more effort into understanding the problem. Also, don’t have a use case that’s too big — start small to make it more understandable. We don’t try to improve our product assortments on a company level — we start at the store level and then build trust so others see that it can be useful for them and approach us for help. I also want to highlight that it’s crucial for employers to trust AI-embedded solutions and tools. My recommendation is to always be as transparent as you can — meaning that you take the time to explain how these AI tools and algorithms work — and of course make sure that you can show the impact!

CR: What’s next?

Minna: Our journey has only just begun. We have a good situation with the dashboard where we’re combining the data, and we’re now moving from insight to action. We’re developing new smart and easy-to-use tools, including analytics and AI components, that help on an operational level – we have some good ones already but the potential for the future is huge, as our digital services, ecommerce, logistics and many other processes can utilise the same approach.

One area where we’ve improved a lot over time is transparency and cooperation within the company and with all the stakeholders — we’ve become much more open in our discussions and ways of working — there are still lots of silos but it’s much easier now as we all have the same data and knowledge, and we’ve already done a lot together. I always like to say, “if there’s a will, there’s a way”; we need to be ready to improve our skills and create more value every day — it’s an endless story.

---

MINNA VAKKILAINEN joined Kesko in 2014 to focus primarily on data analytics, and over time her role has evolved to also include responsibility for AI, machine learning, data development and the K-Plussa loyalty programme. Minna shares how she and her team have developed their tools and processes during her first seven years of leading the company towards customer data centricity. 

For more information about Kesko’s data advancements, take a look at their recent Data Balance Sheet report.