Sainsbury’s chief data officer (CDO) Andy Day says a “farm to fork” approach attracted him to join the supermarket retailer in August 2016.
Day’s formative background lies in mobile telecoms, with a long period at Telefonica, and media, and the data strategy he has started to map out at Sainsbury’s has just drawn the accolade of number one spot on DataIQ’s list of the top 100 influencers in data for business in the UK.
“We collect data, ‘from farm to fork’, on customers and products. We have a lot bright people, and a lot of data, but the data has not been fit for purpose for driving business value,” he says.
His data team of 100 people, based in the company’s Holborn office in London, is being turned out to the business more and more, he says.
“I set out with a deliberately loosely structured plan. I was surprised at how good we were at data, but we have been less good at having a cohesive story to create output in the business,” says Day.
“I had never worked in retail before, so I spent my first 90 days listening. You’ve got two ears and one mouth, so you should use them in that proportion in business.”
Capitalise on data
The Sainsbury’s data team members are also distributed among lines of business, such as the online business, trading and ranging teams. Meanwhile, Day himself reports to the supermarket’s chief financial officer, Kevin O’Byrne, who sits on the retailer’s board.
When he was looking for a CDO role in the summer of 2016, Day says he saw a lot of briefs from companies that were saying, “Come and fix our data and analytics,” but without asking why.
Read more about data analytics in supermarket retail
- Tesco aims to save over €20m a year by using sophisticated analytical technology to ensure its chillers work at optimum temperature.
- How Sainsbury’s digital lab hackathon brings together the IT team and staff from the wider organisation to generate ideas to improve customer services.
- Wal-Mart’s data and analytics investment in Bangalore.
“At Sainsbury’s, we have started from the commercial opportunities and worked back from there. The businesses that have been successful, in the data arena, have been the ones where they think about the application of the data. The right question is: What’s the business problem?”
In Day’s view, organisations fail to capitalise on their data when they fail to recognise that driving outcomes is a business challenge, not a technology one. “Lots of money has been spent on beautiful technology solutions to problems that are, as yet, undefined.
“Our goals are to add revenue and reduce costs – getting the right ranges on the shelves, the right products to the right stores, and optimising our logistics. Can we predict the impact of weather changes better? Can we help farmers farm their fields more efficiently with data science?”
“You’ve got two ears and one mouth, so you should use them in that proportion in business”
Andy Day, Sainsbury’s
Day’s strategy is to focus on bringing customer and product data together in ways that “drive a raft of retail applications, such as in ranging and pricing”.
“There is a big integration piece to this. Our customer and product data has been in silos. For a few million pounds, as opposed to £20-50m, you can point the applications at the right data”, and build piece by piece from there, he says.
Data as a starting point
Sainsbury’s is a Teradata customer for data warehousing, and is also starting to stand up instances of Hadoop and looking at newer master data management tools.
On the business intelligence and analytics side – where Sainsbury’s has a tool set that includes Microstrategy, Microsoft Power BI and SAS – he has a strong interest in applying machine learning on top of business intelligence, “using technology to build data feature layers without human input”.
Of his 100 or so staff, about a dozen data scientists are looking at developments of that nature and using non-linear mathematical modelling on “tricky optimisation problems”.
The aim, for Sainsbury’s, of the work of the data team is to get the company’s business people to turn to data as a starting point.
“We are quite data-savvy, but the plan is for data to be less of an afterthought, in order to create more sales and profits,” says Day. “We want our people to be able to ask the right questions of data. Ultimately, the vision is for a small team to create the data and encourage self-service, alongside a smaller group of data scientists.”