Our industry seems obsessed with looking towards the future: what’s the next big thing, when will it be ready, and who’s about to disrupt the way things have been done up to now.

 

In fact, much of what we do as a company is based on predictive analytics – the science of trying to accurately determine how a shopper will behave in the future.

Yet sometimes it’s refreshing to take a look backwards in time, how it all began – the first ideas, what drove early innovation, how it’s developed over the years, and where it’s headed. In our case, that means zooming into a small niche within the world of supermarket retailing that we term “mass personalization”, or alternatively “personalization at scale”.

To give credit where credit is due, we need to begin our story in 1993, in the hyper competitive grocery market in Belgium, where leading supermarket chain Delhaize ushered in a new era within the retail world by launching the industry’s first loyalty program. It was designed to drive loyalty with their best customers by tracking their purchase data over time and rewarding them for meeting certain targets. This was a unique proposition in the industry at that time.

The trend quickly crossed the pond the UK, Tesco, who in 1994 were the second largest chain in the country, had a market share of approx. 18.3%. Their main rivals, first placed Sainsbury’s, commanded approximately 20% – their highest market share ever achieved. To challenge Sainsbury’s, the Tesco executive leadership had been planning the launch of a loyalty card (to be branded “Clubcard”) which in essence rewarded shoppers for spending at Tesco by giving them points for every pound spent over $5. The points would accumulate and then be translated in currency, sent to members quarterly in the form of vouchers that could be redeemed for products on a future visit.

Initially, Tesco’s marketing approach was similar to that of Delhaize, and revolved around saying “thank you for your loyalty” by giving customers some value back. But it soon became clear that the data being generated about individual shopper histories could be used to group them into segments based on their behavior, and to target each segment with offers that would not only drive frequency / loyalty, but also drive average basket size per trip. Offers were printed per segment and sent to Clubcard members via direct mail. And so, the discipline of “behavioral targeting” was born.

It’s hard to understate the impact of these developments by both Delhaize and Tesco, which undoubtedly and irreversibly changed the industry and its approach to marketing for both supermarket chains and brand owners globally. (Interesting to note that following the launch of this strategy, and due in in no small way to it, Tesco’s market share climbed steadily, reaching a peak of 31.1% by 2008).

While Tesco began using paper and spreadsheets to create the targeting files per customer segment, the industry graduated to the next level with the introduction of technologies using data mining tools which were up until that point considered the domain of much more technically advanced industries. Fast forward to the modern, data-driven era, in which our company (Sagarmatha) was founded with the vision of creating a purpose driven platform that would enable mass personalization for retailers. We did this by adapting a data mining platform initially developed for medical diagnostics.

This technological breakthrough enabled retailers to not only crunch the loyalty data faster, but also create segments of “one”, meaning that each individual shopper could receive a unique set of offers, compiled to drive behavior in sync with the retailers marketing objectives on the most granular level. This was the epitome of “personalization at scale”. Over the last few years we’ve been working with some of the most innovative loyalty marketing retailers around the world to continually develop the mathematical models behind the targeting, with the aim of offering each and every shopper the most relevant, up to date set of content at any point in time, on any and all communication channel that the shopper chooses to use: online, mobile app, in-store kiosks, direct mail and more.

At present, the algorithms and the computing power behind this hyper-targeted personalization platform represent the state of the art….for now.

And that brings us to where we see personalization going in the future.

Firstly, we see a move away from campaign-based targeting towards a real-time focus on each individual customer-journey, and how to impact that for each customer to the greatest degree, in real-time. Even though the frequency with which targeting cycles are currently run is already high (even daily), and regardless of the fact that accurately predicting future behavior is based largely on historical data going back 12-24 months, we have proved that major improvements can be derived by impacting customer journeys through scoring intent as it’s affected by real-time digital and physical interactions on an ongoing basis.

Secondly, AI (Artificial Intelligence) is being deployed to not only improve scoring, but also to enable the automation of offer recommendations. Currently offer pools are developed by marketers, a time-consuming process. Machine learning models currently under development will create a much wider variety of offers and offer types at high speed, including multiple price points per product, eventually leading to virtual personal pricing.

Thirdly, we ‘re planning for the integration of disparate 3rd party data sources, all of which hold clues to shopper purchase intent. Lifestyle and behavioral data, online and social media interactions, micro-location data, and other sources will be fused to create even more accurate pictures of each shoppers’ purchase intent and will further improve the retailer’s power to impact that.

Only 25 years young, the personalization industry, and the technology behind it has already changed so much and had such a profound impact on how we market (and shop).

We look forward to seeing what the next 25 years brings.

Chen Katz