It may be the case that Amazon knows me better than most of my family. It not only knows that I like to read a lot, but it also knows my favorite authors and many of the books I’ve read since 2008. It also knows that I ordered computer headsets three times in the past year and that I buy Avengers and video game merchandise on a regular basis. That said, Amazon sometimes gets it wrong.

Shortly after I purchased my last headset, I started seeing recommendations for headsets because recommendation engines like Amazon’s are still missing one important customer component: context.

Big data and retail: Understanding context

I had the opportunity to see how different brands handle the challenge of context at the Big Data & Analytics for Retail Summit held in Chicago last month. The summit brought together an impressive breadth of expertise where speakers delved into the details of predictive modeling and big data infrastructure. Others turned the discussion toward the needs of non-technical users, what data to use and how brands can use customer information in better ways.

Both sides of the analytics spectrum discussed the need for better insight into their customer and prospect behavior. This level of insight requires brands to look beyond purchase history, demographics and the traditional consumer segmentation. It requires understanding everything from purchase intent and how far away customers are from stores (geo-location) to how receptive they are to different types and frequencies of discounts.

Sri Subramaniam, Groupon’s Vice President of Relevance, spoke to this challenge using Groupon’s business model to illustrate the value of bringing together customer behavior and preference data with time-based offers and local information. As can be expected, there is also a lot of risk. The right offer at the wrong time can flop, and additional factors like willingness to travel can impact how well a specific deal performs. Groupon’s approach is to use a variety of user-based, deal-based and contextual attributes to predict how to best position a deal.

47% of US internet users age 18+ say they like online ads that are tailored to their interests

Mintel research shows that there is a healthy consumer appetite for personalization. According to the Internet Ads: Search, Display and Video US 2015 report, 47% of internet users age 18+ say they like online ads that are tailored to their interests. Consumers become even more receptive when it is tied directly to customer value; three-fourths of loyalty program participants say they like it when customer loyalty programs offer rewards based on their purchasing habits. But building a deep level of context-aware personalization is easier said than done.

Omnichannel: The challenge to better understanding customers

At the summit, Facebook’s head of Retail Measurement, Rick Malhotra, discussed the challenge of personalization-driving analytics initiatives, specifically customers researching products on a smartphone and making purchases with a different connected device or purchasing in-store at a later time.

Mintel research highlights the prevalence of omnichannel engagement, especially for retail. Between 2013 and 2015, preference for using mobile apps to manage loyalty accounts increased from 10% of loyalty participants to 22%. Preference for using retailers’ websites doubled in the same time frame. As Malhotra eludes to above, traditional tracking strategies that rely only on browser cookies are not effective in understanding consumers across so many different touch points.

Of course, Facebook has a unique advantage as an advertising platform. The widespread popularity of the network has led to its integration across many different channels, which gives it a wealth of unified data that would otherwise be unattainable. Other businesses address this challenge by leveraging data in more sophisticated ways. Customer loyalty programs drive many of these initiatives by providing a way of tracking customers regardless of how they engage with the brand.

Synchrony Financial CTO Gregory Simpson and SVP of Analytics Noel Ang refer to this as gaining a 360 degree view of the customer. In their presentation, Simpson and Ang shared how bringing together a range of data types fuels loyalty programs and helps to drive foot traffic in stores – including transactions, customer behavior, multichannel interactions and industry benchmarks. Synchrony has a unique perspective on the topic as it manages loyalty programs for brands by bringing together data as detailed as SKU-level information with broader details such as customer demographics and external factors such as weather.

This highlights one of the current trends in retail: Increasingly sophisticated analytics technology is creating unprecedented levels of specificity for business insights across all departments.

Personalization: Is too much a bad thing? 

During the Summit, I spoke with individuals from companies with two-member dedicated analytics teams and those who are part of forty-member teams. Finding talent and solving the technical challenges associated with analytics have always been key issues for brands. However, the platforms that fuel analytics insights are not only becoming more sophisticated, they are becoming more accessible.

The availability of analytics tools that anyone can use is, in many opinions, exciting, but it may also result in campaigns that come across as creepy to customers. When I spoke to Gregory Simpson from Synchrony after his presentation, I asked him to talk about one of the big challenges Synchrony is looking to solve, and he spoke to the issue of over-personalization:

“I think one of the big challenges is understanding how far to take personalization. There is a fine line between personalization and creepiness. When does it become intrusive? When is it helpful? We want to take that and help consumers see it as a great experience.”

It’s important to thoroughly consider the line between relevant personalization and an invasion of privacy, given that the immense volume of information companies have at their fingertips is also on the minds of consumers. Even with their high receptiveness to tailored rewards, for example, 61% of loyalty program participants expressed concern that retailers have too much information about them. It is an exciting time to be data-driven. However, as more of us gain access to sophisticated tools, it will be increasingly important to make sure we’re using consumer data responsibly, that our insights are truly valuable for consumers, as well as brands, and that we are careful about protecting that data.

Bryant Harland brings almost a decade of experience working in the tech arena, most recently as a Senior Technology Writer with Brafton News, where he oversaw the editorial team, wrote as a trade journalist and prepared a range of industry white papers.