Have Retail Analytics Crossed the Line from Cool to Creepy

A body of new solutions for retail businesses promises to both excite and upset consumers. These solutions deliver in-store (or near store) content to consumers that a vendor has decided might want to buy a product in the store.

The drivers are analytics software solutions that crunch data from a plethora of sources -- social, digital, point-of-sale and customer service -- in close to real time.

Consumers are organized into profiles or personas that can be scored by their propensity to buy a particular product at a point in time. Using these profiles, marketing professionals create and distribute personalized content that engages and, in theory, excites consumers enough to prompt a purchase. 

But could the technology just as easily agitate or annoy potential customers?

Location-Based Marketing

By using existing Wi-Fi and Bluetooth networks and mobile devices, companies can detect a potential customer when he or she walking into or near a store. The technology can then look at the consumer’s profile and push out appropriate content.

It might be a discount, a buy-one-get-one offer or a digital coupon, or a notice about interesting products that a friend has recently purchased. Any content that the vendor thinks is potentially engaging is possible, including video or audio ... even games.

Consumers, encountering the content, are encouraged to buy immediately rather than wait and think about it more, driving impulse buys. In-store content delivery systems may even level the playing field with internet-based retail vendors.

While the advantages are obvious, so is the potential downside.

Companies cannot ignore the risk that technology like this will annoy some customers. If customers feel they are being tracked like animals with radio collars, they will react negatively, feeling both insulted and disconcerted.

If they feel as if they are being manipulated to buy something they don’t want, they may become angry. Either way, the brand (and hence future sales) could be affected.

Avoiding the Creepy Line

Using in-store content delivery in a non-irritating manner is easy. A combination of software and social engineering can create an atmosphere that leaves consumers delighted.

  • Get permission. Always ask permission. Ask several times too. Ask when the data is gathered and ask again when the content is about to be offered. In this way the consumer doesn’t feel that the vendor is doing something they don’t want.
  • Be accurate. Sadly, the accuracy of many profiles is poor. Part of the reason is that it is hard to tell who a person is across channels. But accuracy doesn’t necessarily come with more data. Instead, it comes from a deep understanding of consumer segments that can then be encoded in a data model.
  • Be relevant. In-store content not only has to be broadly relevant to the consumer but also relevant at that moment. A coupon for caffeinated coffee might make sense at 8am but not at 11pm.
  • Don’t use stale data. Stale data is the enemy of accuracy and relevance. People change in real-time and so should the data.
  • Make content interesting. Boring content will not engage a consumer and may annoy them. It shows a lack of effort.
  • Be judicious. Too much content overwhelms and seems heavy handed. If the consumer doesn’t react to content than it’s not the right time or the content is poor. Bombarding them with bad or untimely content will only make them ignore future content.

The mechanics of pushing content to a mobile device is not the tough part. Creating a profile that accurately predicts an individual consumer’s need or desire at just the right moment is still very hard to do but getting easier. It may get even easier as cognitive computing makes it possible to detect subtle patterns in consumer behavior. In the meantime, existing technology can be combined with some common sense to create new opportunities for engaging with consumers and nudging them toward buying something.

Title image by zeitfaenger.at  (Flickr) via a CC BY-NC-SA 2.0 license.