When I think of smart e-commerce or retail firms, Amazon.com is the first name that comes to mind. From robust reviews, to their Prime program, to one-click ordering, they have been ahead of their competition most of the time.
But, Amazon isn’t perfect. Their suggestion algorithm is sometimes off-target. I bought adult dog food, and they emailed me offering the same brand of puppy food a few days later. Wrong product, bad timing.
I also get reminded about items I’ve purchased and am unlikely to buy again. And, when they recently held their first “Prime Day,” a sale meant to rival Black Friday, their offers were widely mocked.
Gizmodo took Amazon to task for offering ridiculous items like a balaclava two-pack (for two-person bank robbery teams?) and a “beard growther” which claimed to accelerate beard growth. CNN.com quoted social media posts with statements describing Prime Day, among other things, as “a bunch of crap nobody wants.”
One of the best tweets about Prime Day came from Doug Davis:
— Doug Davis (@DJD) July 15, 2015
There’s the key point: as smart as Amazon is in knowing its customer’s interests and behavior, Prime Day conveyed the feeling of a neighborhood garage sale stocked with mostly irrelevant products.
Prime day was a fail for me. As an Amazon customer who uses the “Buy Now with 1-Click” button far too often, I found no deals enticing. The few that I might have acted on were sold out.
Before we move on, it’s worth pointing out that the other huge complaint about Prime Day was that items sold out too fast. So, despite the seemingly random targeting, Amazon often met its sales objective very quickly. Perhaps some of the criticism was undeserved, or at least premature.
A better way?
Imagine a different scenario for Prime Day… What if you woke up early, fired up your phone, opened Amazon’s app or website, and were presented with a shockingly on-target assortment of sale items? For example, sale items that were selected not just on your Amazon shopping and browsing history but information from a variety of sources? No Flintstones vitamins, no ski masks… But, perhaps, an unusual wine decanter because they knew you subscribed to Wine Spectator a month earlier? Or maybe a gift item appropriate for a family member with an upcoming birthday?
The day of ultra-personalization may not be far off.
I recently spoke with the founders of Cognistx, a Pittsburgh-based startup that is attempting to create a whole new level of personalized shopping and advertising. Not only will Cognistx take into account your history with a particular retailer, but it will incorporate information from other sources and even data like your current geographic location.
One of the founders is Eric Nyberg, PhD, a professor in the Language Technologies Institute at Carnegie Mellon University. Eric directs the Master’s Program in Computational Data Science. He was very involved in CMU’s partnership with IBM in the development of Watson™ that ultimately triumphed over human competitors in the Jeopardy! Challenge.
Nyberg’s insights into language processing and data manipulation are a key part of the startup’s planned offering.
Cognistx CEO founder, Sanjay Chopra, talked about the ability to do a much better job of customer targeting:
Ideally we are going to get away from a world where a marketing team decides on offers for large segments of the customer demographics, and instead looks at ways of actually coming up with very personalized offers. These offers really depend on who you are, and even what you’ve been doing in the last day or 15 minutes.
Company president Jeffrey Battin talked about natural language. For example, to get a little more information about a customer a friendly avatar with a few simple questions would almost certainly have a higher completion rate than a popup form.
Battin thinks there is a lot of value in “smart” surveys conducted immediately on a mobile device in a retail environment:
Wouldn’t it be great to understand and have immediate context as to every single user’s sentiment and be able to measure that to the store level, to the manager level, to the customer level? We have a customer that is extremely valuable, and, if we’re seeing scores drop off, we can be attentive to their needs. At the same time, if customers are scoring high, why not take that opportunity and instantly offer them an incentive to share the good news with their friends?
One danger of very precise targeting and communication, of course, is the “creepiness” factor. If a merchant knows something about me that I don’t expect them to know, I might be delighted. Or, I might feel like my privacy had been compromised.
To see how clever becomes creepy, we need only recall what happened when Target algorithmically determined which of its female customers were pregnant and targeted them with special mailing. Forbes writer Kashmir Hill describes how things played out in How Target Figured Out A Teen Girl Was Pregnant Before Her Father Did.
Target ended up dialing back the specificity of its offers. Sending a newly pregnant woman a flyer full of pregnancy-related products might freak her out. Including a few relevant items as part of a larger sale would cause no backlash.
Cognistx intends to address promos that are overly creepy or simply unwanted by allowing consumers to immediately indicate a lack of interest in that kind of offer.
With my interest in neuromarketing, I was compelled to ask the Cognistx team whether they were considering incorporating biometric data from devices like a Fitbit or Apple Watch. Chopra indicated that they were looking at the best way to use this kind of data. He said it could be used not only as part of a recommendation, but also to suppress recommendations that might be less appropriate.
(Chopra didn’t provide an example, but I visualize someone walking by a Starbucks and not getting the offer for a triple espresso because their pulse rate was already too high…)
Cognistx hopes to be fully live in Fall, 2015.
What do you think? Is this a potential for Orwellian intrusion, or a chance for consumers to get only ads and offers that are truly relevant in the full context of history, external data, current location, recent activity, etc.?
Roger Dooley is the author of Brainfluence: 100 Ways to Persuade and Convince Consumers with Neuromarketing (Wiley, 2011). Find Roger on Twitter as @rogerdooley and at his website, Neuromarketing.
This article was written by Roger Dooley from Forbes and was legally licensed through the NewsCred publisher network.