First is the assumption that shoppers want personalization — this is not always true. Lots of times, shoppers want a better price or faster availability or a choice of variants. Second, algorithms are deterministic — like the toilet seat letter, they assume consistency of purchasing. While this may be true on some levels, it may not be true on individual products. I might always buy Coca-Cola or Pepsico products, but sometimes I want a cola, sometimes a root beer and, when I’m feeling particularly wild and crazy, an orange soda. AI makes no allowance for variety seeking or whims.

There are two different types of personalization that you are bringing up in this discussion. There is the personalization of offer and there is the customization of the product. Personalization of the offer requires the retailer to care about the customer and want to understand their buying behavior. This is going to require a cultural change. The example of the Amazon toilet seat is more about ad re-targeting than personalization. Amazon’s algorithm actually does a very good job of suggesting what a buyer might like based on past purchases and current searches that did not lead to a purchase. This AI system is improving as information is merged from the Alexa systems. Content providers like Spotify, Flipboard and Apple News do a good job of personalization and providing relevant product based on behavior on their platforms. But for retailers to get it right, they would need to put the customer first and put the systems in place to learn and understand what the customer really wants. Most retailers show the consumer what they want to… Read more »

If every brand had the same advantage as Spotify — a new data point on every customer every three minutes for two+ hours a day — I think they would find a way to execute in this space.

What Spotify inherently has and Vans has done is provide a platform for the consumer to self-identify preferences in a manner that doesn’t seem forced. They also both provide an incentive to do so: with Spotify I get better playlists and Vans gamifies their info journey by giving away points for questions answered. The key to good data is creating an environment where the consumer is actually excited to provide information. It not only creates a better CX, it actually improves the quality of the information provided by the consumer.

Data is not the problem. Not analyzing data and not listening to consumers is a problem. Without doing these things well, retailers give consumers choices that they think are personalization but are not choices that consumers want. Shortcuts to personalization do not work.

Technology is usually seen as solving a problem. That means by definition it is always focused on the past, on mending yesterday. That’s why so many of us are inundated with ads to buy what we’ve already bought. Telling me what I will want tomorrow is still a challenge unmet. We don’t know ourselves what we will want tomorrow — so we turn to algorithms? The best shot retail has at being sustainably relevant is to think, imagine, intuit, experiment, dream. Look beyond your usual sources to what else is going on around the world. Take footwear (since there was a picture of it in this piece). I remember going to the Bata Shoe Museum in Toronto 20 years ago and seeing sneakers from other countries that look similar to what we’re just now seeing in the U.S. Possibilities are metaphysical entities; forms of energy. Everything that ever was and ever will be exists in that way. They are all vibrating around you at this moment though many may think that to be nonsense. We… Read more »

Past purchase is no indication of future interest.

I’ve written about this in this blog post about how data and machine learning tell us less than we think.

Where the idea goes wrong is with this sentence: “The most successful companies understand that personalization is at its core a data problem.”

Data only reveals a very small portion about us. Something that lives up to the idea of “personal” is based on far more than what’s available in the data.

Only when we realize this can we pull back from pestering consumers with ads for things they’ve already bought.

When retailers blindly implement business rules that are driven by technology without data insights, then you get the kind of miss that is illustrated in the toilet seat example. Leveraging data to better understand customer shopping patterns in terms of product associations and timing will lead to personalization that is perceived by the customer as truly valuable.

Awareness of the potential to capture the relevant shopper information and make more appropriate offers is the challenge here. I still believe the majority of retailers and CPG brands are literally unaware of the technologies that are helping brands right now avoid the kind of misdirected offers highlighted in this article. One of the earliest applications of real artificial intelligence in retailing is the 1-800-Flowers.com module that uses personalization to dive deep into their cross–brand assortment to help increase the product movement of not only the promoted items, but also those items that are typically well below the radar of human call center agents.

The problem here is confusing “relevancy” with “personalization.” Consumers want relevancy. I have my own toilet seat example: a wine refrigerator. How many wine refrigerators does a person need? I think, but for a select few, the answer is one. So if I’ve bought one … I don’t need another. That’s not “personalizing” my communications, that is simply recognizing what would be relevant to a wine refrigerator *owner* vs. a wine refrigerator *shopper.* Personalization in the Vans sense is, I think, misleading. To me, that is mass customization. That is a challenge — a supply chain and order configuration challenge — but it’s not personalization in the sense of “knowing me” and catering to what you know. I think the biggest challenge for retailers and personalization is that they refuse to ask customers what they want or like. Instead, they lurk and sneak around and try to infer preferences based on behavior. Pinterest, thank heaven, has started asking: Do you like this ad? Do you want more like these? Do you not want more like… Read more »

Vans Custom program is a great way for customers to create their own color and image design for their shoes. It is a fun way for customers to create one-of-a-kind shoes that reflect their personality. While custom designed products are a unique way to personalize the shopping experience, it may not be a good fit for all product categories. Mass customization is coming! It has arrived in some segments of manufacturing and it is only a matter of time until it becomes a reality in retail. Personalized recommendations and services are another way to enhance the shopping experience for most any product category. What holds most retailers back from offering associate or customer-facing personalized services is access to real-time customer data and enterprise inventory. Combining real-time data and analytics with customer context enables retailers to personalize the shopping experience for each unique individual. Customer context, as defined by BRP, is the interrelated factors of customer insights and environmental conditions that make the shopping experience relevant. With advances in technology (networks, Wi-Fi, mobile, NFC, beacons, etc.)… Read more »

I agree with Nikki! Inferring individual human preferences based on behavior means nothing when it comes to “knowing me” or “knowing you” as an individual. Inferred behavior to recommend a car, a house or a dress? Certainly the brain trust of the world can go deeper than that! After all, the brain trust is composed of humans! We all get the inexplicable feeling of knowing when we love a product. Why we fall in love with things, the space between our thoughts — our unconscious and time-saving ability as humans to prefer one thing over another.

Knowledge of the customer is the first hurdle to personalization. Knowing something about customers who are like a customer — age, income, zip code, etc. — isn’t the same thing as personalized knowledge. A more realistic goal for most retailers would be something akin to mass customization — advertised as mass customization not personalization. You want personalized? Look at effective personal shopper programs. Now, those are personalized.