You can also learn transferable skills from top performers in other fields. If you’re in fashion retail, for example, personalizing your offerings for each shopper is a crucial capability. Netflix, which has made an art and a science out of personalization, are rock stars at this skill set. If you want to apply a “Netflix strategy” to your own site, there are some key lessons to be learned from their success.
Navigating the Paradox of Choice
One of the longtime core tenets of physical retail is abundance (we’ve all heard “pile it high and watch ‘em buy,” right?). It’s a well-established principle that works… up to a point. Abundance certainly brings customers in and makes them happy; the problem arises when retailers offer more choices than shoppers can comfortably process. A landmark 2000 study showed that while customers were attracted to displays with more choices, they were actually more likely to make a purchase when given fewer options.
Psychologist Barry Schwartz described this as “the paradox of choice” in a book and TED talk of the same name: shoppers want variety and lots of choices, but they’re less likely to buy (and less satisfied with their purchase) if you can’t find a way to narrow their choices to a more manageable level.
Managing that paradox — offering a big enough inventory to attract shoppers, but then converting them to buyers by intelligently narrowing the options — is the goal of online retail. That kind of personalization is something Netflix is very good at indeed. Let’s take a closer look at how they do it.
1. Prime the Pump
Every personalization strategy has to start somewhere. In Netflix’s case, new accounts and new profiles on existing accounts are always prompted to pick a few titles they like as a way of jump-starting the recommendation algorithms. The first screen you see will include movies and shows that are, in some way, connected to those preliminary choices you’ve made. It’s entirely optional for the subscriber, and many opt to skip it, but it’s worth providing the option.
On a fashion site, for example, you might present a banner or sidebar on your landing pages prompting first-time visitors with a “quick start” option. If it’s phrased in terms of speed and convenience (“New here? Get to the good stuff faster by telling us what you like!”), many shoppers will opt to click through and make a few selections. It’s a good starting point.
2. Use Social Proof
Whether you’re a new or established subscriber, when you first open your Netflix home screen you’ll see several categories showcasing what other people are watching. The names of the categories are all slightly different — “Popular on Netflix,” “Trending Now,” “Top 10 Today” — but they all leverage a psychological principle called social proof.
The basic concept of social proof can be expressed as “if other people are doing it, it must be safe/OK.” Remember those awkward middle school dances? It’s not easy to be the first one out on the floor, but it’s much less stressful when it’s already crowded. You could even think of it as the positive form of peer pressure.
Giving site visitors the option of clicking a selection of “Best Sellers,” “Top 10 in This Category” or “Most Popular This Spring” is a win-win: it helps ease the path to the initial purchase and it generates clicks that will guide you toward meaningful personalization.
3. Use Categories (Lots of Them)
Another stepping stone on Netflix’s road to meaningful personalization is its categories. How many of those do you think there are: ten? Twelve? A few dozen? The actual answer is over 200, though you might never see them all without a crib sheet.
Some of those categories are pretty universal, such as “Comedies,” “Action & Adventure” and “Children & Family.” Others are more niche, from “Greek Movies” to “Werewolf Horror Movies.” Again, these serve a couple of different functions. One is to create “act now” shortcuts that can be searched for, giving users a quick way to find something they like. A second is — again — to provide the clicks that will inform your personalization strategies.
Netflix tags its shows and movies internally, and the same movie can fit into a number of different categories. That helps the algorithm triangulate on your own preferences, ultimately leading to recommendations as niche-specific as “animated comedy-horror with supernatural elements.” That’s not a formal Netflix category but an inference drawn from your previous choices. In fashion terms, it might be “professional-looking shoes that are still kind of cute but are also super-comfortable for extended wear.”
4. The More Granular, the Better
That leads us to another crucial lesson: the more finely you can distinguish between your products, the more you can learn from each choice your visitor makes. In Netflix’s case its inventory at any given time is around 50,000 titles, which sounds like a lot — and it is — but only a relatively small number come and go in a given month. That means it’s possible for humans to manually tag new titles as they come along, and to do so in some detail.
In fashion it’s harder, partly because descriptions aren’t as clear-cut (though you could make a case for “Drama” or “Action & Adventure”) and partly because seasonal changes affect such a large part of your inventory, four times each year. New titles arrive on Netflix in a small stream; for fashion retailers it’s more of an onrushing tide.
Humans can only tag fashions at the most superficial level, because of the time and cost involved. To get really granular (and to really fuel your personalization strategies) it helps to rely on machine learning and AI.
5. Track and Measure Every Interaction
When and where do you watch Netflix? On a phone during your lunch break? On your laptop, in bed, on Saturday mornings? On your main TV, every night between 8 and 11? You can bet that Netflix keeps a close eye on which device you use at which time of day, what you watch and for how long. For example, if you use your lunch break to watch documentaries about chefs on your phone in 20-minute increments, you won’t see suggestions to binge The Vampire Diaries.
This exact interaction isn’t necessarily transferable, but the principle is. Every time your visitor (or repeat customer) interacts with your site, it’s a data-gathering opportunity. Knowing what they’ve searched for, what they’ve clicked, how long they’ve spent on a given page and which of your on-site tools (site search, categories, autosuggestions, etc.) they’ve used are all nearly as important, in their respective ways, as their actual product choices. These are inputs you’ll need in order to build deep personalization.
6. Be Transparent About What You’re Doing
One more crucial detail about Netflix personalization is that it’s pretty transparent. We’ve all seen a bar of suggestions with a header that says “Because you watched [x]…….” The connection between that first show and those in the suggestion bar may not be as obvious to you as it is to Netflix, but you’ll never wonder why it was offered. They’ve already told you so. The “Top Picks for [your profile name]” category works in much the same way, though it’s informed by your overall viewing pattern rather than a specific previous choice.
You may never have clicked through to it, but Netflix even offers a page in its Help Center that explicitly explains how its recommendation system works. That combination of explicit and implicit transparency is important: if your efforts at personalization come across as creepy rather than helpful, it’ll drive shoppers away. Helping your visitors understand why you offer what you offer sidesteps that pitfall.
7. Images are Powerful (Use Them Wisely)
London-based author and consultant Jennifer Clinehens, a specialist in the application of behavioral science to marketing and interface design, considers Netflix’s focus on science and psychology to be a fundamental part of the company’s personalization strategy: its secret sauce, if you will.
Aside from its intelligent use of large-scale principles like social proof, Clinehens points out fine details such as the company’s use of multiple thumbnail images, so it can serve personalized thumbnails to different users. For example, a diligent viewer of horror films might see a Stranger Things thumbnail emphasizing the show’s creepier aspects, while fans of nostalgic ’80s fare like The Goonies or Stand By Me might see a thumbnail of the show’s ensemble cast of kids.
Images are a powerful selling tool, one that fashion retailers could make better use of. If you have a dozen images of any given garment in your inventory, for example, it makes sense to test them against each other. You’ll quickly learn which images perform best for a given product category, or in a given region, or for individual customers. Once you know that, you can serve them the image most likely to attract their interest and approval.
Netflix Personalization: Data Aggregation Meets Psychology
The underlying principle behind Netflix’s personalization strategy isn’t simply data aggregation or record-keeping. It’s psychology: looking beyond the raw data to find the psychological reasons why you’ve made the choices you have. That’s what takes its recommendations from their initial superficiality (“you said you liked rom-coms, so here’s a row of rom-coms to scroll”) to the full screens of personalized selections seen by longtime users.
It’s also why Netflix’s “Because you watched…” suggestions are sometimes confounding to users. Because they tag their movies and shows in so many ways, their recommendation engine finds similarities that are not obvious to viewers but psychologically sound. As long ago as 2017, Netflix could boast that subscribers discovered over 80 percent of the shows they watched through recommendations from the site.
It’s perfectly possible for fashion retailers to become equally adept at leveraging psychology to drive your personalization strategy. You just need the right tool for the job.
Build It or Hire It: Netflix-Caliber Personalization for Your Site
If you want to bring your site’s personalization to a Netflix-like level, you have a couple of options. One is to invest Netflix-level quantities of time and resources into building your own. That’s not really a practical option for most companies. The other alternative is to source that level of personalization from a third-party. That’s where Lily AI comes in.
The key to intense, psychology-driven personalization is understanding the motivations that cause a visitor to make one choice over another. That requires deep tagging and very granular data). Lily AI’s proprietary algorithm uses machine learning, trained on the retail world’s largest data set, to do that for you.
Deep tagging can create scores, even hundreds, of tags per SKU. That in turn provides you with the rich level of insight you need in order to create psychologically sound, individually tailored recommendations to every shopper.
Contact us today to schedule a demonstration, and learn how Lily AI can support your personalization strategy.