This year's NRF Big Show was a challenging one. Not just because of the explosion of vendors sitting in their booths, but because of what I heard from the show floor. The mass confusion among marketers when it comes to what technologies they need, what they can really do and how they will add any value is suffocating.
So, I'm back to continue my effort to clear up the cloud surrounding personalization and what's really possible (If you missed the first installment,, check out Debunking the 5 myths of personalization: Part 1).
Myth: Personalization is a “nice to have.”
Reality: Personalization is a “need to have.”
Customers today know that brands are collecting information about them, both online and off. With that comes the expectation of a better customer experience - customers believe you're negligent if you don’t execute with the knowledge you have. Your data is an asset.
Brands are coming around to this mentality. Here’s why: Ecommerce is going to increase by 10 percent annually through 2018, and it's going to grow even more quickly after that as the spending power of digital natives rises. The face-to-face, loyalty-producing interactions facilitated by in-store shopping experiences are becoming an endangered species. Smart brands are laying the foundations now for digital engagement.That engagement will be driven by personalization - personalization that, among other things, allows marketers to show customers and prospects content that is of genuine interest to them. Customers appreciate this. They enjoy it. And increasingly, they're demanding it. It can be the deciding factor between choosing one brand, or the competition. By working to meet these needs today, any brand can ready themselves for the continued shift of commerce from offline to online.
Take Alex and Ani as an example, they’re a specialty retailer that made headlines after posting 249.5% growth in online sales in 2013. They attribute that success to better, smarter use of data and keeping the customer at the center of every decision. Personalization is a key player in both of those areas. They’re collecting data from all channels, turning that data into an asset based on each individual customer and providing an experience that few companies expanding like they are can match.
The resulting experience evolves over the customer lifetime for each customer. Tracking back to Myth #1, they’re still expanding their use of personalization across the total customer experience, looking to mobile, in-store and consistently testing new approaches in email and on-site, where they’ve already deployed. Their digital team, led by Ryan Bonifacino, are true masters of modern marketing.
Compare that to brands like J.Crew that sends email with competing calls to action for men and women; catalogues that only feature women’s clothing to men; and even has to ask customers to tell them if they prefer men’s, women’s or children’s focused communications. This doesn’t happen with Alex and Ani (and if you think that men don’t shop there, you’re forgetting how many are there as gift givers, which Alex and Ani knows and uses to their advantage to keep them coming back).
Myth: Recommendations and predictions are the same thing.
Reality: Recommendations are based on the past. Predictions truly anticipate and action against the most likely future actions.
Recently, some in our industry have started to conflate the terms recommendation and prediction. Just because a provider says they have predictive capabilities does not mean they’ve actually built (or bought) the technology to do so; based on the marketing messages I see clearly some don’t even understand the difference themselves. It’s important for every marketer to have these terms down pat, if only to realize how to best combine them for your individual organization.
Most of us think of recommendations as what one person might like to buy or read, based on what other, similar people bought or read after having taken the same actions as the current visitor. Technology allows marketers to know certain things about a customer's recent engagement with a brand, and therefore, can reasonably recommend another product or page to that customer. Amazon was one of the first big merchants to do this widely; now the phrase, "You might also like…" is common across the web. Recommendations become even more effective when they’re specific to that individual user’s previous behaviors, rather than simply others who purchased similar items.
A predictive tool works differently. It's not using historical data to say that a user might buy or read one thing or another. Instead, predictive tools use algorithms to answer questions that get to the heart of what marketers care about. Questions such as: When will they buy next? How much will they spend? Will they read?
Many providers present these distinctly different concepts in slightly to incredibly deceptive ways – they’re essentially saying “we predict that if we recommend this product a user will be more likely to respond.” This is not an effective approach. Yes, in combination they are incredibly effective, but this is not right formula to follow.
The ideal use of these separate approaches is to identify users that are predicted to purchase and predicted to spend a specific amount and then to use those predictions to recommend products that fall within a range of that predicted spend. This way, the user sees relevant products and the marketer can optimize revenue generated. It’s an automated approach to supercharge your marketing ROI.
What’s key is that the predictive technology and personalization platform be one and the same. There are a number of solutions in the marketplace that are quite sophisticated at predicting at the user-level, but in order to make those predictions actionable in marketing, that data has to be exported and ingested.
So imagine you have 100,000 users that you’re running a model on and you want to engage based on the predictions. That means 100,000 API calls if you’re using separate solutions. That’s not realistic, so the resulting predictions end up being batched and shipped out in segments, creating a tremendous divide between what’s possible and what’s practical. Combining omnichannel data, automated personalization and predictions means you’re set to keep moving forward on a 1:1 basis.
Myth: Personalization only drives short-term conversion.
Reality: Personalization is the best method for optimizing lifetime value.
The confusion and efficacy divide is part of why personalization gets a bad rep as a short-term play. Brands use it as a campaign optimization tool, and ignore its omnichannel possibilities. Or maybe they only use it to allow a more fine-grained approach to segmentation (see Myth #2).
I look at personalization differently. To me, it's a mindset, which means that it is always long-term and it’s always about getting down to the individual. Our slogan is “every user is unique” - you need to believe that to win.
There are three information categories I recommend considering to get long-term brand value from personalization. At Sailthru, we fondly refer to these as “BUS”:
- Behavioral information: What are the customer's interests? What time of day or night does she tend to be most engaged? What platforms does she interact on? If she travels frequently to Boston on business, why are you only promoting stores in her home city of Chicago?
- Usage information: It's nice to know a customer is looking for shirts, for instance, but it's even better to know that she prefers peacock to navy and will sometimes substitute a 10 petite for a size 8. Maybe she won't buy at all without an offer for free shipping or a deep discount.
- Situational behavior: Why is a customer on your site? Has she been frustrated trying to find a similar item elsewhere, perhaps in your own stores? There is usually a contextual reason why someone does or does not purchase.
Individually, taking these factors into account may not impact a customer's purchasing habits overnight. But combined, over the long term and across many channels, they can be incredibly powerful to the lifetime value of that customer.
Our view: Personalization at scale is here, today. But to reap the benefits, you have to understand how today’s personalization is far more effective than its legacy predecessor.
Note from Neil: We'll be back next month with a fun/biting piece by Sailthru's Cassie Lancellotti-Young on the context behind marketing buzzwords. Meanwhile, you can read more of my views on marketing technology on my blog.
Image credit: myth and reality word cloud © Marek - Fotolia.com