Some sites are stupid.
They don’t know you; they don’t know what you like; and they don’t know what you want. Even if you’re among the tiny six percent of visitors that log in, the site is the site is the site.
“Unless you put the $4 billion a year that Amazon puts into its technology, you end up with a pretty dumb site,” Joelle Kaufman, BloomReach’s head of marketing and partnerships, told me yesterday.
“We use technology to unlock that potential and make every web experience — mobile, tablet, desktop — oriented around the individual and their need at that moment.”
It doesn’t work how you might think it works.
BloomReach doesn’t track people around the internet via cookies or any other technology. It doesn’t ask you to log in and then survey you on your preferences, likes, or dislikes. And it doesn’t try to guesstimate your social-demographic characteristics and show you high-end mukluks because your IP address shows you to be coming from a tonier district of Anchorage, Alaska.
Instead, the company uses its machine learning technologies, which currently drive over one billion web interactions a day, to get to know you just like an old-fashioned proprietor in an old-fashioned store might.
“All that demographic data doesn’t correlate to your intent,” Kaufman says. “We care about what you do … for example, if you click on ‘what’s hot right now,’ immediately I know what you’re interested in, and I know that you’re interested in social media, and I know that you’re interested in what other people think.”
The company’s SaaS technology integrates into e-commerce and other sites and learns habits over time, fingerprinting users not by tracking them but by watching what they do and look for. That identification is good enough that without asking who you are or violating your privacy, BloomReach knows a particular visitor who comes on a desktop, then on a smartphone, and then on a tablet, Kaufman said.
But it’s not enough to identify visitors.
To be smart, a site also has to understand what you want, whether it be a strapless prom dress with sequins or a high-powered chopsaw. BloomReach does that by watching what you click, and knowing — as a person would — what the store’s inventory is. Click on strapless prom dresses, and you’ll see more of them, even if you return tomorrow from your tablet, not your phone.
Select for shinier, bedazzled versions, and BloomReach will showcase more options.
And then, the site itself stops being a static thing made of files and images, but a fully dynamic application. With its deep integration, BloomReach can change even site navigation, product filtering options, and create customized pages, right on the fly for each user. Even your site search and your search term autocompletion is personalized to what you’re interested in right now.
“We are a big data company with a whole lot of DNA in machine learning,” Kaufman says. “We’re a content-aware pattern recognition machine, and we are able to connect the non-logged-in user across all their devices.”
There’s a line the company doesn’t cross, however.
“We recognize you, but we don’t say ‘Hey John, we know it’s you, now on mobile.’ That’s creepy, and that’s wrong … but we do show you things more customized to you,” Kaufman told me.
Personalized experiences are what we have come to expect as normal and natural from technology, trained as we are by Amazon and Netflix. While big vendors like Amazon or Walmart build their own (or in Walmart’s case, buy their own) solutions, vendors such as Barilliance, IBM, Magiq, and Adobe all offer solutions to help retailers and others make their sites responsive to who users are and what they want.
BloomReach says, however, that no-one offers the “soup-to-nuts” entire solution that it offers.
Interestingly, although IBM offers its own site personalization technology, called IBM Personalized Product Recommendations, it has natively integrated BloomReach into its WebSphere Commerce platform as the “premium, upgraded site search, navigation, and personalization technology.” That integration will be available in a few months.
One thing is clear: The results of personalized done right can be shockingly good.
Sears implemented BloomReach’s technology late last year as the very first beta tester, just before Black Friday. The initial results were so promising that the retailer kept the technology functioning during the most important sales day of the year, and after.
“Sears’ revenue per query was 50 percent higher,” Kaufman says.
BloomReach is designed for enterprise and larger retailers, who pay a monthly subscription fee for a set number of “snap requests” to the BloomReach system. The minimum monthly price is $7,500.