Palo Alto – August 17, 2020 Collecting clickstream data to improve a company website is nothing new.
Every company hoping to compete in the digital age understands that without data, it’s impossible to understand what’s important to consumers on the web. You need this data to report patterns, trends, and pathways to a given page, product, or purchase. Since the early days of the web and eCommerce, businesses have used A/B testing and other forms of analytics to examine the user’s journey, determine different layouts, and optimize their websites.
And while traditional data collection is crucial if you want to provide recommendations that appeal to your customers and keep them coming back, you shouldn’t stop there.
The top consumer companies are leveraging data science to harness and leverage what’s referred to as “data exhaust.” Put simply, there’s all this data customers leave behind when they interact with a digital property. And that data provides nuggets of information about their preferences that can make a site even more competitive and better able to address its customers’ needs.
If you’re not looking at that leftover data, you’re leaving a ton of potentially useful customer information on the table. Here’s how to make the most of the exhaust:
Don’t just focus on what users are doing — look at what they’re not doing.
Whereas traditional site optimization tools track the user journey — what someone clicked and selected — data exhaust looks also at what they didn’t do. The pages, items, or articles they were presented and didn’t select.
It’s almost like a forensic crime scene. You — the data detective — have to survey the entire room, look at all the evidence, and then figure out what was touched and what wasn’t.
Or, imagine you’re at a restaurant. The waiter takes your order and the chef sees that you ordered the steak, the salad, and the souffle. What the chef doesn’t know is that the waiter also suggested the lobster bisque, and that you replied you’re allergic to shellfish. Then (and hopefully this never happens to you) as the chef is making your dinner, he puts a lump of crab meat on the steak just to make it extra fancy. He thinks he’s doing you a favor but he completely ruins the meal.
In other words, sometimes the pathway to the order can be as important as what was eventually selected.
In the world of data, if you’re only looking at the most obvious paths and optimizing around those, you may lose other things that can tell you something just as important. But when you see the whole universe of things users were shown and considered, it’s a huge puzzle piece to help you figure out what motivates them to click.
Instrument your site for real-time data collection — preferably before any major redesign.
Once you’re committed to leaving no data unturned, you have to find a way to harness it and make it actually valuable to your company.
To do this, you should first instrument your website. This is often an afterthought because developers are so focused on getting the site up and running first. We often see this when we talk to customers about instrumenting our type of solution — they want to wait until the site is new and refreshed, and make it part of the big site redesign. But you shouldn’t put it off. You can and should do it before a site refresh. That way you have a baseline or control group to compare the refreshed site to.
But the important thing is to do it as early as possible.
Don’t let the data sit idle.
Instrumenting your site and apps is a crucial starting point, but it won’t do much good unless you’re putting the data to work.
The best way to go about it is to glean information that is going to be most useful to the user in the moment. Today, real-time recommendations are separating the most successful brands from those that are struggling to get attention. AI and machine learning are well-suited for this type of task — because it tends to be repetitive and has to be done frequently as new data comes in. Although intent and desire are very human emotions, ironically, machines can be better at figuring it out for us — especially online. In a brick-and-mortar store, the customer can tell you what they want. But online, it’s much harder to figure out unless you have some real-time analytics.
For example, several LiftIgniter customers now use our AI-engine to power their virtual assistants to mimic the experience from a live service agent. So, if we detect signals that they are currently interested in a specific category of product or content, we will direct the site to dynamically adjust to it — it’s like having an endless virtual showroom right in front of you as you click.
All the best companies are taking advantage of all of their data. If you aren’t, you’ll never be able to grasp consumers’ attention in the age of digital noise.
While conventional wisdom says to track the user pathways and throw away the other stuff, the antithesis is almost as important. They didn’t select the lobster bisque? That can tell you a lot.