« View all blog posts

Lesson #5 from a data scientist: Insight without action is just information


Actionable insights. If you’ve spent any time in marketing, you’ve heard those words from nearly every data and analytics company under the sun. Business intelligence companies sell more than $20 billion of technology that, according to their generic website copy, uncovers “actionable insights”. Hopefully, you’ve been following along in my blog and video series and have already read the blogs about not all data are created equal, using the RIGHT data, the differences between third-party data and first-party data, and most recently, a blog post explaining how the marketer remains in complete control using data science. Without this groundwork, we wouldn’t really be able to articulate what makes an insight actionable or perhaps a more appropriate word would be “activatable”.



What is an insight?

Simply put, INSIGHT = FACT + ACTION. While the equation is easy, it's actually a pretty rare outcome given that traditional business intelligence tools have a significant gap between discovery and doing something about it. On top of that, the individual components can be hard to come by. Not every fact will produce a great insight -- it must be both explanatory and novel, and a system that isn't focusing on signal is unlikely to produce either. Neither will every action produce a great insight -- it must be both contextual and direct, and the increasing complexity of today's marketing stacks can introduce its own hurdles to action.

What does an activatable insight look like?

So what does an activatable insight look like? It's information about your customers that is accompanied by a recommendation around what action to take to impact your goals. And what's even better than an activatable insight? One that automates that action so you can focus on more strategic and creative endeavors. But I digress.

For example, let’s say that I’m a marketer at a clothing and home goods retailer and I discover that people who buy belts frequently buy a pair of shoes within 24 hours. That’s an insight, but is it activatable? Not if I have to go through multiple steps before I can act. In most cases, I would need to open a channel tool such as my email system or push notification tool to target all the people who bought a belt recently. Then I would have to import that list of targets into a campaign tool that might create an online offer the next time they visit the site (since they’ll probably have left the site by the time I get this far) or send them an email with a special offer (again, hopefully within that 24-hour window).

It would be awesome if tools like Tableau just had a “Do It” button to take care of the action needed based on the insights, but they don’t. You’ve got to figure out what you need to do and then once you have an idea, do it yourself, and that takes time.

Now, let’s reimagine that scenario. Instead of having to switch tools once I make my belt-shoe discovery, I stay in the same environment and instantly create a list of customers who have purchased belts in the last 24 hours. As these customers browse my website, they’re greeted with an offer for a 10% discount on shoes. If they leave the site without buying shoes, I can send that segment an email within 6 hours letting them know that their 10% shoe discount will expire within 24 hours.

This scenario is exactly how our Lytics customer data platform works. From insight to action, you don’t have to leave Lytics. And this process is easy to do in Lytics, from uncovering the most important behavioral signals in your data to helping you quickly create targeted campaigns that get exceptional results.

Relevant experiences for the marketer

The other important thing we’ve done with Lytics is to make the experience relevant for the individual marketer. In my role, the belt-shoe connection might be a revelation. But not everyone is in my shoes. My marketing colleagues may be more interested in the affinity between sunglasses and suitcases. In Lytics, marketers can choose what they want to do with their insights, whether it’s cross-selling, up-selling, or just sharing a relevant piece of content based on their content interests, what we call product affinity, to increase customer engagement.

So, hopefully, the distinction between an actionable insight and an activatable insight is clearer now. If you can’t activate the insight at the moment of opportunity, it’s just information with the potential to be acted upon. FACT + ACTION = INSIGHT. What you really want are insights that can automatically be put into action and deliver relevant, real-time experiences to your customers. From where most marketers are standing today, that’s a pretty impressive feat.

Read the other posts in the series: