Predictive analytics is gaining momentum and is rapidly becoming a part of every marketer’s lexicon. The allure is obvious: the term itself portends the ability to foretell the future; in our case, the ability to know what the response will be to your content based on comparative analysis of data. But “predictive” today is often not connected directly to action, except for a certain few vendors in the market.
Let’s look into our crystal ball and see what materializes about predictive analytics.
The features of predictive analytics assume the following capabilities (or “stack”):
- Robust data collection from multiple sources
- The ability to connect and combine these data sources
- The ability to display them in a meaningful way
- A marketer can make more accurate “predictions” about successful content
- Test that content to see if it matched the prediction
- An action layer in the predictive tool itself can automatically serve stored content to a particular user based on criteria living in a database and determined by the output of a predictive algorithm
As you can see, neither of these actually guarantees anything. So perhaps we should call it “pundit analytics” – but that might be giving pundits too much credit.
That said, the difference between the above two models is substantial.
In the first example: There is combination of data and then it remains up to an analyst, working with a marketer (these days the hybrid skill is rather famously called “data science”) to determine what the content should be, given the comparison metrics. For instance, if your analytics tool can combine data from geographic plus behavioral plus calendar information, you might be able to understand that your customer most likely comes from South Bend on a certain Saturday in September; and you would target her with offers that coincide with her behavioral and geographical patterns. And assuming your effort created an uplift in conversions, you’d then be able to take credit for predicting the outcome.
In the second example: The vendor offering itself contains a “decision engine” or “predictive layer” that automatically takes the same data your marketer would have reviewed and then automatically serves up that content to the South Bend user on a certain Saturday in September. And these companies will live and die by the measure of the uplift they achieve over a non-predictive alternative. That’s because it still isn’t inexpensive to engage a fully automated predictive engine – so it had better work!
Predictive analytics is one of the key features of a capability stack (and industry transformation) I’ve called convergence analytics: where customers are demanding the ability to track and take action upon multiple data streams; and where vendors are rapidly taking up the challenge to track multiple silos of data and perhaps even take action on them.
Convergence analytics assumes the following capabilities:
- Accurate data collection from multiple sources
- Combining the data intelligently and flexibly
- Enabling better connections with customers and prospects
The tea leaves in the bottom of my cup suggest that predictive analytics, a subset of overall convergence analytics, is due for a growth spurt as customers look for more consistent targeting in their hunt for ROI. What predictive analytics really does is act as a better targeting device. It weight-balances the bow so that the arrow flies more true. Sometimes the marketer will be holding the bow. In more advanced systems, the arrow is aimed and sent flying automatically.
The same technologies that power convergence analytics – cloud computing, big data, connectors, algorithms, display layers, and sometimes decision engines – also power predictive tools.
It’s not hard to imagine a growing number of vendors working to differentiate themselves by making claims to predictive capabilities. The challenge for vendors will be to make sure they’re not trying to claim a dashboard is a Ouija board; and if they do manage to get some magic into their algorithms, to allow for easy testing of the results.
Either way, it’s customers driving the pack. Nearly every marketer today is feeling the pressure to see more data at once and do more with it. Predictive is part of their future.
Prediction image on home page via Shutterstock.
Marketers create personas to better understand their target audience and what it looks like. If marketers can understand potential buyer behaviors, and where they spend their time online, then content can be targeted more effectively.
What’s behind a successful data-driven marketing strategy?
One of the major challenges in the martech industry is getting the attention of prospects in a world where they are bombarded by content and emails on all sides.
Facebook is addressing one of the biggest missing pieces of its chatbot offering: analytics.