According to Forbes, most Enterprise companies (78%, to be precise) are planning to implement a Customer Data Platform.
If you’ve been keeping up with the latest marketing news, that probably comes as no surprise. Between new data privacy regulations in California and the EU and increasing customer demands for more and better personalization, more and more companies are getting serious about their customer data. And early adopters are enjoying incredible success.
So, if your company is like most, you’re probably already planning to add a CDP to your MarTech stack.
Which begs the question: Which CDP should you choose?
Once you’ve committed to harnessing your customer data, how do you choose the right platform? Which features are optional…and which are essential? What do you need in order to hit the ground running, identify your top customers, and start seeing real results?
Data science is the key
There are lots of opinions about what’s essential and what’s optional in a CDP, but here’s a foundational truth: CDPs with built-in data science perform better than CDPs without.
Let that sink in: 20 times as likely.
What kind of marketing results could you produce if you could identify users 20 times as likely to interact with your campaigns?
These results are why we vehemently disagree with anyone who calls data science an optional feature. It’s not a nice add-on; it’s the heart of a CDP.
Lytics’ data science and behavioral scores
So, what does effective data science look like in a CDP? What kind of behavioral scores and intelligence should you be looking for? Here’s what we at Lytics focus on and provide to our clients straight out of the box:
This measures whether your user’s engagement is increasing, decreasing, or staying the same.
For example, let’s say you have two users. Both are reading two articles per day. Their weekly active metrics look the same. They spent about the same amount of time on the site. They look similar. But if you look back at previous weeks, User A was averaging five articles per week and User B was averaging one.
They might both look the same this week, but past data now tells you that one user is more engaged than they used to be and the other is less engaged. Suddenly, their metrics look very different.
Momentum is how we measure these changes over time. Customers who get more engaged week by week have high momentum scores while customers whose engagement is dropping off get low scores.
Based on an analysis of all of your users, this metric identifies the average amount of engagement and lets you identify users who are above or below that average.
This measures how consistent a user is over time. How often do they interact with your brand? The more frequent, the higher their score.
This is a more robust way to look at the last time users were active on your site. Where most systems might give you a single date (User X was last active on October 12, 2018, for instance), recency gives you a weighted average based on the last three time stamps.
This means that people who accidentally clicked into an email after being inactive for a year won’t show up in a segment of your most active users, but users who consistently open their emails, click through to the site, or visit their profiles will.
This one’s simple: The more a user engages with your brand, the higher their quantity score.
Finally, to measure propensity, Lytics predictive models look at lifetime engagement and predict whether each individual consumer is likely to come back in the next 30 days.
Customer use cases for this data vary by business, industry, and goals. Some use the data to identify their most engaged customers and target lookalike audiences on platforms like Facebook. Others re-target customers whose engagement is dropping off. Still others combine these metrics with specific purchase data to drive add-on sales.
The truth is that the sky’s the limit; these are just your starting points. They’re just a few of the ways data science can take static data and use it to make our marketing smarter.
Can we predict acquisition costs and audience responses?
On top of the out-of-the-box features, Lytics offers the flexibility to build custom predictive models and return metrics like prediction accuracy, false positive rates, etc.
You can also look at the results of one campaign—identifying which customers purchased—and create lookalike audiences from that data for future campaigns. And once you create a predictive model, it can be set up to re-train itself and/or re-score users as needed so that your team can set it and forget it, moving onto other campaigns and priorities.
Learn how Lytics can support your marketing
If this all sounds exciting, we’d love to chat. Schedule a demo today, download our free CDP Buyers’ Guide, or contact us with questions. And if you’re already a client? Give your rep a call to find out how our services can help you take your data-driven marketing to the next level.