Customer segmentation is the cornerstone of an effective marketing campaign. By grouping audiences according to behaviors, concrete attributes, purchase habits, and affinities, marketers can create campaigns that resonate with audiences and provide a strong return on ad spend. While the goal of customer segmentation is to create better customer experiences, boost ROI and reduce customer acquisition costs, many segmentation efforts fall short of this. The reality is that even small to medium-sized businesses can spend upwards of $100,000 per month on Facebook or Google ads, yet get underwhelming returns on that investment because of poor or no segmentation. Other channels also benefit greatly from improved customer segmentation, including display ads, email, and mobile, as well.
What’s the difference between good segmentation and bad segmentation? It often comes down to the data. Many brands rely on demographic data to drive their customer segments and campaign decisioning. This includes both first-party demographic data—the data that brands collect through their own interactions with customers—and third-party data purchased from a data reseller. Of these, first-party data is preferable, as it’s more reliable, fresher and less expensive.
Why behavioral data builds better segments
But the best data for building segments is behavioral data. This is the data that marketers can trust the most because it’s based on users’ actions and behaviors, enabling marketers to draw reliable conclusions from that data. For example, if you’re an online media company and you know that a specific user keeps visiting your renewal page but has not renewed, you can reasonably conclude that they’re thinking about renewing their subscription. That, or they really enjoy clicking the “renew” button.
Best practices for customer segmentation
Beyond data, there are best practices for customer segmentation that marketers should follow. Here are a few that stand out from my experience in helping customers achieve segmentation success:
Build a customer segmentation team.
Some of our customers have seen great results by creating a center of excellence (COE) for customer segmentation. This usually starts with a subset of the marketing team tasked with building and maintaining customer audiences. The team should include a technical stakeholder as well as marketers who will build segments for activations. Generally, the time commitment of team members is about five hours per week during CDP implementation, decreasing once the data streams and mappings have been set up for most use cases. A COE not only leads to a better segmentation strategy, but also has benefits down the road, as the new audiences being built can be synced with downstream marketing tools and used by other marketers for campaigns without requiring them to even touch the CDP.
Let’s say your marketing goal is to acquire new customers or reduce customer churn. Your audiences should align with those goals. For example, you might create a segment of customers who are likely to churn and set up rules to define that segment, such as “anyone who hasn’t clicked on an email in six weeks” or “a subscriber who hasn’t visited our site in two months.” Lytics CDP has easy-to-use rule builders to help with this, including access to explicit behaviorally-based rulesets, as well as implicit data-science-driven likely-to-churn (or likely-to-purchase) scores.
Don’t worry about perfection.
Getting your data into perfect shape for customer segmentation can actually be counterproductive. It’s more important that your data be prioritized and clearly labeled rather than perfectly clean, especially when you’re using a CDP. Otherwise, companies run the risk of going down the data warehouse rabbit hole again.
Mix it up by trying new segmentation tactics.
Experimentation is the key to getting the most value from your customer segments. If you have a hypothesis about a segment, test it. Measure the results along the way, and don’t be afraid to fail. If a segment performs poorly, then it’s an opportunity to apply your learnings to testing new segments. Also note that buying habits can change over time and segments should too.
One final thought to leave you with: using behavioral data to build segments doesn’t mean tossing out your old audiences. The objective should be to both create new audiences that you didn’t have before and improve the audiences you already have. Better customer segmentation is the result of better insights, and the best way to get those insights is with behavioral data and good data science.
If you’d like to learn more about how marketers can apply data science to their efforts, download our guide, Data Science for Marketing today.