3 Steps to Proper Segmentation
by Jason Haddock
posted on 11-29-2015
Is personalization a key part of your digital marketing initiative? What does that mean exactly? When somebody gives you a personalized gift or a personalized letter, it’s tailored specifically to you. Oftentimes digital marketers have a dream of doing the same on their websites, offering the perfect mix of message and product that resonates exactly with the visitor. This is possible, but not a realistic starting point for most organizations.
When structuring a personalization initiative, are you considering efficiency?
- Fact – targeting a segment reduces the size of the population
- Fact – there are unlimited possibilities that influence value
- Fact – every company has a finite level of time and resources
When beginning to plan your personalization strategy consider, how narrow do you want to go? Is narrow even better? Broadly speaking if your segment is under 10 percent of your population, it’s probably not worth your time or effort, unless that small portion provides significant revenue lifts to your business.
A good place to start is taking a broad approach and working your way to more focused segments of traffic, over time. A common mistake among digital marketers that hurts their efficiency is that they assume to know what slice of traffic they should be targeting towards. Often they jump to targeting the assumed segment instead of discovering the proper segment and thoroughly vetting it. Without first using the data, you may do a lot of setup work, only to find out that the slice of traffic doesn’t behave the way you thought. I suggest a three phased approach:
We are looking for behavior, exploitable behavior. We want to see how the general population acts and then compare that to different slices of that population to see if they act differently. Do they browse different pages? Do they add different items to their cart or download different white papers? Comparison is the key! A segment by itself doesn’t teach you anything, it needs to be contrasted with another segment. Firefox browser users may be a segment that behaves differently, but you won’t know unless you compare them to Safari users. Return visitors will act differently on your site but you can only see that when you compare to new visitors.
As you start broad and slice of big chunks of your population and include them in your optimization tests, you will start to learn things about them that you didn’t know (Limit your assumptions and let the data tell the real story!). Perhaps people from the west coast behave differently than people from the east coast. So you may want to create a segment based on geo-location, but after some analysis, you may now see that west coasters are high mobile users in comparison to our east coast friends. Now you can develop a scoring model related to purchase, and continue refinement. People who use mobile, who come from the west coast may have a higher propensity to buy product X. We further refine this segment by presenting them with product X messaging and creative, and watch to see if they respond in kind.
Now we target to them. Now we give them the messaging that resonates with that group and reap the rewards of a higher conversion rate. But the digital sand dunes are constantly shifting in the winds of change. We need to constantly vet and test these groups. Keep comparing that segment to contrasting segments. Keep throwing new creative and messaging at them to see if their attitudes are changing. Keep monitoring the efficiency and ROI of going after that segment. Make sure you are spinning your cycles and resources where it matters.
The good personalization programs put in the front-end leg work to really understand their visitors. If you take the time to figure out who your visitors are, you might be surprised.