Ever-increasing Personalization

Posted on 26 November 2014 by

Let’s assume that your online business has a really sharp customer marketing and retention team. They are constantly dreaming up new offers and incentives to entice your customers to spend more, keep them happy and keep them coming back again and again. Let’s even say that the team is great at slicing and dicing your customer database to identify particular segments of customers worth targeting with particular offers. Overall, you are pleased with their performance as you watch your customer retention and lifetime values increase. What more can you expect?

Ever-increasing PersonalizationActually, you can expect a lot more! The fact is that even the smartest marketers are incapable of thinking in an unlimited number of dimensions. It’s simply too difficult to juggle the myriad variables involved in accurately measuring the effectiveness of every campaign, determining which campaigns are most effective for which customer personas, and identifying ever smaller and more focused customer target groups to get even higher campaign response.

In other words, there is a high cost of generalization when it comes to customer marketing. Until the marketing team has discovered the ideal matches between every campaign and every customer persona, they are leaving money on the table.

The solution is to implement an approach that systematically replaces relatively crude campaign-customer matches with ever more fine-tuned and personalized marketing efforts. This is where the big money is!

Systematically Optimize Your Customer Marketing

Dear marketer, coming up with great campaign ideas will always be your purview. The next evolution in maximizing your marketing effectiveness is for you to focus on sending the most effective campaigns to each and every customer.

The optimization cycle you should be working towards looks like this:

  1. Start with your best ideas for customer marketing campaigns.
  2. Select the customer personas that you think will most likely respond well to those campaigns (ideally, this will include segmentation based on predictive behavior modeling).
  3. Run the campaigns to the selected target groups using test and control groups.
  4. Measure the results of every campaign in monetary terms.
  5. Identify sub-personas hiding within the recipient audience according to how each one responded to the campaign (in monetary terms) and isolate the sub-personas for which the campaign was extremely effective, somewhat effective and not effective.
  6. Divide your original target group so that the top responders will continue to receive that campaign (with further tweaking and testing, as necessary), while creating (and testing) different campaigns for those sub-personas that did not respond as well.
  7. Go back to #3: measure, optimize, improve and repeat!

To read more about this process, and see an example, see our blog post, Divide and Conquer for Killer Marketing Campaigns.

This highly-effective approach to ever-increasing marketing personalization has proven to consistently and dramatically increase customer spend, customer retention and customer lifetime value.

Ever-increasing Personalization

Getting Practical

While straightforward in theory, the above process is one that does require some sophistication on the part of the marketer (or some good software to manage the process and perform all the required calculations). Here are some practical challenges that you will have to address to incorporate this process into your marketing:

  • Continuous dynamic customer segmentation – Customers are not static, and thus neither can your segmentation be. Clustering customers into small, homogenous segments (preferably using the latest predictive models!) needs to be done frequently, if not continuously. After all, every customer interaction may have a major impact on what persona best describes that customer.
  • Measure the financial uplift of every campaign – As mentioned above, it is necessary to run your campaigns using test and control groups so that you can calculate the actual financial impact of each campaign. In other words, you can only ascertain the true value of a campaign by determining how much more revenue you generated than you would have received anyway from the same group of customers had you never run the campaign.
  • Prevent simultaneous cross-marketing to targeted customers – In order for your uplift calculations to be reliable, you need to make sure that no customer receives more than one marketing campaign at the same time (or even during overlapping response measurement periods). The only way to accurately determine that a particular campaign led a customer to a particular action is if no other campaign influenced his behavior at the same time.
  • Identifying sub-personas within campaign audiences – You’ll need to analyze campaign response data (again, in monetary terms) in order to identify recipient personas for whom the campaign was extremely effective, somewhat effective and not effective. Once you’ve done this, you need to develop and test new campaigns with the various personas in order to improve response. You should iterate this process until you reach a large number of small groups, each of which demonstrates high response rates to the campaigns they receive.

A Quick Plug

If you’ve read this far, the thought might have crossed your mind that implementing this approach may be a little more than you can easily handle. Don’t give up just yet – technology is here to help!

Everything described above (besides coming up with the great campaign ideas!) is exactly what our Optimove retention automation platform does for customer marketers. This systematic approach to marketing optimization promotes dramatic ongoing improvements in campaign effectiveness, both by identifying those customers who best respond to existing campaigns and by isolating those customers to whom marketers need to develop more effective campaigns. Within months of starting to use Optimove to manage their customer marketing efforts, marketers typically reach the point that they are regularly and automatically sending highly-targeted campaigns to hundreds of distinct, finely-selected customer groups.

Segmentor

About the author

  • David

    Thanks Pini for the really insightful article! A follow up on how this campaign result can be used for campaign predictions/recommendations would be great.