Try to google A/B test your personalized marketing messages and what you get are three kinds of results–first, on A/B testing your emails (not all kinds of marketing messages), second, personalizing your marketing messages and third, tools that claim to help you do the previous two. Unfortunately, you do not find anything on the confluence of these two marketing strategies. Now why is that?

When both these tactics are so important for CRO then, hasn’t anyone thought of carrying them out together and fetch even better results? Let me tell you–a lot of people have tried them together, but failed. Why? Because these two concepts are fundamentally and foundationally different.

Why Split Testing And Personalization Cannot Go Hand-in-hand

If you think of A/B testing, touted as the #1 CRO tactic, you need a great volume of traffic on your website to perform this test and get accurate results, because this test speaks on behalf of the complete traffic that is coming to your website. For example, if you want to test two copies of an email on at least 1000 subscribers, then you need a database of at least 3000 subscribers so that the winning copy can be shown to the remaining 2000 people.

A simple representation of split testing for URL (source:

Take a close look at this–you are sending the same copy of email to 2000 users, which readily qualifies this email as a mass mail. Now, even though your copy is well-tested and tried, it will fail to resonate with approximately 5% of the email receivers. And here walks in personalized marketing messages. Let’s talk a more about them.

Personalized marketing messages are a product of advanced segmentation of users based on not just their demographics and location, but also their online behavior, activity on websites other than yours and device hopping. So your segments are now micro-segments with little scope of split testing and then sending the winning message to the remaining.

In short, A/B testing uses a large sample size which is representative of the target audience, whereas personalization is not. Plus, the insights gathered from the personalization efforts cannot be extrapolated to the whole traffic and scaled.

So most marketers are like: How am I going to do it? Send him the same copy at two different times and see when he responds better? Or send him two different copies of the same proposition and see which one he clicks? Assuming I do that too, then what am I going to do the next time?

One cannot continue to test all the time, or cannot afford to make generalized assumptions about each one of your users. None of these is feasible for any ecommerce. Can’t deny–the marketers are right about this!

So how to go about it? How to A/B test your personalized marketing messages?

Frankly, you cannot go about it!


As the image shows, personalization is the tactic to lift conversions and generate more customer lifetime value. A/B testing, on the other hand, is a tool to verify whether something will work or not. Still, one can design campaigns on the confluence of these two and that combination is called Dynamic messaging.

Combining Power of Personalization With A/B Testing

What you can do is A/B test your users with a goal in mind and then, personalize the messages sent to them based on their individual profiles. Another way of doing this is creating personalization scenarios for different user segments and crafting messages best suited to them.

For instance, you can design a message for a person who used your filter, checked out one product, but then left without performing any further actions. Such a person (or group of such people) can get notifications (could be in the form of emails or in-app messages) that would take him directly to his filtered results to continue his buying journey.

The Five Step Approach

Step 1: Create micro-segments (not segments)

Segment your current user database on various parameters such as, purchasing behavior, traffic sources, geolocation, device type, and everything that you can think of, not leaving the demographics out.

Step 2: Craft hyper-personalized messages

Create different sets of messages for each segment based on their position in your sales funnel, their past purchase behavior, their predicted online activity and events that take place on your website. Do not forget to craft messages for cross-device conversion optimization.

Step 3: Send and A/B test them (for real)

Send your messages as per your plan to your micro-segments and analyze the conversion results. Pick the winning messages for future while identifying the kind of users that do not show any signs of conversion.

personalized user profiles

Step 4: This is the game changer!

You now have your tool-belt ready to hit all kinds of customers. All you need to do is use Google’s customer match tool to identify similar audience. Why you need to do this is, because the previous step exhausts your micro-segments and so you are not left with any similar customers to send those winning marketing messages to. Hence, you need to find a new bunch, but one that is identical to the audience you tested on.


Step 5: Personalize winning messages

Now that you have your winning messages and your target audience ready, start sending messages. However, before sending any do not forget to personalize these messages based on the individual details of every user such as his name, age, gender, geolocation, need, etc.


And voila! You did it. With a little help from technology you have combined two seemingly uncombinable water-and-oil like concepts–A/B testing and personalization. Now go on and try to use this approach to get the best conversions in 2016.

P.S: Do not forget to share your results with us!