Slice Your Ecommerce Site into Cake Wedges For AI

It’s the dawn of AI and that’s here to stay for eCommerce.

“One Size Fits All Split Test Just Does Not Cut It Anymore”

What this means for us marketers is that the one size fits all split test just does not cut it anymore.

Look at the tests run over the past five years on sites like which test won and you will see big AB split tests mostly across all traffic.

e.g. split testing two different category header images is now dead. We need to go deeper.

How?

“We Need To Think Of Our eCommerce Site As Not One Site”

With artificial intelligence adding to the equation we need to think of our eCommerce site as not one site, but many sites for all the segments of users coming to it.

Given that, for now at least, the magic number of conversions per month needed for an AI engine to optimise on Facebook and Adwords seems to be around 50. We need to slice the website into slices of at least 50 conversions a month.

This might look like the following segments:

1) Existing customers who bought within the last 6 months
2) Existing customers who bought within the last 12 months
3) New Visitors
4) Repeat Visitors within the past 30 days but never bought
5) Lapsed customers

If we have at least 50 conversions per month for each of these segments, then we should be able to optimise them further.

Once you are done, you might have very different user experiences for each segment, as what converts one segment might lower the conversion rate in one and vice versa.

Using free tools like Google Optimize we can start to segment the traffic into silos like the above and find out exactly what makes them buy.

You will find out that the traffic to your site is very different in motivation and buying trigger. For example, we found that for segment 1, making the reward points on the checkout MASSIVE had a big difference in their conversion rate. However, reducing the reward points prominence for segment 3 had no effect and instead putting in testimonials here was much better.

The key eye ball real estate on your eCommerce store is quite small. The places that people focus on are narrow and they will miss your main sales points unless positioned carefully. For example, an existing customer knows you have free returns and free delivery and so might respond better to other messaging, whereas this is key for first time buyers.

“Find Out What Converts Best”

So, start cutting your traffic into segment silos and throw some variants at an AI engine and let it find out what converts best.

It’s quite exciting times.

Have a look at our Google Shopping video to see what segmentation and working with the AI (rather than against it) can do to your Google Ads results  https://go.markhammersley.co/get-started

Thanks

Mark (Cake Slicer) Hammersley