A/B and multivariate tests are super important when it comes to eCommerce. My team and I have sworn by it for years. I have spoken about it on stage and on podcasts for years. We test everything and anything you can. For the past several years, most of our website testing has been done through Google Optimize. We have always used a partner for the execution aspect of this.
Historically, our process has been something like this:
- We (internally) come up with a list of things we want to test on our site over the next several months.
- Partner comes up with a list of things they want to test on our site as well.
- We stack rank the list based on the likelihood we think it will make an impact.
- Let’s say combined we come up with 20 things to test.
- We will likely have 4-6 things we are super excited to test out of the list so they go in the top spots on the ranking which decides what we test first.
- We take the remaining things and place them on the bottom of the list but not necessarily immediately after the 4-6 top things. Why? Because we know with certainty that we will gain some learnings from the initial tests that will, in turn, give us some better testing ideas to put above the bottom part of the list. Also, we might decide after these learnings to pull something from the bottom of the list to the top as well.
- We all agree on the course of action and the partner executes via Google Optimize.
Thank goodness they know how to use Google Optimize because nobody on our side knows how to use such a complicated platform! Yes, my team and I always just assumed that Google Optimize was complicated and required vast technical knowledge.
For the past year, our development agency partner handled Google Optimize. Prior to that we had an advertising agency partner handle it. Our development agency partner, Hulk / Praella, does this with all of our maintenance and normal initiatives as well. We pay them a monthly retainer for a block of hours. At all times we have them working on all kinds of stuff besides testing. Because of this, we are testing in a slight jog, and I like to run. To be fair, previously, our advertising agency partner was power walking at best. A jog is likely fine for 99% of eCommerce brands, but again, I like to run.
At the beginning of the year, on our weekly catch-up call, Walter and I were discussing Google Optimize. Walter likes to run as well. We decided we were going to carve out a day in the near future and try and get a basic understanding of Google Optimize. We are both pretty smart guys. Surely, if we block out 6 hours of a day to focus on learning Google Optimize, we will have a decent foundation.
A few weeks later, on our weekly catch-up call, Walter asks me what day the following week were we going to try and learn Google Optimize. That’s #2422 of the reasons Walter is a Rockstar. He is always moving the ball forward and holding me accountable when appropriate. We agree on a day and move around scheduled meetings so we can focus on the goal.
Fast forward to the day of learning. Walter came over in the morning and we both worked out of my place for the day. In preparation, I found about 3 hours of videos that I thought would be a great starting point for us. OK, we are doing this.
We watched the first video which was about 13 minutes. It was horrible. I felt stupid. We were not going to be able to learn this. It was by a marketing growth agency and I guess they did it on purpose. They wanted to maintain the value agencies yield by knowing the complex platform known as Google Optimize. The platform you will NEVER understand!
We watched the next video and sponged up all 8 minutes of it immediately. Wait, what? This is super easy. Ridiculously easy. (this video https://www.youtube.com/watch?v=l3thoiqvH6E).
We followed up that video with a super corny super basic video that Google put out (this one: https://www.youtube.com/watch?v=cFVnkSYPYXI).
We had planned for 5 hours of education with the hopes we would have a good understanding afterward and could schedule some more time to actually play around with the platform and eventually take over management. We were 30 minutes in and we were ready to jump into Google Optimize. We jumped in, quickly got our bearings, and set up some future tests in literally less than an hour (we had tests already running via our agency partner). We spoke to the development team, explained we would be running point on Google Optimize, and made sure the line of communication was set up correctly for us to do so. The main takeaway was that when we add the line of code for a Google Optimize test, we needed to shoot development a note about the code so it was backed up in our repository (so it was not lost when we pushed live a new version of the site as we do this almost weekly).
Walter and I were both shocked (and still shocked) and how easy the platform is to spin up tests. Needless to say, we are about to run! I will likely share some interesting tests as we come across them. The last test our agency was running ended last week and our new test (that we executed 100% of) is currently live on the site.
The takeaway here is simple. Google Optimize is very easy to use. Should you be running these tests yourself like we plan to? Maybe, maybe not. It depends on you. Agencies definitely bring value by being able to manage these tests for you. They also have insights into specific things to test that you might not be thinking of.
Another takeaway, which most will know, but I want to say anyway…
It is natural to have a ton of ideas to test and running multiple tests at one time is possible, BUT know if you are running multiple tests in the same area of your site you are adding in additional variables which might make the results not super clear. We are sticking to a single test per area of the site so that the results are super clear and actionable.