After analysing the context we decided to split up the conversion challenge into three pillars – V2U (Visitor To User) U2C (User To Customer) and C2AC (Customer To Active Customer) where a Visitor is new to the site, a User has filled in the registration form and registered, a Customer has added a product to a basket and an Active Customer is someone who has checked out and bought something.
We decided to focus on V2U for 30 days to see if we could increase conversion at the top of the funnel (TOF).
We also decided our one and only growth metric would be zero day conversion, in other words the percentage of people who convert the same day they first visit (as opposed to metrics like 30 day conversion that incorporate people who come back and register later).
We decided to A/B our ideas as we iterated through the tests to try and cancel out any macro effects on the conversion rate.
The starting V2U experience was a generic home page with some vague “lifestyle” messaging and a CTA (Call To Action) box that linked to a second page with name/email address/phone number form on the left hand side with supporting text on the right hand side.
We decided a cross functional growth hacking team to attack this challenge and given the front end of the web site was WordPress we set up the core of the team as 1. A gro.team Growth Hacker 2. A WordPress Developer 3. A customer experience Marketeer. Other representatives from Operations, Technology, Product, Sales, Finance etc would be called upon as required. We started looking at daily same day V2U numbers for new visitors and we were ready.
Top Tip…restrict the test to new visitors only…returning visitors are much less likely to click a signup Call To Action (CTA).
We got access to Google Analytics, Google Search Console, WordPress, Maxymiser (for A/B testing), we set up a Slack channel, decided to have a stand up every day at 10 am and we were ready..
Top Tip – It’s very difficult to draw meaningful conclusions from raw conversion time series data from analytics tools due to the impact of extraneous factors like the weather (which effects conversion a lot by the way). That is why we need to A/B test…theoretically the impact of all other factors are cancelled out leaving the only difference in macro and micro experience being provided by the experiment.
Oracle Maxymiser is expensive. Google Optimize is another (free) option.
We had a kick off meeting with the standing team to talk about how we wanted to work and what we wanted to achieve…a cross company ideas generation brainstorm was set up for the next day
We invited representatives from across the company with roles ranging from customer support to UX Design to get in front of a whiteboard for an hour and collect some ideas on what we could/should do.
We set up the meeting up with the context of reminding ourselves that sites have less than one second to tell Visitors What the site does, Why they should trust us and WIIIFM (What Is In It For Me). We looked at the marketing flywheel Attract –> Engage –> Delight…and we were ready to start generating ideas…the first half of the session (known as divergence phase) was about generating ideas and then the second half of the session was about getting a feel to what the group felt were the most important ideas (known as the convergence phase). It was a great session and we all left with great energy and enthusiasm to get cracking…)
All the 40 ish ideas from the Ideas Gen session were transcribed onto a Trello board at a status of “On Ramp”. We also decided to use the gro.team growth board statuses of “Parking Lot”, “In Progress”, “Complete”, “Live” and “Blocked”.
The ideas from the Ideas Gen session could broadly be broken down into three buckets…
We were now ready to launch Variant 1 (our first experiment) which was to improve the WIIIFM messaging on the home page, start the sign up process on the home page and to lose the ask for a phone number during account creation. We used Maxymiser to launch the A/B experiment to 50% of new visitors and ended the week happy we were in the game.
Top Tip…Maxymiser is powerful but not particularly easy to use…remember to permission accounts so that they have permission to launch experiments to avoid hours of frustration…
Top Tip…Writing new copy (text) for sites needs to be done in context…the person from the Commercial Team that was writing our new messaging ripped it up and started again once they saw how the text would fit and look on the page…context is everything…
After releasing our first experiment we were all eager to see the results…it was very early days but Maxymiser was reporting a conversion uplift of 101.76% for Variant 1 with an uplift of 113.21% for engagement! At the standup we decided to focus Week #2 on creating Variant #2 as a longer more engaging experience for the three main product lines. We left the standup eager to engage the Marketing Team for the new copy and the Design Team to help with the new UX…
We spent today working on Variant 2 which was to bring much more WIIIFM information about the three products on to the home page and invite visitors to select which one they’re interested in. The uplift in conversion being reported by Maxymiser has dropped back to 68% but still looking positive overall…
We spent today doing more work on Variant 2 to bring much more WIIIFM information about the three products on to the home page and invite visitors to select which one they’re interested in. Asking the Visitors to self select what they’re interested in like this will allow us to provide a more more engaging journey for them before we know too much about them…
We spent today doing more work on Variant 2 ready for launch tomorrow. Most of the trouble today is getting Maxymiser to run 3 experiments in parallel (33%/33%/33%)