A/B Experiments on Landing Page Conversions
What I learned doing some fun A/B Experiments on Landing Page conversions & more at Goodbox.
Context:
Goodbox is a platform where businesses come and create a mobile app presence by creating a Mini App — which is a readymade app within the Goodbox Mega App. We had a landing page where we explained what our product offering is. There is a Call to Action (CTA) on this page that will take the user to the onboarding process. We ran an adwords campaign (people searching for ‘make an app’, ‘create an app’ etc.) that directed traffic to this landing page. We measured the conversions from this page to the Onboarding page (no. of CTA clicks) and tried to improve those numbers as it was a very important funnel for us.
Experiment 1: Girl vs. Woman
Our target audience were Small and Medium Businesses in India. Lot of them are not tech savvy. So we wanted to have a friendly, humane feeling to our website and not show some cool tech gadget or a mobile app. So, we used an actual human picture.
When we first launched, we had a girl holding a phone. She looked kiddish and we assumed that people may not take our site seriously. So, we experimented by changing the girl to a woman and the conversions improved!
Experiment 2: Woman vs. Confident Woman
Then we said, lets try with a more cool and confident woman but we didn’t see a significant change in the conversions.
Experiment 3: Woman vs. Man
Later, we tried the same experiment with a guy pointing his hands towards the CTA and our conversions improved again! Somehow, the bounce rate was also down a bit and avg. session duration increased!
Experiment 4: CTA Text
One version of the design had the text on the CTA as “Create your Mini App”. We tried a variant where the text says “Create in 5 minutes”. Our hypothesis is that introducing a time component will give some sense of urgency and gives a feeling to the user that ‘Ok, if I click this button and go ahead, I will get something in just 5 minutes. Let me try.’ We were able to see better conversions with the variant thereby validating our hypothesis. Sorry, I don’t have the exact numbers to share on this experiment.
Experiment 5: Create vs. Get
We thought may be ‘Create’ is a strong word. It sort of gives a feeling that the user has to invest some time and effort in ‘creating’ something. We wanted to try with something that sounded simpler — ‘Get’. You don’t have to create anything — You just get something in 5 minutes.
But we actually didn’t see any significant change in conversions in this case.
Experiment 6: CTA position
Using Mouseflow, one typical behaviour pattern I observed was that people come to the site, click on pricing, the page scrolls to pricing section, they see the plan amount and they either drop off or they scroll up and down to find the CTA to click. They are not really interested in learning about the plan details.
A.) If they are scrolling randomly to find the CTA, it’s likely that they find the price cheap or reasonable and would want to try out our product. What did we do for these people? Put the CTA right above the pricing table so that users can immediately find what they are looking for and lesser chances of getting lost.
B.) If they are dropping off, it is highly likely that they find the price too high. What did we do for these people? We introduced an exit dialog, meaning when the cursor moves out of the browser window, we started showing a big flashy screen that says ‘Creating your own app can cost anywhere between Rs. 5–50 lakhs! Mini Apps are awesome!” that will essentially tell the user that may be you didn’t read our offering fully, we are a much better option than an own app, come back, read more and decide.
Doing both A & B showed a huge jump in our conversions! We later included a sticky CTA as well on the bottom right that bumped up our conversions a bit more.
I thank my bosses, Mayank and Nithin for helping me in conducting these experiments.
Tools used: Google Analytics, Google Optimize & Mouseflow
Clap for this article and follow me on medium for interesting design stuff :)
