Find $100m Lost Revenue Without Creating a Single Experiment

Abi Hough

It’s a bold claim, but entirely doable. How? Functionality testing websites and finding bugs that create barriers to conversion will find you a bucket load of cash for very little return on investment… in just 3 years I have recovered around $100 million worth of this sort of revenue, lost money no one knew about because nobody has bothered to look.

Functionality is the foundation

Every website needs to be built on a solid foundation. In fact, Eisenberg puts functionality at the very bottom of his hierarchy of conversion, it forms the base of his pyramid, the foundation which is why it needs to be robust and strong.

Think of optimisation as if it were a tree. At the top of the tree we have fruits, flowers and leaves. These represent the things that optimisers do on a regular basis to form research that allows optimisation to begin, such as heuristics, eye tracking and heat mapping (to name a few).

Analytics represents the soil that the tree grows in. We have to ensure the soil is of good quality so that the fruits, flowers and leaves can grow well in order for us to harvest them.

Functionality (usability and performance also) is the root system of the tree. If the roots are dying, the soil is pointless and there will be no harvesting of anything.

An illustration

A travel website was experiencing a loss of conversion rate on iPads. iPad users were served the same experience as those on desktop, but the analytics showed they were converting at half the rate of those visiting the website via a traditional desktop browser. So, what was going on?

I carried out a heuristic smoke test on the website using a variety of physical devices to try and discover the problem. The testing period lasted 3 days. The website was tested on 2 tablets, 1 mobile and 6 desktop browsers and unearthed 99 issues.

What was discovered?

A particular button in the funnel that allowed a user to finalise their purchase did not activate when a user tried to purchase a multiple destination journey. The user was unable to checkout even though they had made valid choices.

How much was it worth?

This individual broken button was losing the company £65k a day, or about £24m pound a year. Think about this. How many A/B tests could you run for £24m? Quite a few. Now think about how long each of those would take to research, build and run and what the potential return on investment would be. What works out better? And, why are we not functionality testing as part of the course?

Was it a one-off?

Absolutely not. It’s an epidemic. We suffer from an infestation of bugs and it is guaranteed that no matter what website you look at you will always find them, provided you look hard enough and in the right places.

The problem with bugs is that they’re a lot like having a bit of spinach stuck in your teeth. Not many people will tell you it’s there, but they’ve seen it and let’s be honest it doesn’t look pretty. However, it only takes one person to tell you, and you can look in the mirror and fix the problem. Always listen to that one person, but better still make regular checks that the spinach isn’t there in the first place.

The lesson

  • Start with the best foundation you can, functionality test websites before you start optimising it and make some easy money before the hard work starts
  • When you do start optimising and creating A/B tests make sure you test your test

Top tips

  • Always have your own in house device lab and keep it well stocked with physical devices.
  • Simulators are not reliable enough to carry out accurate functionality testing.
  • When testing, use virgins! Seriously, you get better results by using someone who has never interacted with the site before but has enough technical know-how to break things.
  • Manual testing reaps the best results.
  • Automatic testing is OK to use for repetitive tasks but it cannot simulate the qualitative results human testing provides.
  • Ensure functionality testing becomes part of your optimisation process and your clients day to day business.