5 diabolical split testing mistakes

May 12, 2017 by
5 diabolical split testing mistakes

Marketing decisions are often based on assumptions, but if you don’t put these assumptions to the test, you’re definitely leaving money on the table. After all, it doesn’t matter what you think will be the best decision – it’s your customers who matter. They’re the ones keeping your business afloat.

Split testing allows you to find tune your PPC advertising, business site and value propositions to gel perfectly with your audience’s desires, with the ultimate goal of maximising your revenue. In fact, Barack Obama used split testing extensively to receive an additional 60 million USD in funding for his presidential campaign!

But split testing isn’t necessarily a simple matter. We’ve put together some of the more common mistakes, so you can avoid them.

1 – Not running your tests for long enough

In the third quarter of the Super Bowl, the Atlanta Falcons were up by 25 points and had a 99.7% probability to win the game. Just like the Falcons, your split testing variation could show enormous potential before going on to lose eventually.

You need to account for variance from day to day – the longer you run the split test, the more accurate your data will be in terms of predicting long-term trends. Running split tests for at least a week is recommended, although it depends on the level of traffic you are generating.

2 – Arbitrary testing

Rolling the diceRunning random, arbitrary tests is a surefire way to waste your budget and achieve very little. Sure, you might get lucky and find that a second “Buy Now” button located underneath your reviews improves product page conversions – but it’s far more effective to analyse where people are getting stuck in your funnel, hypothesise how to fix it and then run a test to verify the results.

Peter Boyle, writing for Crazy Egg, states that there are only two steps involved in creating a successful hypothesis:

1. Find the problem or sticking point for your customers.
2. Figure out a way in which you can solve that problem to increase conversions.

3 – Not accounting for a lack of consumer demand

If no one actually wants your product, then you can split test all you want – you still won’t be able to achieve a good product page conversion rate. One of the best ways to check whether your product is going to sell is by finding a similar product on the market and then looking for indicators of sales.

Try looking up your main competitors on SimilarWeb and see what kind of traffic they’re receiving. If all of your competitors are receiving low traffic, then the problem is your business idea rather than your split testing campaign. Likewise, if you’re marketing a physical product, sales on Amazon and other online marketplaces provide some indicator of consumer demand.

For more information about validating a business idea, check out this informative article by Pat Flynn.

4 – Testing multiple elements at once

If you test multiple elements at once, this can lead to confusing data. It’s more tedious to test one element at a time, but this is the only way to do it if you want to know, unquestionably, which elements are improving your conversion rate.

If you’re testing the price of your product, don’t go changing other elements of your landing page within the same time duration of the test. It’s easy to become impatient and want to change things if results aren’t going your way, but it’s more important that you allow tests to run their course, since this will give you the insightful data that you can use to grow your business for the long-term.

5 – Giving up too quickly

Not every split test will produce positive results. In fact, you might find that with your first handful of split tests, your new variations are consistently yielding negative results! That’s perfectly fine, not every split test is going to impact the fortune of your business. According to Conversion XL, only one in eight split tests drives major changes.

Split testing is a long, ongoing process to which you need to stay committed. Instead of one change that turns your business around, it’s far more common to make small, incremental changes that eventually result in a more profitable business. If you can avoid chasing the immediate gratification of sales and stay committed to the long-term process of testing, you will be rewarded for it.