Split Testing Data Proves It Does Make a Difference

For some years now I have been telling all my students and clients that split testing is an absolute must for any serious marketing professional or for that matter anyone who has an interest in seeing one of their ads perform better.

I have based this advice on not my only common sense but on my own personal experiences designing, implementing and cross testing various types of promotional media ranging from printed media to search engine sponsored links such as Google Adwords.

But in all honesty up until now I have never actually seen any empirical data on the topic that supports (or refutes) my position on split testing. Thankfully this week I finally found some that I'd like to share with you from Marketing Sherpa.

The pie-chart above shows the data from the Website Optimisation  Benchmark Survey from 2012 with 789 respondents.

There are some key takeouts here to be addressed. Firstly and sadly I think is that over half the respondents (53%) could not or did not calculate their respective ROI. When you read data like this you just know that it reflects the reality of our own local situation here in Australia as well.

However on a brighter note of the remaining 47% who did calculate their ROI figures, split testing was shown to have a positive or neutral effect in 38 of the 47 percent of in real terms, 81%.

In simple words 4 out of 5 people who did it, at worst had no loss and usually an increased ROI.
It's staggering in fact that the  remaining 3% actually ended up with a worse effect and this may have been for unrelated issues such as lack of budget, media selection or timing of ad.

It's just reassuring to know that occasionally we get to see the proof of what we know to be right.
So don't delay and if you haven't already done so, start split testing now to deliver a more effective and impactful promotional message.

So until next time, good luck and good marketing.
Regards,
Daniele.