Improving Your Testing Mind Set
The subject of testing your advertising is one that usually starts and ends with a chapter or two in a Marketing 101 text book. And while it may seem easy to accomplish in print, it’s somewhat more difficult to achieve in the real world we all operate in.
Try convincing the local newspaper to share two different versions of your advertisement across its print run to see which one is the best. Or your radio station to switch alternate adverts at the same time of the day to pick the winner. Perhaps you could re-shoot the end of the next TV commercial with three different calls to action and try to keep it all within your modest budget. Achieving any one of these with a view to improving your own testing methods would be a struggle and possibly prohibitively expensive to achieve.
Online the story is a bit different. Here, with the low cost of production and easy access to free tools to manage the testing process, marketers have the chance to really unleash the tester in them that may have lay hidden for years. The only issue is that all this new capability can be easily squandered by marketers who struggle to capitalise on any gains this opportunity brings with it because they lack the experience in how to best make testing work.
To help those who are new to this challenge, I’ll outline four points you can apply. It’s by no means an exhaustive list; just enough to help you avoid the common mistakes we see and, by doing this, ensure the time you invest in testing is well spent.
But first, why should you embark on testing anyway? For those not convinced of the benefits, try reviewing this article I wrote earlier on the Multiplication Effect of Testing: http://tinyurl.com/56ansj. The improvements that testing bring can help you achieve dramatic and long-lasting effects on your marketing. For instance, all it takes are a few tests that return an extra 3% increase in conversion rate for each and you can be well on the way to doubling your advertising effectiveness.
Convinced of the value but not sure what to test? Here are a few ideas. Our testing methods usually follow the paths a prospect takes through our customer’s marketing funnel – going from the top to the bottom. So if they are running a simple lead-generation campaign using Google AdWords for traffic, we would test the following in this order: Google Ad Words headline, body copy and display URL, and then the landing page headline, lead-in copy and body copy.
These four tips illustrate the process we follow to ensure our efforts with testing will be worthwhile.
Start your testing at the front end of your marketing. (In the lead-generation campaign example the front end would be the AdWords campaign.) All tests will require a steady stream of traffic to run through them before they reveal the winner. In most cases, there is more traffic volume towards the front of the sales process, which allows your tests to be completed faster and, therefore, improvements can be applied quicker.
Consequently, any gains in efficiency can then be passed onto tests further down your sales process. So, once you have an ad that produces a sizable number of clicks at a tenable cost, then focus your mind onto the possibly more complex tests you may need to apply to the next step – the landing page.
Start by testing the big things. The other day I heard a great analogy for this – first test the forest, then the trees, then the branches and, finally, the leaves.
So, start with two vastly different Google Ads or Landing Pages (forests), then find the theme of one that works and needs further testing against quite different copy (trees). Then test the headlines (branches), followed by the font type and background colour (leaves).
Concentrate on testing one thing at a time. The ease with which you can alter your online advertising copy frequently trips people up here. For instance, when we are called upon to review Google AdWords accounts that include multiple text ads, quite often we will see two ads that are different in multiple ways (say a headline and the display URL). Now, if one is pulling 30% more traffic than the other, which part of the ad is doing this extra work?
The same applies to landing pages. Compare headlines between otherwise identical versions, rather than changing headlines, body copy text and images.
Yes, all this takes time as you have to run more tests. But each test builds on the one before it. Also realise that some tests will take you backwards and reduce your results, so knowing what worked before makes it easier to back-track to the previous winning version.
Allow your test to run long enough to provide you with accurate results. The more results you collect, the higher the probability that what you see accurately reflects reality. Fortunately, there are some great free online tools that will help you make this statistically based decision. For instance, this tool will help you determine if your AdWords ad is a winner: http://www.vertster.com/adwords-tool/
So there you have it. Four pointers to ensure the time you invest in testing is well spent. Why not set up some tests this week and see what the results tell you? There could be some big improvements just waiting for you to discover.