I’d be a rich man if I got a dollar for every ad tech vendor that suggested I “test” their application. “Let’s do a test,” they all say. They’re confident that their technology is the best-in-market and will smash any other tech out there. This is a rather bold statement given that there are about 150 different ad tech companies in Australia, all saying a similar thing.
Putting aside the oversupply of vendors, it’s safe to say that “tests” have a number of shortfalls. Stop and consider the following five factors before suggesting a test.
1. Tests done poorly are stressful
School kids don’t beg their teachers for maths tests. Patients are usually unhappy when they ask doctors to run blood tests. Tests, by definition, are stressful. They require loads of prep and crazy amounts of focus. Next time you suggest a test, think of stress endorphins rising in the other parties.
2. Tests should be executed by professionals in labs
The FDA (Federal Drug Administration) tests new medicines very carefully before calling them safe. Imagine if they only did one test on one subject and then called the drug safe. Their tests are done over a long period of time and they analyse millions (if not billions) of data points before any conclusions are made. We should have the same type of rigor with ad tech and media products and solutions.
3. Tests should be replicable
If a test is successful (or not) then we should be able to see why. If we don’t know why a test succeeded or failed, then the test won’t be replicated. Too many times I’ve seen tests done on a single campaign, and the results were good but inconclusive. How can we have confidence if we don’t know why it worked?
4. Tests should be done once and the results should be published
Ad tech companies should pay independent companies (such as PWC) to review and publish their test results. If they did this, I could read these results and I wouldn’t need to run the test. Instead, I feel like a guinea pig because I’m always asked to run tests. I can’t be the only person globally performing these tests!
5. The person who suggests the test should pay for the test
Lastly, but not leastly, the ad tech company that suggests the test should stump up the money. They should be one that is willing to pay for the test. As the old cliché goes, put your money where your mouth is. Why should the advertiser be forced to pay for tests that prove the effectiveness or otherwise of your product?
So for all the ad tech companies out there getting ready to suggest another test in their next meeting with a media agency or advertiser, please consider how the test should be run before you suggest it.
Tim Whitfield
Director, technical operations
GroupM
