True Split Tests Return to AdWords – Whew!!!
At Vitruvian, we’re big fans of Google AdWords. It’s the best medium we know for immediate traffic and accountable results.
But we love AdWords even when it’s not a great medium for generating leads and sales.
It’s the market insight that AdWords provides that’s its most valuable benefit. Because it’s a big world out there – much bigger than AdWords. When our clients experiment and hone their message and strategy with AdWords, they can then apply that message and strategy everywhere else, with improved results.
Why is AdWords so good for this? Because it was built as a testing machine that just about anyone could drive.
For the first time in history, you could run competing messages, offers, and price points and get the market to choose its favorite in days, not months; for pennies, not tens of thousands of dollars; and through actual behavior, not focus group hypotheticals.
And creating split test ads has been so easy, even non-technical people like me could jump in and do it. Just click “New Ad” and start filling in the text fields.
The key to accurate testing is randomization. Meaning, if you are testing two ads, each searcher has roughly a 50/50 chance of seeing Ad #1.
Sadly, AdWords has never made this a default setting. If you don’t know better and go in and change that setting, Google will show your “better performing ads” more often.
What’s wrong with that? After all, don’t you want better performance?
The trouble is, your definition of better performance (more profit for you) is not the same as Google’s (more profit for them). By “better performing,” Google means ads that generate the most clicks. Clicks is how Google gets paid. Conversions is how you get paid.
Google has set up the system to favor ads that make Google money. And most advertisers stick with this option, either because it’s buried behind two sub-menus in Campaign Settings, a page few people ever look at anyway, or because when you choose a different option Google warns you that you’re doing something that isn’t recommended.
But until recently, you could opt out of the “More Money to Google” setting and run true split tests, showing each of your ads equally, regardless of performance. A few months ago, Google quietly removed this option, replacing it with “rotate ads evenly for 90 days, then optimize for clicks.”
I’m pleased to say, after who knows what backdoor machinations, Google has returned the valid split test protocol to the menu:
Though it comes with disclaimers and warnings (which, arguably, should make you happy if they discourage your competitors from engaging in valid tests), it’s there.
And until you’re ready to optimize your ads for conversions (which could take a while to build up enough traffic to your website to get sufficient data), this is the best setting to use if you want to get the benefit of AdWords’ market insight.
Think of it: in a few days, for a few bucks, you can find out whether you do better aiming your ads at men or women; students or their parents; professionals or amateurs.
You can determine whether your prospects are more eager for a free trial or an online demo.
You can figure out whether they’re driven more by hope or fear.
And you can apply that insight to the rest of your marketing; website, offline media, elevator speech, tattoos, you name it.
Thank you, Google, for bringing back the best of AdWords!