Best Practice on Google ACE
Google ACE is an excellent tool that came out in October last year, which allows for scientifically significant testing across many features within an Adwords Campaign. In this blog post we won’t discuss how to actually implement ACE tests (Google has that covered), but will focus more on the functionality and best practice when setting up one of these tests.
So what is ACE?
ACE stands for Adwords Campaign Experiments, and it is a tool that allows you to split potential Google traffic at its source. It is applied at the campaign level, and traffic can be split in a number of ways (usually we opt for 50:50) between a ‘control’ group and an ‘experiment’ group. You can also put parts of the campaign into both groups at the same time. Best practice here is to have everything you aren’t testing in both groups, the new part of the campaign you want to test in ‘experiment’ group, and the current part of the campaign (i.e. the part that you want to replace if the test goes well) in the ‘control’ group (this is explained in more detail further down).
What can you test?
You can currently test anything you can change within a campaign. You can add new keywords, negatives (Adgroup level), placements, different ad copy and alternative Adgroup structures. However there is currently no functionality to test anything applied at the campaign level. I have heard talk that the next thing to be added to ACE is the functionality to test site links, which would be nice, but our wishlist also includes GEO-targeting, campaign level negatives, networks, devices, location extensions, dayparting and campaign budgets.
How do you test?
We have found that the biggest limitation with ACE is its reporting. Google was obviously keen to get this useful tool in the marketplace as soon as it could, so reporting seems to have been neglected. The way we get round this at EF is by building duplicate Adgroups (a generic example is below).
Campaign: Holiday Destinations:
Control Greece Experiment Greece
Control Portugal Experiment Portugal
Control Spain Experiment Spain
Control France Experiment France
Control Italy Experiment Italy
This is the starting point for an ACE test (the above experiment adgroups are exact replicas of the control groups). We would then make the changes we wanted to test (i.e. a new landing page URL) to the experiment groups only, and start the ACE test to split the traffic 50:50 between both groups. Many of our account managers now maintain a continuous ‘test’ campaign, purely for ACE testing as they always have several tests running at any given time. This ethos of continuous testing is important in PPC (and a policy that Google itself has always championed) because it is the only way to keep learning ways to increase the performance of your account / business in the long-term.
So what is the benefit of using ACE?
In short, flexibility and accuracy. Previous tests relied on the before and after technique of making a change and comparing last week’s data with this week’s. The problem with that approach is that no two weeks are the same. Anyone who has been running a PPC campaign for some time (particularly those working on large-scale, complex accounts) can tell you that performance changes because of a variety of conflicting variables, including consumer demand, seasonality, the weather, even the quality of what’s on TV! Using ACE means both the control and experiment group are being exposed to the same market conditions. You can write off all of those variables because they should affect both parts of the account equally. You can be more confident that any differences in performance are due to your changes, rather than a statistical fluke.
As well as being more accurate, ACE is more flexible. Previous ad copy tests relied on putting the whole campaign on even-rotate. This approach is inefficient as your worst performing ads in other ad groups are now being shown more often than they should, dropping quality score, inflating CPCs and decreasing conversion rates. Using ACE allows you to test just the two ads you want to test, leaving everything else in the campaign as it is.
Used correctly and extensively, ACE is a great way to discover actionable insights into what drives performance in your account. It doesn’t take long to set up, is fairly intuitive and can help you drive performance. I would highly recommend anyone not using ACE to start using it as much as possible.
Michael Taylor, Associate Account Manager, Efficient Frontier