Best Practices for Google Display Ads Testing

Key Takeaways

we've covered some of the best practices for search ad copy testing on this channel before and if you're interested in that video you can check it out right here but one of the things we didn't cover was around display ad testing now a few of the things for search are going to be similar best practices in display but display is inherently different there's a lot of different factors that go into it so there are a few best practices that are not applicable to search so in this video I want to talk briefly about the best practices that overlap between search and display ad testing but then dive deeper into some of the more specific nuances of display ad testing to get you started down that path in your account as I mentioned in the intro I just want to do a brief recap of the ad testing best practices that we've talked about for search in a previous video that also applied to display first we want to make sure that we test in regular Cycles a really quick recap of this is that I usually start a very large test with very different patterns of messaging in different variants than depending on the winner I'll iterate on the winning variant and then again I will iterate one more time to come up with a third more narrow test for the third time around and then start back over with a completely different large message change which also carries into number two here I want to make sure that I'm constantly testing high level message changes as well as breaking down the different pieces of each message to understand what makes one perform better than the other so we can carry that over into different campaigns efforts channels what have you we always want to make sure we're running at least two variants at the same time and on the high end I would probably say up to six but this is something that in conjunction with number four we need to make sure that we have enough volume to run a certain number of variants if you have less volume probably stick to the two or three range but if you have lots of volume coming through your campaigns and you might want to be able to test a bunch of different things at the same time you could probably go up to four or five or six different variants and then you need to let the test run long enough so that it can gather data and then you may or may not use statistical significance to determine the winner with that data but if you have a relatively low volume account two weeks is not going to cut it you're going to need to run the test for longer than that but you basically want to have enough data that you feel confident making a decision in a winner so that you can start the testing cycle all over again so with that let's start talking about the handful of distinctions that I want you to pay attention to when it comes to display ad testing as compared to search the first is to compare within ad formats and not across ad formats for this account we're running two different types of AD formats responsive display ads which I'm calling RDA days and then the banner or image ads which I'll just call images we're not using HTML5 in this account so we're not going to look at those but although we can run both of these ad formats in the same campaign I don't compare these to each other when looking at performance there's a few reasons for that that I've listed down below responsive display ads are extremely flexible and images are not Google is able to put a responsive display ad in every different placement around the Google Display Network and images only fit in their one specific spot so the performance is almost always going to be quite a bit different this account specifically is not looking at conversions but I did leave that column on there the biggest piece we're trying to look at is the click-through rate because this is just focusing on engagement so you can see that the click-through rate for the responsive display ads is quite a bit higher but also the cost per click for the image ads is lower so in these different areas depending on which metric you're focusing on you might have one format as the winner versus the other when they're both providing quite a bit of a value so rather than comparing all line items in a campaign all individual ads regardless of AD format I actually break them down and compare only responsive display ads to other responsive display ads to determine a winner since they're the same format now images are going to be a little bit different and this is another Nuance within image ads I only compared to other image ads but I also compare only within the image ad sizes that are comparable to each other here I have not segmented by anything other than the add size nothing about ad messaging device anything we'll get into that in a little bit but here you can see there is quite a big difference in the click-through rate and the cost per click for these different ad sizes by themselves 300 by 250 is a very common size it has a huge click-through rate cost per click is on the higher end but still pretty good it has a lot of volume whereas the bottom line 970 by 250 has a pretty terrible click-through rate the cost per click is really high so it doesn't make sense to say 300 by 250 is the winner and 970 by 250 is the loser because that 970 by 250 still gained 15 000 Impressions that we would not have gotten if we only ran 300 by 250. so if you're going to run image ads for your campaigns don't compare one ad size to the other instead we need to compare themes across these different ad sizes so here in this account we have eight different themes running and each of them runs all of the different ad sizes that you saw on the last screen now none of these words should mean pretty much anything to you you shouldn't be able to recognize anything here but you can see that the click-through rate varies for each of the different themes that we have here and so does the cost per click the 2x4 has the lowest click-through rate as well as the highest cost per click so this to me doesn't necessarily seem like the right fit because we're running either responsive display ads and image ads or just image ads to this theme and other themes are outperforming it if we turn this off we'll still be able to capture those Impressions because we're not turning off an entire placement or ad size we're just turning off one of the messages that doesn't perform as well here now I know that's a lot to keep track of but what I like to do to make analyzing my display ad copy tests easier is to make a label that's associated with each of these different pieces so I might have a label that is for the theme and for some ads it'll say boost others it'll say Trends all the way down the line it might be easiest if you just include the word theme at the beginning of the label so it'll say theme Dash boost theme Dash bear whatever makes the most sense for you but that way you can easily identify which labels are associated with themes versus something else the ad sizes same thing here pretty easy just label it 300 by 250 728 by 90 and although you can filter by add format in the add interface I still like to put labels for rdas versus image ads versus HTML5 when I'm using it because more often than not I'm going to be in the label report within Google ads and it's easiest for me to just use the same tool rather than going from one interface to look at one component to a different interface to look at the other the nice thing is once you have all of these different labels here and they're all separate labels you can not only compare how each theme performs but you can also see how each theme performs with each ad format and what image size works best for each theme there's a lot of different overlap in these reports that you can find exactly what's performing best so again going back to that testing variant you can create a new version an iterate on it for your next round aside from all the different components to look at when you're trying to analyze your ad test you also need to know what metrics to look at as well just like with search campaigns or social campaigns anything across any platform your campaigns might have different goals and depending on those goals you need to make sure that you're looking at the right metrics I have some really high level examples here for Branding maybe you need to look at just Impressions viewable Impressions and your cost per thousand Impressions or CPM because for brand you're really just trying to get in front of people there might not necessarily need to be a secondary step for you to see success with it but with engagement maybe you're just trying to get people to engage with your brand in one way shape or form so the metrics that you might be comparing in your ad test could be click-through rate the cost for those clicks just the clicks themselves or different engagement or engagement rate metrics as well and then lastly my guess is that many of you are focusing on conversions where you're going to want to focus on conversion volume cost per conversion and conversion rate but then I also have one of my favorite metrics in there impression to conversion where instead of calculating your conversion rate by taking clicks divided by Impressions you calculate your conversion rate by taking conversions divided by clicks for this metric you actually take conversions divided by Impressions and find how many Impressions it takes on average for you to get a conversion with display where there's tons of Impressions available I like to use this metric sometimes more often than the click-through rate because I find that display has a little bit more of a need to catch somebody's eye and of a certain theme or ad size or add format does a better job of catching somebody's eye and then they eventually convert I want to lean into that with my display campaigns now a minute ago for The Branding section you might have heard me mention viewable impressions for display I think these are really important as well as viewable CTR here I have metrics broken down by device but you can easily see the difference in performance for Impressions and viewable impressions in the second and third column so for the total view we have 156 000 Impressions and only 81 000 viewable Impressions that also changes our viewable click-through rate quite a bit overall click-through rate is 0.42 percent whereas the viewable click-through rate is 0.54 now these metrics don't matter nearly as much for a search campaign but for display makes a big difference here are the definitions for those metrics a viewable impression counts an impression only when the ad is viewable and the way that Google defines viewable is that that at least 50 percent of its area is visible for one second on the Display Network or two seconds for video ads viewable click-through rate is how often somebody clicks your ad when it is viewable and then it again gives the definition of a viewable impression I highly encourage you to use the viewable Impressions and viewable click-through rate for your display campaigns because I think that 50 of your ad being visible for one second is a very low bar to hit and very easily missable and as the example that we just saw if almost half of the impressions are not even viewable I don't want to compare that performance because that doesn't make any sense to me I'm trying to understand which theme ad format image size messaging all this stuff trying to understand which of that performs best and if the user didn't even see it then it doesn't make sense to use that impression for my calculations a couple of other things to keep in mind I'm using the same table because you might have seen here that the performance by device also very varies differently looking at the viewable click-through rate tablets have a pretty high rate at 1.17 mobile phones are next at 0.85 and computers are at 0.27 so if you're comparing performance across different device categories or even campaigns that have different device settings make sure that you're taking that into account when you're reviewing performance and not determining a winner based on incomplete data where maybe one campaign is heavily focused on computers and the other one is heavily focused on cell phones and it's causing your data to be skewed I would say the same is true for your audiences for search everybody is actively searching for something so whether they're new to you a returning customer somebody who's still shopping they're probably still looking for you with almost the same amount of intent behind each of the searches for display it can be quite a bit different since this is a push strategy here you can see that for these campaigns prospecting has the highest viewable click-through rate of 0.61 retargeting which are people who are not yet clients but are are somewhat familiar with this brand are only at 0.25 and clients are way low at .04 so again if we're trying to promote this webinar as you can see in the client campaign name across all these different audiences probably makes sense for us to take into account at a high level that the click-through rate is very different and analyze the ad messaging format sizes all that stuff differently based on each audience because they do perform quite a bit differently overall display ad copy testing is not too drastically different from search there are just some key components that you need to keep in mind because the fundamental differences between the platforms and how the users will engage with your ads if you have any additional questions about running display ad copy tests or just ad copy testing in general feel free to leave us a note in the comments below thanks for watching our video we really appreciate it if you liked it give us a thumbs up below if you really liked it maybe think about subscribing to the paid media Pros YouTube channel and you'll get alerted every time a new video drops if you really really liked it you can help support the channel by checking out some of the t-shirts that we're wearing on our merch shelf as well as looking at the super thanks button all right