Strategies for Reviewing Auto-Placements on the GDN: Part 2 – Performance
This is the second in a two-part series on how to review automatic placement reports for Google Display Network campaigns.
In my last post, I discussed how to review the content of the placements and evaluate if they’re a good fit for your account. In this post, I’m going to focus on the performance from each of the placements our ads have run on.
In almost every case, you’re running Google Display campaigns to get some sort of performance from them. Very few people run campaigns just for the sake of running them and gaining impressions. Nearly all of my accounts are focused on conversion actions from the GDN, even if those actions are more top of the funnel.
As we go through each of these performance filters, there are two key things I’m looking for: underperformance and unrealistic metrics.
Here they are, in no particular order.
High Impressions, No Clicks:
Sometimes it can be great to have your ads running on lots of sites and seeing few clicks. Arguably, if it’s the right site content and the right audience, you should still see branding benefits. But odds are if you’re not seeing any clicks, it’s likely a bad match.
Start by filtering your report for any placement with Clicks = 0 and then sort the sites descending by Impression count.
Depending on the volume of impressions you typically see with your campaigns, a high count here could be in the thousands or tens of thousands.
I typically like to do a thought experiment like this: If this site gets one click on it’s next impression, is the CTR acceptable or still unbearably low?
Depending on the answer to that question (and the content match for the site), I can choose to add sites racking up impressions as negatives or keep them running.
This experiment gets a little hairier if you’re running CPM campaigns as you’re being charged even if no clicks occur. In that instance, you’ll have to ask yourself if the impressions are worth the spend even if no one is engaging with your site.
CTR over 100%:
CTR is an important metric to know whether your ads are resonating with your target audience or not. CTRs on the GDN are nearly always going to be lower than Search (unless you’re doing a particularly good or bad job depending on the network), but one thing that always stands out to me are sites with abnormally high CTRs. Specifically, those over 100%.
Many of the sites that end up with 150% or 200% CTRs usually have low volume, but I’ve also found that for whatever reason, many of them are also lower quality sites.
You can quickly filter your report for CTRs over 100%, give a quick scan, then exclude as you see fit.
One of the benefits of Display campaigns is that they typically have lower CPCs than other targeting options. But at times, there can be some sites that are highly competitive that cause you to pay very high CPCs. Unless these sites are delivering results or are such a perfect fit for your audience that they’re worth it, these can be budget eaters that aren’t providing much value.
A quick descending sort of your CPCs column will show you where you’re paying the highest amount to show up. Vet these placements. If they’re not giving you a strong reason to stay, then they gotta go.
High Cost, No Conversions:
This one is probably pretty self explanatory. If we’re trying to drive conversion performance from our Display campaigns, then placements with high spend and 0 conversions should be a pretty easy one to exclude.
Simply filter your report for 0 conversions, then do descending spend and exclude all of those with so much spend that even if they converted on the next click, you’d be disappointed with the CPA.
Typically, I give a little bit of leniency and let spend of 1.5x to 2x the CPA goal slide, but you can set your own limitations.
Let’s close out by looking at conversion performance.
Similar to looking for high impression placements, I always like to take a look at the sites that are driving the highest number of conversions. It’s a pretty easy filter in the report to include all rows and sort by descending conversions.
Based on the results, I like to look at a few different things:
- How widespread are conversions? Are they coming mostly from a couple of sites or are they pretty evenly spaced out?
- Do the sites that are generating conversions look relevant?
- Are the converting sites of reasonable quality?
- Are there any sites I can leverage further? Add to a placement campaign, use for Custom Intent Audiences, etc.?
Depending on what you find in your top converting sites, there are lots of ways you can leverage those insights to either grow the success of your campaign, mitigate lower quality leads, or expand display strategies in other campaigns.
Converting Sites, Ascending & Descending CPA:
Lastly, we’re going to look at which sites are converting for the highest and lowest CPA, regardless of volume.
The easiest way to find these stats is to filter for Conversions > 0 and then do the sorts just like we’ve done for other views above.
This one is relatively self explanatory. If placements are converting at a CPA much higher than your goal, it’s likely a good one to exclude.
That said, just like with high cost, no conversions, it’s worth giving some flexibility on individual placements. At a campaign level you may be hitting your goals, but within that, you might find some placements with CPAs well below and others well above your target. Having this range on this granular level will allow you to scale, but there is a breaking point.
My rule of thumb is almost always 2x my target CPA for individual placements, but only if others are also converting and offsetting. Based on your individual account performance, you might need to set a different benchmark that makes more sense.
Low CPAs always seem like a good thing, right? But sometimes, just like high CTRs, they can be a sign of questionable activity.
Check out the placements that are converting at the lowest CPAs. Are they suspiciously low or is it just lower than average? Check out the site itself and make sure it’s content suggests a better CPA, especially if conversion volume is relatively high.
If you believe the content might be relevant enough but still question the conversions, maybe try adding a placement parameter to the end of your URLs so you can track lead quality or revenue in your CRMs. Then you can let the data speak for itself whether you should keep leveraging a placement or not.
Reviewing a lot of placements can honestly be a huge headache, but with a game plan, you can at least get the big issues out of the way a bit quicker. Hopefully between this post and the content focused piece you’ll save yourself some management hours as well as improve performance.
What parameters do you use to review placements? We’d love to hear your tips in the comments below!