July 24, 2013
There is an enormous amount of information in a placement report – some may even say an overwhelming amount of information. I’ve known some overzealous individuals who have spent full days on a single placement report, emerging on the other side bleary eyed with an enormous list of sites to exclude and a deep understanding of the weird sites that have opted into the GDN. That’s too much time to spend on a placement report. I’ve actually found that by planning on spending a relatively short amount of time on my placement reports I’ve increased efficiency and the performance of my campaigns.
According to the blog post introducing placement reports from over six years ago, “the Placement Performance report provides site-by-site performance metrics — including clicks, impressions, cost, and conversion data — for domains or URLs in the content network where your ads have appeared.” Even though the Display Network has undergone countless changes since this announcement, the placement report remains largely the same. It’s a place where you can see site-by-site performance metrics for wherever your ads are appearing on the Google Display Network.
It’s nearly impossible to police every site that your ads are appearing on one at a time, so being efficient in the ways in which you use your placement reports can go a long way toward making your display campaigns as streamlined as they should be.
Here is the process that I follow when running through a placement report to maximize performance for my display campaigns.
Download an account-wide placement report on the Display Network tab for however long a date range makes sense for your account. For the most part I use the last 30 days’ worth of info, but that range should be longer for lower volume accounts. You’ll basically want enough volume to be able to judge your placements on your standard CPA/KPI goals as you would with search keywords.
As you can see in this screenshot of which columns you can add, ad group and campaign are both required to be in there. Those are your two options when it comes to excluding placements, so it makes sense for those to be the determining factor in how performance is reported.
Before you start deciding what to exclude, identify what your motivation is for excluding sites. Are you going to be mostly performance based? Or are you trying to protect your brand by avoiding some of the weirder sites out there on the GDN? I tend to focus on CPA over brand protection (because I love leads JUST THAT MUCH), so this post will for the most part focus on improving performance for performance’s sake. Eric’s post about excluding categories on the GDN is a great way to cut out a lot of stuff that may damage your brand, and if your main aim is brand protection that’s a great starting point.
My first step is to filter my spreadsheet campaign by campaign and see at a glance what’s wasting money for me in each ad group. I sort by cost and find what’s spent well over CPL goal without converting or what placements just have a high CPL(using our oft-cited CPL matrix).
Heading into a placement report, I plan on mostly adding in ad group level exclusions. Depending on your targeting, you can see vastly different performance from one ad group to the next. For that reason I tend to give each placement a chance ad group by ad group.
After going through each of my campaigns and finding placements to exclude for each ad group within those campaigns, I’ll next look for outliers. These are the instances where CPL doesn’t set off any alarm bells, but other metrics may. It’s about finding sites that are doing something weird, and for that reason I filter my performance campaign by campaign so I can establish somewhat of a baseline. It’s hard to know what weird is without knowing what normal is, so don’t let different display campaigns make you think poorly of what could be a perfectly good placement.
Here are the different statistics I like to look at:
You could lose a lot of impressions by excluding really high volume sites, but you could also see some big improvements in CTR. It’s all about deciding which metric your campaigns are in service of. This could also be a cue for you to take a look at those sites and see where the ads are actually placed. If everything is well below the fold, this might be a way for you to exclude traffic that doesn’t really have a shot of getting to you.
Are the sites with above average CTRs legitimate? Or are they a candidate for click fraud (or at least something very close to that)? Make sure to filter your spreadsheet to more than just a few clicks (I use greater than 10 clicks or more). That way you can see percentages that are a bit more meaningful than 100% or 50%. Is the CTR that much better because the site is just a perfect fit, or is the ad placement on the site possibly misleading? You don’t want people clicking to your site by mistake (or at least I firmly believe that you’ll always want users to buy in voluntarily and not via confusion).
Just as different keywords have different qualities, different sites have different qualities. Especially if you use CPA bidding (as I do in this example), you can see some wildly different CPCs. Are the conversion rates high enough to justify the high CPC? Is the site worthwhile enough to be paying that much per click? Take a look at the top CPC sites in your campaigns to ensure you’re spending wisely. Even if it’s just a few clicks now, at that CPC it could get out of control quickly.
You’re looking for sites that are trying to sneak cheap clicks into your accounts. I actually didn’t have anything exclusion worthy in this campaign (based purely on conversion rate), but there were a couple of alarm bells. I’m going to keep an eye on those sites, and if current trends hold up they’ll probably show up with one reason to exclude them or another.
Following this process I also usually take a look at stuff with a bigger lens. All of these optimizations were ad group by ad group, but I want to see if there are any plain bad placements in general. There are two different ways you can do this:
1. The dimensions tab. Add in your metrics, take a look at campaigns without ad groups and then take a look at placements without campaigns. Use these numbers to find campaign- or account-level negatives.
2. Do this same process, but with a pivot table. You can do this even within your same spreadsheet!
Once you get used to this method you should be able to do an entire placement report in an hour with a heck of a lot of insight (at least in my opinion) into what’s actually going on in your account. You’ll know which of the thousands of websites in your reports are worth further investigation. You’ll also even have some easy to make projections and AdWords Editor Uploads with your campaign info already on hand.
How about you guys? Do you have any tricks for making sure your ads are only appearing on the best of sites?