Every day we meet with media companies who spend a lot of time and resources on digital advertising, and yet they leave an abundance of money on the table. Why?
Well, the DevOps group in most media companies are really good at monitoring your sites and apps in order to assess if they are working correctly -- from a technical point of view. But when it comes to ad serving? Not so much.
Part of the reason is that publishers have a supreme trust in the systems they’ve set up. If GoogleAnalytics or Adobe Analytics says this web page received four million unique visitors, few will question that number. Ditto for their ad server. If four million users visited a page, then surely the ad server served up ads for a corresponding number of visitors, right?
Hardly. We’ve seen alarming discrepancies between disparate systems in ad tech. It’s rare for a media company to see identical numbers between its page views or video players and ad tags fired.
Those discrepancies are more than just accounting errors...they represent real revenue that’s been lost to the publisher. If you don’t want to leave money on the table, then you’ll need to pay attention to ad-serving diagnostics.
Reasons for Discrepancies
There are lots of reasons why these discrepancies occur. To begin, it may be an indication that your ad tags aren’t firing correctly, and you may need to optimize them. Or you may have too many ad-tags on a page, and that bloat slows down the page load. So although your web analytics counted the visit, in reality, the user got frustrated and clicked away before the ad (and page) could render.
Viewability, which may discount an ad if at least 50% isn’t in view for one full second, or even the presence of ad-blocking software in the user’s browser
Counting methodologies, which often differ from platform to platform, especially when it comes to exclusion criteria. Is there a gold standard? If so, what system?
Targeting criteria, which may come from multiple sources, including the page itself, or a third-party like Oracle Data Cloud, all of which may be mangled, leading to more discrepancies.
Many of these reason can be eliminated, or at least reduced, and those reductions means higher yields for you. That’s why it’s critical that you look at key metrics holistically across your organization. Discrepancies are everybody’s problem to solve.
What to Look For
There are key metrics that monitor the health of your ad-serving environment. Look for:
Areas of your site and apps where content renders without ads. Once you’ve identified the areas, you can begin to analyze why ads aren’t rendering. Some of the issues, such as the presence of ad-blocking software, require complex solutions. For instance, many sites won’t allow users with ad-blocking software to access their content. The New York Times makes appeals to such users, explaining how it hampers their ability to earn revenue, and requests users to add its site to their white lists.
Some issues, such as problems with ad tags, are easier to address.
Ad unit implementations or page designs that are viewability unfriendly. Many publishers and advertisers use a separate system for counting viewable impression, so right away there will be discrepancies between what those viewability measurement tools and your ad server report. But you still have a lot that’s in your control. Test your pages to see how viewable your ad units actually are.
Latencies. Increasingly, users are referred to an article from a Facebook post, Tweet or email. If that article takes too long to load, they’ll click away, squandering your opportunity to earn revenue. Latency can occur because of:
- Header bidding, which goes through a complex process of selecting the best ad to show to an individual user (where “best” means highest CPM). Although this process is typically occurs in less than 1 second, hiccups do occur, and you need to know about them so you can fix them.
Pixels. These are the necessary evils of advertising. You can’t show and an ad -- and consequently earn money -- without your advertisers’, ad networks’ or other demand partners’ pixels firing on your site. Often, your demand partners “usher” their own partners onto your site, who bring their own pixels to your pages. All of these tags slow down page loads.
Page design. A page may be beautiful once fully loaded, but think of that visitor who clicked on it from Instagram or from a search page...does he or she have the patience to wait?
To recapture all the revenue you’re probably leaving on the table, you’ll need a strategy for assessing whether your ad-tech stack is working as expected; a test that’s completely independent of ad campaigns themselves (e.g. take a hard look at the inventory).
How do you do that? One way is to align datasets, such as your web page and ad server data. Are they relatively in sync or off by factors?
Or you can engage an external service, such as Dynatrace (FKA Gomez), to devise your own tests. Some people even do their own analytics of their logs to assess how the various parts of their ad-tech stack are performing. Of course, this could quickly turn into an exploratory Big Data exercise, but one that is justified given the amount of revenue at stake.
It Takes a Village
Assessing and addressing any health issues in your monetization efforts requires a shared discipline between all parties. One of those parties are the beneficiaries of the monetization, which is typically someone in your revenue/Ad Ops department, but they’re not to only ones.
Often, the solution falls outside of the revenue/Ad Ops group, which may need to engage the folks who are responsible for maintaining or designing the site. We recommend that media companies build a structure for collaboratively analyzing potential issues, so that each group understands exactly what’s at stake, and why it’s important they they fix it.
Advertising revenue is more important now than it has ever been before. And it’s getting harder to earn, as advertisers demand better results at lower costs. This means it’s a downright shame to walk away from money that rightfully should be yours.