To Protect and Serve: Buying and the Third-Party Ad Server

Since the dawn of online advertising, there has been a quest for standardization. At first there were only a few on this trek, but now countless self-proclaimed knights and ne’er-do-wells have saddled up and combed the countryside in search of the Grail of Webvertising.

Until only recently, much of the web has been a Tower of Babel in terms of what to call things, how to quantify them, and how to report on the information those things provide.

The industry first began with naming conventions. What do we call things on the web? Hits was all there was at first, a terribly inaccurate measure of what was going on at a particular web site.

Soon there was the emergence of third-party auditors who began to bring some commonality to the words the industry used to describe the online advertising space. Groups like I/PRO and NetCount (R.I.P.) began to standardize and simplify how we talked about web site activity. Terms like pageviews and impressions were injected into the webvertising vernacular. These labels started to bring sense to the web, giving agencies and potential advertisers something to begin quantifying.

Third-party auditors like those mentioned above and others, including stalwarts of the print world like ABC Interactive Audits and BPA International, now had something to put in their monitoring sites and actually audit.

But this didn’t end up making anything clearer for those of us trying to read the results of our online advertising. Though the arrival of third-party audits helped to bring standard terminology and an appearance of propriety to a web site’s reported traffic, it didn’t do anything to bring standardization.

Most web sites were using, and still use, their own proprietary ad-serving technology and software. And these technologies all record their data differently. And they all report that data in a variety of forms.

Advertisers began to find themselves confronted by inconsistent data from a multitude of sources. Month to month, sites on which advertising was being run would report wildly variant results. In addition, there was more and more data that needed to be managed.

Though the industry could not agree on a standard count for certain basic online activity (and still has not), it had come upon a myriad of reporting formats for this uncertain data. So not only were advertisers being buried under piles of apples-to-oranges data, it was coming to them in dozens of different baskets.

Enter the third-party ad server. What the third-party ad server promised was not only centralized campaign management, making collecting and reporting data easier, but, for the first time, an apples-to-apples comparison of campaign activity.

Even if the data were somewhat inaccurate, the counting methodology was the same in every instance, meaning that my cross-site comparisons would be flush. Finally planners and buyers had one place that could show them how the campaign performed and how each property on a buy related to one another.

But did this bring standardization and, thus, peace in our time? Far from it.

Since most sites already had technology and software in place that they endorsed, they were reticent to accept outside technology for the serving of ad banners on their sites.

And when it comes to reconciliation of activity at month’s end, advertisers now have a whole new problem that will erode any time efficiencies gained from the use of a third-party ad server.

As is the case with most sites, advertisers and publishers reconcile activity and thus billing against what has been delivered. Either a short pay or a makegood are the result of reported underdelivery.

An overdelivery results in a happy advertiser and quite possibly a renewal. But when a third-party ad server has been used, there are often discrepancies between site-reported results and those reported by the third-party ad server. Though the possible reasons for these discrepancies are many, to date there has been no satisfactory conclusion. Or resulting solution. Discrepancies may vary from a few hundred impressions to a 50 percent difference.

So, if you are using a third-party ad server to handle a campaign, what do you do about these issues?

Here’s the deal.

First, make sure the use of a third-party ad server is acceptable to the site during the RFP process. If you are going to use one, you might as well know upfront if a site you want to do business with will accept it. If at first the site says no, follow up and talk to them about it. You’d be surprised at how many don’t actually know what a third-party ad server is, and once you’ve explained it, they will change their mind.

Next, find out from the site what kinds of reported discrepancies there have been with other third-party ad-served campaigns. This will also help to determine whether or not you will run with the property and what you can expect if you do.

Finally, negotiate upfront what kind of discrepancies will be acceptable and how both site and advertiser will reconcile against them. There are invariably discrepancies between a site’s reported figures and those reported by a third-party ad server.

Remember that your efficiency analysis depends on real costs, so be sure you know ahead of time just how those costs are going to be determined. I’ve used ten percent as the standard acceptable difference between site and third-party ad server, but sites like bigger margins. See if you can negotiate a “split the difference” scenario for a given discrepancy. Say, if the site reports 100 percent and you report 80 percent, then you agree to 90 percent.

There are other ancillary items to be aware of when negotiating the use of a third-party ad server with sites, such as what kinds of creative units can be served into the site. Will they accept redirect URLs? Will they move around third-party ad server tags on the site to optimize the campaign?

These are just a few of the things to be aware of when you are buying sites and using a third-party ad server.

Related reading