ellipsis flag icon-blogicon-check icon-comments icon-email icon-error icon-facebook icon-follow-comment icon-googleicon-hamburger icon-imedia-blog icon-imediaicon-instagramicon-left-arrow icon-linked-in icon-linked icon-linkedin icon-multi-page-view icon-person icon-print icon-right-arrow icon-save icon-searchicon-share-arrow icon-single-page-view icon-tag icon-twitter icon-unfollow icon-upload icon-valid icon-video-play icon-views icon-website icon-youtubelogo-imedia-white logo-imedia logo-mediaWhite review-star thumbs_down thumbs_up

How we screwed up online advertising

How we screwed up online advertising Eric Picard

Online advertising is now about 15 years old. Certainly you could claim longer if you look back on what Prodigy was doing in the '80s. But modern online advertising -- essentially display ads -- have been around in the current model since about 1994, when Rick Boyce sold the first banner to AT&T at HotWired. And it's only now that I'm realizing the extent to which we screwed up when we created it.

There are many mistakes I could point to from the early days of online advertising, and mistakes surrounding the formation of a new medium -- and the advertising business models associated to it -- shouldn't be surprising. But in this case, online advertising was the first new media industry that was created in an age of software. Shortly after the first banner ads began appearing on the web, ad servers began appearing -- with the intent of managing delivery and enabling billing.

Since the new sales models and metaphors for online ads were based on traditional media, we should have been OK. But we weren't.

Traditional media sales processes are designed around throughput. It's exceedingly easy to spend a few million dollars across TV, print, or radio. Spending a few million dollars in the online advertising space is hard by comparison. One major CPG advertiser told me that the company's online ad buys are eight to 10 times less efficient than its traditional media buys.

In traditional media, the currency is the gross rating point (GRP) and targeted rating point (TRP). This currency enables immense substitutability of inventory, meaning that there is a way to value an ad across media types -- there is, essentially, an exchange rate. When planners compare ad buys across several TV shows, or several magazines, they can clearly understand the opportunity. When they negotiate a buy, and fail to come to terms on the purchase, they can easily find another vehicle to get the ad in front of the target audience.

You can imagine that in traditional media, the RFP process is pretty straightforward. I'm trying to reach a specific audience, I've contacted you because your show/magazine/radio station has the right mix of that target audience, and I want to buy an ad. In TV, you want an ad in a specific pod of a specific show. In magazines, you want a half- or full-page ad associated with a specific article, column, or section of the magazine.

In online advertising, the RFP process is much more complicated. There is a much greater expectation that the salesperson at the online publisher will be creative, will read a very broad RFP, and return a comprehensive plan to achieve advertiser goals. Essentially, we've made the sales process so complex that the buyer has an expectation of the salesperson doing work that would sit with the buyer in traditional media.

When the first salespeople figured out how to sell online ads and the first ad servers were built to service that market, the requirements for how to handle online advertising were pulled from the sales teams. Unfortunately, media salespeople had no background in providing software requirements, and the engineers building those early systems didn't have the foresight to build much more flexible architectures. The sales models were influenced mostly by magazine sales, but were designed to take advantage of the strength of online ads at that moment -- strengths that should have been short lived, but that were crystallized in code, locking us into a very inflexible set of sales models and metaphors.

We essentially hard-coded the software that manages online ads based on very small media buys with extensive controls at a granular level to manage optimization of campaigns. The second part sounds great, but in reality it causes serious problems. Since we've optimized our workflows and buying processes around small granular buys, it's hard to do anything but that in online advertising. There is frequent discussion about the amount of management required on the buy side during the life of a campaign. But there's very little public discussion about all the work done behind the scenes in hundreds of ad operations teams across our industry, just to ensure campaigns deliver properly.

Next page >>

Salespeople and media buyers in online display advertising have to work much harder than their counterparts in traditional media. In traditional media, there are a few months of work to line up all the buys, buys are executed, then reviewed when the campaigns end, and a few weeks of reconciliation takes place. In online, the planning and buying process is more complex -- lots more work planning and writing RFPs, much more time assembling a comprehensive buy across publishers, and most of the work in online begins once the buy is executed. There is a ton of work required in measuring, optimizing, and managing existing buys during the process of the campaign, both on the buy and sell side. And the billing and reconciliation process is much more complex than in traditional media.

We're at a pivotal moment in the history of online advertising. The legacy publisher and advertiser ad serving systems are showing their age. With the purchase of DoubleClick by Google and Atlas by Microsoft, both companies are investing heavily in re-architecting and rebuilding the advertising platforms. The question is this: How can we change things to ensure that we don't repeat the same mistakes that were made in the past? We cannot continue to support the same models that have existed for the past 15 years.

The critical change that is needed revolves around the currency and requires that we automate and streamline each part of the process: research, planning, buying, selling, inventory management, billing, and reconciliation. We cannot afford to leave things the way they are. Fragmentation of media, new inventory attributes like behavioral targeting, and the massive shifts in the agency compensation structure simply require that things change.

Moving to a model that supports substitutability of inventory has been highly contested by salespeople across online advertising due to the fear of some sort of commoditization of ad inventory. Branded entertainment and highly custom ad implementations have been pushed by sales teams as the "future of online advertising," but it's completely ludicrous. I'm not saying that there isn't a place for that kind of work -- but it's just a small piece of the puzzle. Without massive influx of scalability and streamlining of processes, we're going to be relegated to small media buys.

I would suggest we start by looking at the GRP and TRP models that function quite well in traditional media. Moving to a currency beyond the baseline impression -- one that inherently captures substitutability of inventory and is designed to flexibly evolve, incorporating new targeting and optimization methodologies without fracturing -- is critical. Dave Smith, CEO of MediaSmith, has been a vocal champion of getting the industry to adopt GRP and TRP based models since the industry first started selling ads. I've been having this conversation with him for a good 10 years.

We further have to start investing in automating the sales process, thereby enabling buyers to discover, reserve, and purchase inventory without going through a salesperson. This is highly controversial -- for salespeople. But the reality is that we simply cannot scale this industry by relying on human order-taking and negotiation. Salespeople serve a critically important role as evangelists of the value of a publisher's inventory, and in taking control of that small pool of highly customizable inventory such as branded entertainment and page takeovers. But the friction created in our human sales process is simply too great. And the vast amount of time spent on managing and optimizing campaigns during the course of a campaign must be automated.

The time to move on this is now. New architectures and systems are being put in place at the major ad platforms. We can't wait another 10 to 15 years to fix this.

Eric Picard is the advertising technology advisor to the Advertising Platform Engineering team at Microsoft.

Eric Picard is Vice President, Strategic Partnerships, at MediaMath. He was previously the Founder and CEO of MediaMath-acquired Rare Crowds, an open source ad technology company that provides a completely open advertising technology stack for...

View full biography


to leave comments.

Commenter: Eric Picard

2009, May 14

"online advertising" - I'm not sure that you're commenting at all on my article. Looks to me like spam.

Commenter: Kurt Indvik

2009, May 13

Great summation, Eric, of what I agree is the ultimate pain/friction point of this industry. No matter what scale & targetability we as an industry are able to offer to marketers, until we can make the whole process of planning, implementing and measuring digital campaigns (love that word, substitutability) as easy and effectively in relation to other media, it will be THE stumbling block to further growth for digital.