The Greycroft team had a great agency experience last week.  We were hosted by Digitas, on behalf of their largest client American Express, over a night of dinner and drinks to discuss the opportunities in social media marketing.  On our end we invited two of our portfolio companies, Klout and Ad.ly, to lead the discussion.  The conversation topics trended from content swarms, to database marketing, to the expansion of the personal web.  At the end of the night Amex walked away happy, our companies walked away happy, and Digitas came out looking great.  The only complaint I heard all evening from American Express was, “I don’t understand why Digitas doesn’t hold events like this more often.”   I thought that comment was a good inspiration for a blog post.

First off, it is worth pointing out that it is not for lack of effort.  Over the last few years we have met with almost every major agency and tried to arrange for connections between their clients and our companies.  Digitas was the first to pull it off, and we owe them a huge thank you.  Our door is wide open to any other agency who wants to experiment with us and our companies.  I have to admit, however, that most large agencies don’t seem all that anxious to pursue this sort of interaction.

One casual observation is that agencies have a hard time interacting with start-up companies.  To understand why, you need to first understand how an agency bills it clients.  A typical agency makes money two ways – through project fees and media spend.  Advertisers spend a few million a year in project fees, for which the agency delivers tangible products like websites and advertising creative.  On the media side the same advertiser may spend another ten million dollars, sometimes a hundred million, on paid media for television commercials and Internet advertising.  The project fees are billed at cost plus, whereas media fees have built in profit margins and scale quickly.  For instance, on a hundred million dollars in media spend an agency can make ten million dollars in contribution margin.  The typical agency supports all the intangibles (creative briefs, strategy, project management, etc) on the margin they make on these two revenue streams.

The effect of new technology is easy to understand on the project side – automating processes reduces billable hours.  On the media side, however, it is not automation as much as the shift to digital that has agencies rethinking their business model.  Over the last five years the online world has developed systems for electronic media buying, which add both precision and complexity to the system.  Take one look at the now-famous Terry Kawaja chart and you can get a sense of the inherent complexity.  This is the world I have lived in for the last five years.  I will endeavor to explain this simply, without using any acronyms:

Google, and a handful of other companies, have built technology that allows anyone to buy banner advertising online.  All you need is an account (also known as a seat) on Google’s advertising exchange.  It is designed to be like trading stocks on the stock market.  Companies have built versions of this product that are simple to work with, as well as versions that need to be run by Math PhDs.

At the moment, electronic media-buying platforms have four shortcomings:

1.)    The systems do not support banner ads with animation or video (known as rich media)

2.)    The systems do not support traditional post-campaign brand measurement

3.)    The media inventory is weighted towards long tail websites and media that was not purchased in advance (generally less desirable inventory)

4.)    It is hard to reach consumers with the required frequency* of most media plans

*there is a lot of math around ad frequency, but for the sake of making it simple the goal of advertising campaigns is to show you an ad enough times that you remember seeing it.  In the online space that means you need to see an ad A LOT, and they need to do it in a relatively short period of time.  That is hard to do.

The major conflict on the horizon is that agencies want to direct media budgets towards their internal efforts.  However, I have seen quite a few start-ups that consistently outperform agency trading desks.  If a company has access to proprietary media inventory and uses advanced rich media (read: Oggi Finogi), than the performance is even better.  The agencies are now at a crossroads where they have an inherent conflict between their revenue model and the relationship with their clients.  And the stakes are high because this is where the agencies make most of their money.

Because of friction in the media buying landscape, I have found there to be an almost schizophrenic approach from the agencies and start-up companies.  Agencies want to meet with new vendors, particularly data vendors that can make their media buying systems perform better, but they don’t want to take the relationships too far in case the technology disrupts their business model.  In practice this creates confusion, and it is sowing seeds of distrust in the digital media world.

Where are we today?

Today, I recommend that start-up companies go straight to the client initially.  This is a tightrope walk because every vendor competes to get traction and inevitably they face staunch opposition from the client’s agency of record.  The goal is that, at some point, a company can reach scale, the agency takes them seriously, and they figure out a working relationship.

I would much rather have agencies represent the companies right out of the gate, but we have tried that a few times and the response hasn’t worked for us.  Meetings get pushed out a few weeks, then a few months, and eventually the company runs out of cash waiting for introductions.

A better solution

I believe there is a better answer for agencies, spurred by the same technology that is causing all this digital media angst.  There are two facts of the new marketing paradigm:  1.) the world is adopting marketing tools that are born outside of agencies, and 2.) the amount of data online is increasing at an exponential rate.  The agency of the future needs to become a hub between clients and technology, with a way to seamlessly plug-and-play vendors, as well as becoming the brain behind next generation marketing by crunching all this data.

There is one company in our portfolio today that solves the technological challenges required for this change.  Last year we invested in a business called Tagman, which on the surface is a platform for managing Javascript tags.  Tagman’s secret weapon is that it also enables agencies to be a hub for vendors and the brain for client information.  Tagman allows marketers and agencies to swap vendors in and out quickly, so clients can experiment with new ad servers, analytics providers, email marketing vendors, etc.  Second, Tagman allows marketers and agencies to collect large streams of attribution data.  Attribution is the science of what leads a customer to purchase, and it is a major opportunity due to the endless list of email providers, search agencies, ad networks, affiliate networks, etc.  Tagman pulls the relevant data out of these systems into a single repository.  In the right hands this data reveals the complete conversion path for all of a marketer’s customers, and it can yield detailed insights into how to optimize that funnel.

Unfortunately, technology is not a panacea for all of an agency’s woes.  Clients want access to best-of-breed solutions, which means that a “not-invented-here” bias has to be eliminated.  Second, agencies need to bill for high value intangibles.  Media dollars are starting to shift towards centers of excellence, and that will leave a lot of large firms exposed.  I don’t profess to know how to solve either of these issues.

Long story short, the pressure is high on agencies right now.  Agency relationships depend on trust, and if clients are getting restless about the perceived pace of innovation you can bet that this will have ramifications in the years ahead.