Wednesday, January 30, 2008

Automation Of Media Buying Through Ad Exchanges: Good Idea! | MediaPost

 Over the last month I've stated more than once that I feel ad exchanges are increasing in their importance, so I felt this week I'd spend a little more time explaining just why I see them as useful.

First off, let me give a little back story. I've been involved in online media since 1995 and I've built online media departments for more than one agency. During that entire time I've stayed true to a few ideals. One ideal is that media and creative cannot be planned separately in the digital space; they are uniquely intertwined and one cannot succeed without the other. Another is that you could not automate the planning and buying of media and run a successful campaign. This is the ideal I am straying away from, now that I know the situation we're in.

I still maintain that planning media must be intertwined with creative, but for the most part the buying of media has become a commodity business. In the old days an agency could maintain that they had leverage over their competitors and could secure rates lower than other shops. That is no longer true. In the old days buying media was not standardized, so the bigger shops could execute larger relationships than the smaller shops because they could afford to throw bodies at the problems; but in today's market, the absence of the bodies affects everyone equally. Meanwhile the business has become standardized and the majority of the media which is bought and sold can be automated, with much of it able to be purchased on a marketplace model with a digital front-end. Salespeople are not obsolete, because they still need be in place to explain the systems and detail the value of one companies' inventory over another -- but Google has proven that the ad marketplace model, or the ad exchange model for managing the buying and selling of actual "tangible" inventory, can succeed.

Back in 1996, Flycast launched and tried to provide a dashboard for online media planners to buy and manage their own campaigns on one network. The Flycast pedigree is seen all over the industry, with the folks responsible for the development, launch and management of that tool running all sorts of companies. Each of these companies has a piece of that promise -- of a digital dashboard enabling more control over the ad budgets and subsequent spend. The Flycast Mafia (as I affectionately refer to them all) had a great idea, but it was about 10 years ahead of its time. Flycast withered away under the management of Engage and CMGI and became part of a business school case study, but the Flycast Mafia model lives on. Companies such as AdBrite, ADSDAQ, Right Media Exchange, Traffiq, and AdECN are all providing services that can useful to agencies looking to minimize the staff managing a client's media dollars, but still maintaining the level of targeting and effectiveness that they have seen over the last 10 to 15 years.

For practical purposes, let's look at it this way: Once the campaign has been planned and the paperwork has been signed off on, the campaign is placed into your ad server. The ad server is now designed -- maybe not perfectly but close enough for today's purposes -- to interface with all sorts of other systems: a bid management tool, an external ad exchange system, and a behavioral targeting system for starters. Companies like Connexion A can build actual dashboards for integrating these into a single location, or you can manage them all on your own. Google Analytics can be layered into your client's Web site to track the post-click activity, and you can manage through to a sale for all search, network and basic portal buys. If you're tracking through to the sale or action you intended and you see conversions happening as a result of a network buy, you access the dashboard of the network and ramp up spending against demographic A while decreasing spending against demographic B in another ad network. You switch over the exchange system for the CPA campaigns you're running, and you allocate a higher bid to the inventory here because you've recognized these efficiencies are not offset by volume. Then you log into your bid management system and you increase the CPC you're willing to spend because you've also identified that at a slightly higher CPC you see volume increases that outweigh the increased CPA you are effectively spending. Overall you increased your exposure to the possibility of spending, but you calculated that a higher CPC of 15% actually drives an increase in volume of 20%.

It sounds very complex, but with a simple database managing the data, you really aren't too overwhelmed! You need to know what to look at, what to ignore, and try not to get too focused on the minor data points that can cloud your judgment. It's basically more like day trading than media buying, but the model works -- and with the tools at your fingertips, you can manage twice the inventory with half the people.

 Of course these examples are slightly theoretical and are based on the idea that your media team is trained on these tools and probably trained in statistical analysis, which is training most media teams haven't received. This requires training and time, which are two things the online media professionals today are sorely lacking. Agencies will need to invest in this education in order to find the efficiencies that will come from these systems, and they will need to focus their energies on the companies that offer these tools rather than the ones who don't.

If you doubt my ideas, take a look at the biggest player in online, Google. Google is spending oodles of money trying to build a marketplace model for radio, TV and print to supplement the one it has built for online and is already using, to command upwards to 60% of the search business. Add this to the portion it is controlling with Ad Sense and its other products, and then see if you still doubt the viability of these models.

And of course the naysayers will state, "You can't automate the integrated sponsorships that are so effective in this market" and to that I respond, "You are correct." However, those dollars only account for up to about 25% of the total online spend, which leaves 75% of the budgets open to this sort of a system, which is enough to make me want to make it work. Don't you agree?

By Cory Treffiletti
Wednesday, January 30, 2008

No comments: