Monday, September 16, 2013

Man vs. machine, the advent of electronic buying and the death of the media buyer

If you read about what is going on in media land, I am sure you have heard of “programmatic buying” and DMP’s. You may have understood that it is something to do with digital media buying. And you are right, but also wrong.

What is really happening is that all media are becoming digital, as a result of which all media are generating reams of data, real time. And it is this data that is allowing or causing many disruptions across the field of connections strategy, planning, buying, optimization and post analysis.

In this post I want to explore what this means for the role of the media buyer so you can better prepare yourself for the inevitable: the machines are taking over.


(man vs machine visual found here)


Before I gaze into my crystal ball, let's first get you primed on what we are talking about. Here is a link about DMP's courtesy of MediaPost's Joe Mandese. And if you follow this link, you will get a primer about Programmatic Buying. Note that the programmatic buying article seems to focus on digital media properties, but that reflects only today's status.

Here are my predictions/conclusions:

1. Most media buys will become electronic and similar to day trading over the next 5 years

When I started in media back in the mid 80’s, media planning and buying was a service provided by full service agencies. It was in fact their most important source of income, charging marketers 15% of the media cost for all services, including the development of creative executions.

This is of course now very different. Nobody pays 15% anymore, and media is no longer a service offered by full service agencies. In fact, there are basically no more full service agencies, except in Brazil, where by law media needs to be bought by the creative agency. Media agencies are “illegal” (talk about a sweet law for the industry).

Agency groups like WPP, Publicom/OmniPub (or whatever they will become after the merger), Dentsu/Aegis are all clamoring to lower the cost of doing business by developing ever smarter back office systems for the mundane tasks that come with buying, booking and billing media. IPG’s Magna Global are placing probably the biggest bet, or at least are talking about it publicly the loudest. Their quest is to have 50% (yes, five-oh!) of all media transactions automated within the next 3 (yes, three!) years.

Madison Avenue is going to trade like Wall Street. The revolution to automate the stock market happened a long time ago, starting from the 1970's. A friend of mine took me onto the trading floor at Wall Street around the same time as when I started out in media in the late 80's. There were a LOT of people shouting, trading, heckling, waving… it was mayhem, and it was a business between people.

Today most of the trading volume is handled by automated trades. And when I say automated, I mean it is not traded by people but by algorithms. The number of actual traders on the floor has decreased from well over 3,000 by the end of the last decade to around 1,000. And their numbers are still decreasing.


The algorithm based trade is not a fully safe practice (yet), as P&G found out, as did Facebook during their much publicized public offering.

The large majority of trades however happen without glitches or the need for human intervention. It is the same as our planes today, which basically can fly themselves, and the pilot is just there to program the computer and in case something goes wrong during the flight.

Media is following the same path. And Magna Global is betting they will be the ones leading the pack.

2. If you're a media buyer, you better learn analytic skills and/or algorithm programming – fast!

If your current life consists of jockeying excel spreadsheets, buying schedules, orders, bookings, ratings data, etc. chances are you are a media buyer. Your job will become extinct, largely.

It is going to take fewer people who will be able to handle large volumes of media buys and manage the systems that will do the actual trading. It will also become inevitable that many of these trades don’t happen in NY or London. Some may come from the “countryside” (i.e. New Jersey, San Francisco, Chicago or even further afield). Of the top 10 trading firms mentioned in the article linked in the previous sentence we find firms located as "far" away as Florida or Wisconsin.

Media trading might even move abroad to low cost labor markets where there are a lot of university educated mathematicians and statisticians. Any places come to mind?

3. Despite the electronic market, some things will not go away

Before we buy, we need to know what to buy. This means that analysis will continue to be important. And I don’t mean just the analysis of how to optimize the current media schedule for ROI. I am talking about knowing the relative value, contribution, role and place for each of the potential touch points on a schedule. The analysis before the client approves and the buyer hits the “execute plan” button.

People working on that side of the business will utter a collective sigh of relief: their lunches and court side tickets are safe… for now. The buyers will start losing these privileges as the machines take over. There is no point in taking an algorithm to the US Open.

(image from here, where you can buy the original)

A final piece of advice? The data deluge is here. Programmatic buying and DMP's are here. Don't be a dinosaur...



No comments:

Post a Comment