"Data-Driven Thinking" is a column written by members of the media community and containing fresh ideas on the digital revolution in media.
Click To Enlarge
It’s that time of year again when business and trend forecasting articles start hitting the magazines and blogosphere. I’ve contributed a blog post each year since 2006 about future trends, and I’m going to change it up a bit this year. Historically, I’ve aggregated many different trends into one cohesive resource, but this time, instead of being the “techmeme” of trends, I’m going to actually contribute one.
If I’m wrong with this forecast, I’ll be drowned out by all of the other forthcoming opinion pieces and you won’t remember mine. My prediction is that because the API-driven buying platforms that have proliferated in the past few years are overwhelming the market, in 2011 we will see these platforms drop out, be consolidated, or be acquired by agencies. Furthermore, businesses will begin to build up the services associated with the platforms in order to create more value. Note: these are my opinions and not necessarily reflective of the companies by which I’m employed.
2008 through this year  was all about the API-driven buying “platform.” We’ve seen platforms label themselves as Demand Side Platforms (DSP), which is a much better self-imposed name than “Behavioral Targeting,” and quite a few of these platforms have bloomed. Invite Media, TerminalOne, Trade Desk, Turn, Triggit, AppNexus, DataXu, X+1, LucidMedia and others have all built (or pivoted) platforms that are, in some cases, both self- and managed-served. There have been quite a few acquisition rumors over the past few years about DSPs, but we’ve really seen only one come to light: Invite Media to Google. There are some rumors and bets that AppNexus will see a Microsoft term sheet in the near future, but that’s pure speculation at this point.
The DSP business model is simple: service a client, either by letting them use the platform themselves (“Self Service”) or by doing all the work for a client (“Managed Services”). This model has created conflicts with some of the DSPs’ clients as the DSPs could theoretically compete with their clients. I won’t go into that too much here, but that’s an offline conversation we could have.
The promise of API-driven buying (inclusive of exchange, RTB, etc.) allows marketers to match data to a media impression that delivers on a) determining bidding value, b) forecast performance, c) decision whether or not to bid, and, if so, d) bid/buy. This is all accomplished through a technology “platform” as described above. As technology has penetrated Madison Avenue’s walls, this has opened up much opportunity around both agency and client value creation and deliverance.
The select few (but growing) agencies that are taking advantage of this new paradigm of buying are becoming truly technology-driven shops. Bertram Cooper wouldn’t get it, but that’s OK--Harry Crane probably would.
While these new platforms offer great opportunity, I believe we are currently seeing a glut of platforms. Look at the now-infamous Terence Kawaja chart of Display Advertising Technology Companies. There are more platforms than there are agencies. The issue with this is that in order for a platform to build a BIG business, they will need to be leveraged by multiple agencies (and brands), so with a glut of platforms … the supply & demand curve doesn’t work in their favor.
My prediction through the end of 2011: consolidation of DSPs at lower valuations, agency acquisitions of financially constrained platforms, a few joining the dead pool, Managed Services at DSP reigning king (especially to brands directly) and pivoting of business models (again). We also will start to see all of the search bid management platforms adding display capabilities.
Networks of course are another story: I’ve probably been too harsh on networks publically but still feel that I have good reason to be. The trend we’ve seen from ad networks is that a few of them are morphing into DSP businesses, but I think that we’ll see quite a few of those turn into 2011/12 Trading Desks. That is a bold prediction, but it is based on my predictions for current DSPs outlined above and what you are about to read below.
The Service Business
To really succeed, DSPs will need to provide good services rather than merely technological advances. Noah Elkin (@noahelkin), a Principal analyst covering mobile over at @emarketer recently Tweeted (11/3) a very insightful statement: “Fragmentation great for innovation but maturity brings need for servicing entities (opp’ty for agencies).”
I don’t think that Noah was intending for me to use this quote outside of the mobile space, but it certainly has meaning to the adtech world within our ecosystem. Platforms are exactly that: platforms. Platforms in themselves are not social, they don’t attract or build client relationships, they cannot say they are sorry, and they are limited to only what they are programmed for. Platforms are an enabler, but the team and business built around them drive ultimate value creation.
Now that DSPs are getting embedded into the daily activities of the Advertising ecosystem, I predict the next few years are going to be focused on building world class trading desks (internally) and driving auxiliary companies and freelancers in specialized fields such as Algorithm Design, Data Visualization Specialists, Market Analysts and Data Management. These services will allow them to build a stronger client base and create more value.
Trading Desk: This currently exists as separate entities within the large holding companies, and this is where people turn the levers and knobs of the platforms to drive the best performance for clients possible. Generally staffed by an Account Management, Yield Management, Trading Specialist, and Business Development teams. Trading desks currently range in size from 4 through 80+ and sometimes lean on partners for delivery of specialized areas of expertise. The trading desks may maintain their specialized positions as next-gen buying desks (to go well beyond display buying) or may morph into SWAT teams engaged at actual agency-levels to drive collaboration and understanding to all who work within agencies.
Algorithm Design: Get ready for the term Algo-Specialist. If we’ve thought MIT and Stanford have graced the AdTech ecosystem already, get ready for people strutting their rocketship-engineering, DNA strand-testing, scientific super duper algorithm-design credentials. As Platforms make available the BYOA option (yes, that’s Bring Your Own Algorithm), we’ll see Trading Desks investing in different algorithms and testing their performance across different campaigns. While there may be a website or app-store dedicated to buying different algorithms for the different trading desks, you will also see custom-written algorithms by client/campaign before launch and Algo Specialists on staff (or freelance) to tweak for performance. Nerds.
Data Visualization Specialists: If you have read my blog or my published articles in the past, you will know that I’m a huge fan of infography. I think I even coined that term. As our world increasingly needs access to lots of data quickly, and our traders (and account teams) will need to make rapid decisions, visualization will be key. This is an area where Don Draper might get very excited. Trying to figure out data visualization dashboards for trading desks is going to be lots of fun, and I bet many firms will use history as a guide and build similar terminals to those from the financial markets. Data Visualization will be needed not only for making trading decisions but also around targeting and insights opportunities. As we move to use 3rd-party (and even learn about our 1st-party data), consumer insights can be gleaned, and we can become much smarter about whom we are targeting and whom we should be targeting.
Market Analysts: Research. The stock market moves either up or down based largely on what research analysts say about certain stocks. Think about the ad exchange and private exchange ecosystem: a ton of research will be created. Will Trading Desks create Institutional Research houses? Quite possibly. I do not know the exact format in which they will exist, but I know our roadmap for 2011 will include some people focusing solely on understanding supply, demand, and all factors that influence data, media pricing, and availability.
Data Management: It’s not new, but we’re going to see increased presence of it. IBM already made their move by acquiring Neteeza, but we’re going to go beyond actual servers and storage into software layers that help manage the glut of data that we’re all creating. This will go far beyond media impression level data and into total data management that will allow us to model and forecast/predict potentially, dare I say, “media mix” models across many channels as, in an ideal world, all media is bought and sold through trading desks. Ideally the data management layer will interface with the buying platform and visualization, but all three could be independent.
As businesses mature, scaling revenue, staff, and building a real business is the focus, as Noah mentions above. These businesses have significant client relationships, partner relationships, and internal processes, and we’re going to see the shift in focus from the platforms to the Trading Desks over the next year. I’d have to think that the hundreds of Ad Networks who didn’t position themselves as DSPs would pick up the Trading Desk moniker.
My friend Mike Nolet of AppNexus fame challenged the term DSP and vowed to change it to Display Engine Marketing with the obvious reference to Search Engine Marketing and the parallels of each of these. I think that Display Engine Marketing, while it works, sells this entire new paradigm short. GoTo/Overture pioneered the biddable search business, and it’s taken 12 years to seriously move this model into the display business. I don’t know how many years it’ll take to get into television, print, radio, and OOH, but it’s coming. This is, at least, my vision of the future--maybe yours too?
Using vast amounts of data to purchase and optimize media and creative requires a full infrastructure that can be owned or licensed. The platforms have been developed, but a lot of work remains to develop the business and service aspects of the infrastructure. We’re in the early innings of game one of the World Series, and once all the pieces of the ecosystem fall into place, we’ll be able to deliver on the promise of putting the right message in front of the right audience at the right time with consumer buy-in.