There’s something happening at Facebook.  Almost under the radar it seems that they have established one of the strongest online advertising platforms available which is disrupting the established digital media buying landscape.  A simplistic view of digital media buying is that there are two very clear disciplines, ‘search’ and ‘display.’  Search is 100% platform traded whilst a growing portion of display is now starting to be platform traded via real time bidding (RTB).  This certainly seems very neat and Google have been in pole position to offer an over arching digital advertising management platform, the acquisition of Invite Media being the latest display addition to the existing, ubiquitous Adwords search platform.

However somewhere in the middle of this neatly bisected landscape, Facebook have started to build their own advertising empire, ignoring this equilibrium, and fast creating what would appear to be a completely new channel. In 2008 this quiet revolution started with the release of their self service buying interface for ASU’s (Facebook’s proprietary ad unit) and to complement this came a buying API on which they have allowed a limited number of third parties to build out campaign management tools.  At the start of this year, the walled garden approach started to pick up pace with the removal of all third party banner advertising and the under reported but very significant release of the Facebook conversion pixel.  Add to this an unwillingness to accept third party view tracking (at least for self service buyers) and the platform becomes a hybrid of the existing search and display models that are already prevalent but powered by Facebook using micro-targeted demographic and interest data.  Walled advertising gardens are the preserve of the audience or perhaps more crucially the data rich, and Facebook has both in abundance.

Rumour has it that their standard ‘banner’ CPM’s were incredibly poor with response rates for advertisers to match.  With the new system, advertisers now have the opportunity to tap in to the vast (and bettered only by Google) data treasure trove that Facebook holds on its users and create highly targeted campaigns.  Our experience of advertising on Facebook via their self serve platform is that it performs incredibly well for advertisers and justifies the ‘channel’ label with an emerging trend (in the UK at least) being Facebook specific pitches separated out from the rest of digital media buying.   Revenues are growing exponentially as well, up to $700m in 2009 and predicted to be over $1bn in 2010.  It’s safe to say that with this type of revenue and growth that banner advertising will not be returning to Facebook any time soon and they will push on with new ways to mine user data for advertising purposes all within the confines of Facebook.

So an interesting dynamic is emerging with Paid search, RTB traded display and now Facebook all having bespoke buying systems that are needed to operate them.   A couple of platform companies from both the search and display space backed up by large amounts of VC money are trying to solve these interoperability problems with the vision being a universal buying platform.  However, the further down the walled garden route Google and Facebook go, the more difficult it will be for this to become a reality as data is not portable between these environments.

From a media buyers perspective this is frustrating, but then it’s only from this side of the fence that interoperability make sense.  After all, would you share your data, if you were sitting on the monopolisitic advertising goldmine that both Google and Facebook are or would you keep it behind closed doors? It’s a smart play and the early signs are that Facebook could well be here to stay as a channel in its own right.

There’s been no shortage of informed comment and speculation recently about what Google are planning to do with their latest acquisition - Invite Media. The announcement this week that Google has announced a partnership with Omnicom to build them a global trading desk in return for ‘millions of dollars in display ad spend’ offers more than a few clues as to what Googles strategy may be.

Let me throw one possible scenario out there.

Invite becomes the UI gateway/ optimisation engine for Google’s display business. They integrate it with their dynamic creative solution Terracent (another recent acquisition). They give this away for free to their large advertiser and agency clients (such as Omicom), under the guise of a supply agnostic platform, and slowly start to integrate and sell access to proprietary data and audience segmentations (Google Analytics, and Search data cover pretty much the whole of the funnel)!

It makes operational and commercial sense that they would….

Adsense and Google owned and operated (e.g YouTube) inventory represents a massive slice of supply in AdX, and generates ‘high margin revenue’ for the Big G. They also make money (less) from every other supply source flowing through the Adx (e.g Doubleclick inventory) .  Seems quite obvious that Google would offer media buyers a free tool-set to help them buy more from them. The cynical of you out there may take the view that packaging it up and spinning it as a platform/ supply agnostic DSP makes their clients less suspicious of their motives.

Another reason why Google may want more control of the DSP trading interface market is that it will allow Google to take more control of the audience segments (data) that ad buyers build when they trade via a DSP. Presumably someone at Omnicom has thought about this, and has a tight contract with Google around data ownership and usage.

Microsoft (AdECN) and Yahoo! (Right Media) have got some serious catching up to do. At the time of writing, both are still in closed beta trials of their RTB capabilities. What happened last time Google were given the space to develop a lead like this…?

There are many knock on implication of this recent acquisition and potential strategy outlined above. Here are a few of them….

1 - It commoditizes core DSP technologies and RTB supply integrations.
2 - It changes the exit scenarios for other DSP’s and will undoubtedly intensify pressure for quicker exits from major DSP investors. Will Microsoft and Yahoo! make similar moves for other DSP’s?
3 - It changes the focus of the value proposition for other DSP’s and demand side trading organizations away from predominantly tech focused, towards smart data and service layers (on top of good technology and infrastructure).

I may be totally wrong about all of this, and no doubt Google are already thinking 15 steps ahead of anyone else, but it will be interesting to see how things play out in the next 6 months.

Picture the scene. It’s 2004 and Scott Ferber, one of the founders of Advertising.com, is giving one of his inspirational companywide speeches via WebEx. The slide is showing “Ad.com 2.0” when Ferber delivers his vision: “We want to be the eBay of online advertising!”

He went on to explain about how their AdBid technology, which allowed customers to set their own CPMs, CPCs and CPAs every day, had ended up increasing rates as customers saw for themselves the effect on volume as each rate was entered into an auction for each impression. The natural progression for this technology was setting up an advertising marketplace where buyers and sellers could come and trade inventory, setting their own rates depending on volume and ROI goals and all benefiting from AdLearn, Ad.com’s proprietary optimisation algorithm.

Fast forward to late 2007 and I’m sitting in a Right Media introduction meeting whilst working for Media Contacts (Havas). They talk about marketplaces, buyers and sellers trading inventory and auctions for each impression, all benefitting from a unique optimisation algorithm. All the agency people around me are becoming excited and start discussing how best to make use of this new concept. I’m sitting in the corner looking at my peers wondering where they’ve been for the last 3 years.

Then it hits me. #1 Ad.com never ended up releasing its marketplace to the masses. Whilst it had the technology, it instead used it to bolster its own network preferring to keep the concept proprietary. #2 as a result agency people had never been given the ability to buy inventory in this way before. This 2nd revelation turned out to be a bit of a curse as well as a blessing, but that will be the subject of another article…

So fast forward again to 2009, Real Time Bidding is all the rage, and I’m once again scratching my head wondering what all the fuss is about. Surely exchange bidding is real time? And if not, surely this is a simple natural progression of the technology that would happen behind the scenes with no big song and dance, after all, exchange bidding had certainly been sold as real time up until now. Then I read Mike Nolet’s excellent blog on the subject and, happy to be away from marketing hype and down to some proper techie speak, he provided the gem that made everything fall into place.

No more redirects. Proper, back end integrations. Smart software running on big iron. It all seemed so much more elegant than our current discrepancy ridden, chaos ensuing redirect methodology. By doing all the complicated stuff in data centres housed in military style bunkers rather on users’ desktops, you not only make the whole process much more efficient, but through the easy and open exchange of data, companies can become truly specialist in a particular niche. Instead of all adservers having to have their own retargeting technologies, you can just plug in someone else’s data. Similarly behavioural companies will not have to build their own adserver in order to sell their data.

Mike Nolet talked to me about the problem with the Right Media ecosystem. For it to work, everyone had to be using Right Media. This was the same with Advertising.com’s version a few years previously and both these companies have suffered as a result. RTB is different as no one company “owns” the standards that build this ecosystem. Much like the growth of internet, the use of open, agreed standards should provide the basis for every kind of company that wants to trade in the ecosystem to flourish, instead of being held back by a poor adserver, reporting system, or other commoditised technology service.

So what’s holding it back? Well until a critical mass of companies buy into the concept, it will remain niche. Up until now, trying to be the one stop shop for online advertising was the goal for many a network (and agency for that matter) and this may take time to unlearn. However, much like Microsoft, the once great monopolist, is now seeing the benefits of using open standards across at least some of its products, so too will networks, data companies, behavioural specialists and dynamic creative providers (to name but a few).

Of course, these companies could follow Apple’s lead and try to keep everything proprietary. It certainly works for Apple (most users do not care that they are locked in to a single platform since the product is so good). But are there really any online ad companies that would boast that their products / services are as good as Apple’s? I await with interest…

Martin wrote an interesting piece recently for AdExchanger on the European 3rd party data market.

On a recent trip to New York, Infectious Media met with Adexchanger amongst others to chat about the exchange space.  The sector is growing rapidly in the UK at the moment, with advertisers, agencies and publishers all showing a real appetite to grow their involvement.

One of the biggest differences that surprised everyone we spoke with in the US was to hear that exchange liquidity and more specifically inventory quality was a problem in the UK.  To illustrate this, I’ll give you a tale of two premium publishers that we’ve been dealing with.

The first has gone in to the exchange space head first.  They are not someone who we would have expected to be trading on exchanges, and if they weren’t, we wouldn’t be doing business with them. They have dedicated time and resource to understanding how it works.  They have made some mistakes, undoubtedly sold some of their inventory for a much lower price than it’s probably worth, and I’m guessing it’s been a real headache for them.  But over time something magical has happened, the volume we buy from them has increased steadily as have the CPM’s that they are receiving for their inventory. From our perspective they have proven, through a strong ROI, that their inventory is good quality, and in a depressed and over supplied market, this is a great story.

The second publisher has taken a different approach to exchanges.  They constantly flirt with the idea but channel conflict and a fear of low CPM’s have dominated their thinking and hamstrung their entrance into the space.  In a same depressed market they publicly maintain their site is fully sold out every month at an astronomical CPM. We’ve yet to do any business with this publisher and I’m not sure when we will.

So as the first publisher gains learnings and future proofs their business, the second is watching the market change around them and doing little about it to protect their (extremely) short-term yield.  When you examine things more closely, there have been a number of recurring themes that emerge from publishers we have talked to that have held them back, these are in no particular order:

1)   High entry costs.  Many publishers have been put off by monthly minimum charges and long contracts
2)   Low CPM’s. There’s a perception that CPM’s on exchanges are very low, our experience is that good inventory still elicits good CPM’s
3)   Skillset.  Some publishers put their toe in the water but couldn’t make it work for them as they didn’t have the right skills in-house to build a platform trading business.
4)   Physical distance from the companies and people making this happen, predominantly the US.  Even the UK sales teams of large corporates that own exchange platforms have barely heard of them or have any awareness of the tidal wave of change that is coming.
5)   The economy.  No one wants to make a costly error for their company that could make them look bad so everyone sticks with what they know.
6)   Education, lack of.  Exchanges are creating a real buzz in the UK at the moment but there’s a chronic lack of education amongst publishers about how best to engage with them.
7)   Channel conflict.  Publishers are scared to sell their remnant direct for fear of cannibalising their premium sales.  Ironically, one of the principle roles networks are fulfilling in the UK is one of anonymity for publisher inventory.
And finally…
8)   Rumours and misinformation circulating about all of the above.

However, two big developments should change this over the coming months , accelerating adoption.  Firstly, the easy integration of DFP customers into the Google Exchange and secondly the aggressive arrival in the UK of the publisher yield optimisers. Both of these should help bridge the education and technology infrastructure challenges, bringing many more premium UK publishers up to speed.  More than this however, it is adopting a mindset of innovation and embracing change that will differentiate the winners from the losers as the new landscape emerges.

As technology and data shape the next wave of digital media buying, the need to approach the entire process from a statistical mindset is growing in importance. In the not too distant future it won’t be surprising for a day in the life of a digital media buyer (or trader should I say?!) working on a response brief at a typical agency to involve monitoring campaigns on multiple trading systems whilst working with a team of analysts, crunching campaign data using analysis packages such as R.

Google Chief Economist Hal Varian in an interview for The McKinsey Quarterly earlier in the year mentioned that statisticians will be the sexy job in the next ten years and with Numbers pulling in prime time audiences on TV it certainly seems that he’s not too far from the truth. This could happen a lot faster in the quickly evolving media buying industry and gazing into a crystal ball reveals that skill sets will also need to also evolve to keep pace with the technology infrastructure that is powering the industry.  Gone will be the Head of Trading whose lunch ability knows no bounds to be replaced by the Head of Data whose skill and value will be in interpreting the vast amounts of data that agencies will be generating through their trading platforms, and developing meaningful insights from this. Think of it as being more Wall Street (minus the suits of course!) than Mad Men. All in all, a combination not just of statistical techniques but also marketing know-how will be an integral part of the new agency product. It doesn’t take a huge amount of insight to spot the increasing importance of data but what does it mean on a practical level for your average digital media agency? Using statistical techniques to help analyse and optimise campaigns has multiple benefits ranging from improved ROI for advertisers through to greater operational efficiency for media agencies. In short it’s in everyones interest.

Infectious, as a ‘new breed’ agency has been built around this core belief. Our underlying infrastructure, skillsets and services are designed to get the most from this new world. It will be interesting to see how quickly the wider agency community embraces this shift.

There’s an article on Business Insider today which in turn quotes a great Business Week article from 2000 that insightfully asked ‘How will Google ever make money?’

The crux of the article is this:

‘The company’s adamant refusal to use banner or other graphical ads eliminates what is the most lucrative income stream for rival search engines. Although Google does have other revenue sources, such as licensing and text-based advertisements, the privately held company’s business remains limited compared with its competitors.’

This is interesting for a couple of very different reasons.  Firstly, as a historical reminder that Google flipped the industry on it’s head in breathtaking fashion.  They did away with the predominant image based ad, borrowing the business model of Overture to serve text ads linked to an auction pricing model and took something new to market.  At the time (as a media buyer) I certainly didn’t believe it would work, preferring, instead, the glamour of AltaVista (RIP) banner advertising to the three line text ad on Google.  Fast forward to 2009 and Google is the biggest ‘media owner’ in the world, generating $6bn of advertising revenue per quarter.

The second interesting point to come out of the Business Insider article is linked to comments made recently by Eric Schmidt, speaking at the Morgan Stanley technology conference.  When asked about where Googles growth in revenues would come from, he replied:

‘Where is [our] next source of revenue? [The] next source is current business functioning better. Next and adjacent is a set of display businesses and an exchange being built from DoubleClick business.’

A slight change of tactics from their stance back in 2000 but the reality is that the market for text ads is maturing and growth is levelling off.

There are few details on the Google exchange product but this time they have bought the business that they will use to to leverage display (Doubleclick), instead of borrowing the model and being sued at a later date. So what could this product look like, given that Schmidt sees it as so crucial to revenues going forward?  Well, instead of thinking about Google as a purveyor of text ads on the search results page, think about them as the largest owner of intent data in the world through their search engine.  One way they’ve used this intent data is to serve text ads against search tems, but this is only one execution of an ad against a hugely powerful data set.  Through a display media exchange it becomes possible to intelligently  serve display ads against that data on any site, which will open up a huge new market to Google in highly targeted display advertising.

To make this a reality, a big challenge for Google to overcome is the issue of data privacy. Two weeks ago Google made a significant announcement around a basic ‘behavioural’ targeting product across You Tube and its Ad Sense network. It was only a matter of time before this happpened, but they have clearly gone to great lengths to introduce it in a simple, (relatively) transparent and cautious way. The media whipped up an ‘information privacy’ storm around Phorm in the UK, and it seems that Google may have been waiting for that to die down a little before dipping their toe in the water in a very controlled way (with plenty of user control and opt out-ability). The obvious omission from the announcement was the use of Googles search data in the targeting formula.  Make no mistake, this will come in the future (Yahoo! announced their own Search/Behavioral retargeting solution last week), but Google may be waiting for people to become comfortable with this simple behavioral model first, before ramping up the sophistication with the reams of data they hold.

Much of this is still speculation, as things are heavily under wraps, but the Google Exchange isn’t too far away and there’s little doubt they have both the data and the distribution channel to make this the market leading product in display, flipping the market back on its feet and with display advertising on a more even footing once again.

We’d love to hear what you think…

I like data.

It has stood me in good stead in this industry as I’m sure it has many others. Our campaigns produce tons of the stuff and increasingly agencies are reliant on companies that process and present this data to push our thinking forward such as Atlas and Doubleclick.

But data isn’t (and shouldn’t be) the be all and end all. It’s easy to forget that data is a product of something that has already happened and as media planners we’re often trying to find something new, especially in digital. Take search for instance, the ultimate data driven medium. We analyse search patterns and use algorithms to determine bid strategies; all based on what has happened in the past. There’s nothing wrong with this but it can be a dangerous obsession and lead agencies to a one dimensional and commoditised offering.

To illustrate this point, I’ve seen and given hundreds of presentations that are based around a marketing funnel. Where data driven techniques are at their most potent is in pulling customers through that funnel and converting demand that is already there; think PPC search and retargeting. But this is an ever decreasing circle as it does nothing to actually create demand for a product or brand, it is just squeezing the most out of demand that’s already there.

Advising a brand on how to act strategically to create demand requires a different type of thinking. This is insight driven, takes an understanding of a advertisers business, their product, their brand, their market and shock horror: ‘people’ and how they behave. There shouldn’t be a spreadsheet in sight for planning purposes, no formulas or macros and definitely no definitive answers. Overlaying and translating all this thinking into media terms is a wonderful exercise and for us is what media planning is really about.

So as a digital media agency, the skill sets you need for a complete offering are hugely diverse and need truly different people, but that’s what agencies should be about:- a collection of people with different specialisms that integrate seamlessly for a client. It can be difficult to reconcile some of these skillsets but these are our challenges. Data skills are clearly hugely important but if that’s all it’s about then perhaps it’s time we started employing monkeys and bought some typewriters.