Rob Leathern is CEO of CPM Advisors.
AdExchanger.com: CPM Advisors... so I take it you believe the CPM is not dead?
RL: CPM Advisors ("CPMa") typically works with advertisers who are optimizing to a CPA or CPC metric, but realize that they cannot get volume and scale unless they can tap into CPM advertising options. We are smart about how and where to find quality inventory at fair, market-driven prices. Today, that is driven largely (but not exclusively) by the exchanges and some of the aggregators.... we want to automate the process and systematize the knowledge. CPM Advisors - as a name - seemed to capture that ethos fairly well, though we are certainly not wed to the notion of CPMs in the long-run.
The current online advertising system does all come back to a CPM and will for a while. A publisher or network has to assess any CPA or CPC deal and compare it to what they will realize on a CPM basis, and CPM is thus the "base currency" for selling space online. I do believe that the fractional-user-attention "impression-based" CPM as we know it should have been dead by now if true innovation had occurred in the online ad market over the last 5 years, but it isn't gone and won't be for quite a while: changing the way people do business always takes time. Audience- or attention-based metrics are still a ways off.
How has your buying evolved in the last 12 months? What about the next 12?
As we've built out our software and systems, at the same time running billions of impressions and millions of dollars of ad campaigns for clients, we have continued to become smarter and figure out more and more ways to automate the different parts of the media buying and optimization process. For CPMa, it's an exercise in continual learning and improvement as it should be.
We will continue to build on our early success. For the next 12 months, there should also be several new exchanges that finally become commercially interesting. The more sources of good-quality inventory that become available with API interfaces the better for everyone, not just us. Display advertising is fragmented and inefficient. Since the learning curve to even the most basic display ad buying system is steep, via the economies that exist in multi-system integration CPMa will grow further as a buying option for advertisers and agencies.
Are you using your own buying platform or using a third-party? Any observations regarding who will win the "buying platform" race and why?
We have our own buying system, which is now in beta at http://www.cpmatic.com. CPMatic is a technology-based ad buying service with an interface. I think that there are/will be companies with multi-exchange buying ad servers, for people who want to tweak dials and levers themselves, and this is a backend component of what we have as part of our service. We provide a front-end that allows the user to set up their campaigns, upload creatives, get reporting, retargeting or conversion tracking tags and so on - and that's helping solve a lot of the campaign initiation inefficiency that exists today. Then of course, very important for us and the advertiser are the smart bits on the back end: the technology to figure out where to set up ad campaigns and how, how to achieve consistency across channels and optimizing frequency and spend across those channels.
Figuring out the differences between what can and can't be done in the various ad systems is not trivial, grabbing reporting, making changes, tracking budgets, propagating creatives and knowing how and when to do it. It's a difficult set of problems encompassing both integration and optimization. What I helped build at NexTag is quite relevant to my view on things: in creating an adserving, media management, analytics and creative management infrastructure, we were able to become a top 5 online advertiser by spend - automating as much of the display media buying and optimization process as possible to a point where the company could have statisticians looking at data to decide whether to run certain media buys. We got the optimization piece put together, but back then our 'integration' with third parties was mostly scraping reporting data from publisher adservers so we could track costs accurately, and API opportunities for buying media in an automated way make the space much more exciting now and allow a company like a CPMa to exist.
I'm not convinced it's a winner take all (or even most) market: the market is so big that there are going to be a number of players who can cater to differing marketer and agency needs, and whose techniques work and have application in different subject areas. I think the biggest race and one we all need to lend a hand in (and I applaud adexchanger's part in this) is to educate our clients and potential clients about how the marketing world is moving to a new kind of technology-driven accountability, and help move more dollars online.
Is there any value in placement anymore or is it all about finding the right audience?
There is absolutely value in the placement. Because placement is about more than a place on a page spatially, it's often also predictive about what the user is doing at the time that space is occupied by an advertisement, and what they're going to do next. Search in many ways is easier than display because there are fewer behavioral modalities involved, and much less variation in how/where the placement shows up to the user - in consequence response is easier to model and predict. So analyzing placement is hard but worthwhile, especially when you can combine it with a notion of who the viewer is and his/her interests at the time. Advertisers and agencies need as their partners the experts who can programmatically *apply* behavioral data from a variety of partners to a detailed knowledge of placements and user modalities. I count CPM Advisors among a small group of providers who are doing that and building the tools to do so.
What's your view on ad exchanges? Did your experience with Root Markets or Consorte Media help inform your opinions and strategies regarding the exchange model?
I believe in making online media, in general, and display media, in particular, easier for advertisers to try, to target and to buy in volume. Ad exchanges frequently fall prey to false comparisons with financial markets. A lot of it centers around the false notion that impressions are commodities and can be "traded" interchangeably. It's just not so (and definitely not true when it comes to leads either...). But take a look at two leading "exchanges" - Right Media and Doubleclick AdX v.1.0 . The former is a venue where there are some buying exclusively, some selling exclusively, and some entities that buy and sell from one another. It's a matchmaking service and an adserving technology with a shared cookie space that I think has opened the eyes of a lot of people to the possibilities and the perils of multi-party buying/selling, and as such must be judged a success, even if it appears it is not (in its current form) going to be the definitive central place to buy or sell most media online. Doubleclick is more of a venue for buyers to quickly buy across a larger number of (mostly) direct-publisher sellers, where Doubleclick plays the middleman and absorbs payment risk, provides liquidity and anonymity in some cases.
There are many other parallels between the Root Markets experience and ad exchanges - and funnily I spoke with some of the Right Media people on several occasions when I was at Root Markets. There are a lot of parallels between what we were trying to do in the lead generation space (unsuccessfully) to what Right Media has both succeeded and failed at in the broader online advertising market. We both wanted to create a marketplace for the actual buyers and sellers of leads to meet, and disintermediate some of the middlemen, but realized that to get liquidity initially we would have to work with the aggregators (leads) or ad networks (display). In both cases there weren't always incentives for the sellers with their captive salesforces to regard the exchange as anything more than an outlet of last resort to sell their unsold inventory. I think Right Media may have had the same initial vision of disintermediating some networks who don't add value (of course there are many that do add value).
Both RMX and Root made efforts to reach out to quality suppliers - we drove leads ourselves by buying media and did deals with pubs like the New York Times; Right Media created their own ad network and also started encouraging standalone advertisers, agencies and others to use their exchange, and some networks started using it as their sole platform and source of supply.
And of course the same issues come up in these environments regarding quality (that pesky commodity point again), transparency, channel conflict, and so on. But they can all be overcome with enough effort and investment.
How do you make the case for brand marketers to use ad exchanges? Or, do you?
There is a lot of value for brand marketers here, although being able to reliably predict volume is still a concern - but this is improving all the time as well. You need to work with a specialist like CPM Advisors who knows where the good quality and the bad quality publishers and networks are. The shared cookie spaces and increased data available mean that things like cross-platform frequency management are possible, something novel that brands should be thinking about as they strive for the right balance of reach and frequency within their target audience. At any rate, I would make the case to pretty much all brand marketers not to try to do it themselves. And even for most agencies representing brand marketers, it is still too hard to do efficiently themselves. Where financial analogies make more sense - you wouldn't advise a retail investor to go to the NYSE to buy 100 shares of Microsoft; brokerages and others have grown up in this market as on-ramps to help the end-user get buy they need, whether that is an individual stock or a mutual fund - so firms like ours can help be the on-ramp for agencies and advertisers to these new markets.
A recent CPM Advisors blog post said, "There is so much 'untargeted' display inventory out there, much of it of dubious value, growing at a fast rate. The costs to trying out all this inventory, however, is going to be very high for advertisers or ad networks looking to benchmark it." Will real-time bidding and the ability to bid on single impressions mapped against advertiser data unlock display and overcome the "untargeted" issue?
My point here was commenting on someone's idea that as inventory expands, the price goes down and approaches zero which I don't think makes sense because impressions are not a commodity and are often of unknown value (so inventory spots with known characteristics are more valuable). Real-time bidding is still a cool idea whose specific implementation details and value is uncertain. It certainly doesn't make sense to bid on every single impression given latency, overhead and cost, but for highly desirable users there needs to be some kind of bidding/auction mechanism in place. In whatever way real-time bidding makes sense and can help advertisers, we will be doing it. But it's not the panacea to make display magically work for everyone.
Do you feel you have effective attribution models that can leverage the scale of display to drive conversions through search? What more can be done here?
CPMa's goal is to get every online marketer who is buying search to at least do some display advertising in the form of retargeting. We've created a quick self-service way for them to do it with a credit card at http://cpmatic.com that allows them to create their pixels, set up their campaign and everything without having to talk to a salesperson or fax an IO - the ROI is fantastic as you know, but volume can be low so reducing overhead is key. Attribution is an extremely complex issue, and it will remain contentious for some time to come. There is little doubt that display and search work together to create value for marketers, sometimes leading to unfair lower ROI for display media at the cost of search. But impression-based conversion is also fraught - witness past tactics by certain companies to buy lots of untargeted super-cheap display units to get view-conversion credit for large-scale campaigns. Once you add retargeting into the mix... it gets even more confusing since you could potentially have multiple display partners claiming success, but the case for search + display retargeting is an easy one for us to make.
How do you see the ad network model evolving?
One scenario is a bifurcation probably - a smaller number of very big, either technology-driven or sales-driven networks... the smaller sales-driven ones falling away and then a large number of targeted, focused specialized networks (either vertically or some other targeted orientation). Technology and lower serving costs means more people can get into the ad network business. Whether they should, remains to be seen. In some ways, more networks could mean more friction, but it could also mean more people able to service unique advertiser needs and the agency and ad network models start to blur. There's some balance or equilibrium in there.
In your opinion, what do media agencies need to do to prepare for the increasingly automated media buying environment?
Hire developers and math PhDs and invest in technology in addition to what they are doing now, or focus on their relationships, the creative and the cross-media aspects of the puzzle and work with companies like ours who are building the platforms and technology components. But I honestly don't think anyone has the answer yet as to how to properly resource the online ad business yet and what business models are going to make sense long-term, so we would like to talk to everyone from agencies, networks, data providers to publishers about how to create real long-term value for advertisers. I'd rather talk about how we move many more dollars from less accountable media to the more accountable media world we live in, than grabbing with others for the relatively small share that is already here.