Ad Server Maturation, Part 2: The Future
ADOTAS – Let’s examine the possible evolution of these two new functions of ad servers we discussed on Friday:
The ad media exchange feature’s future is discussed in length in Tim Cadogan’s article “When Ad Exchanges Talk to Each Other” and I concur with most of what he has written there. This future is bright and is a natural evolution of this near unique example of a “perfectly competitive market.”
If it manages to overcome the huge technological barriers (or rather when it overcomes these barriers) it will move one step closer towards creating such a model and all will benefit (publishers, advertisers, ad networks and ad servers) as the possibility to deliver a much better match will emerge from the mere fact that there will be many more options and much more information to create the perfect match.
As for the ad-delivery optimization function, the story is completely different. As I wrote earlier, the ad server’s targeting and optimization technologies are always (at least) one step behind. In this situation they can either try to catch up with the technology or give up altogether.
Giving up altogether is not really an option, but trying to catch up with all types of optimization technologies and integrating them requires a huge effort. The reward however, is dubious at best. It has a potential of keeping their technological edge and it improves their revenue slightly as most of them are rapidly becoming result-driven.
The risks, however, are enormous as I described earlier (performance issue, loss of focus, etc.), so how then do you solve this catch-22 where you cannot stay put on the one hand, and moving forward is highly risky and not very rewarding on the other? The answer: You use other, specialized vendors to do the work for you.
In this way, all you need to do (as an ad server) is create a platform for these new technologies and vendors to easily integrate into your system and the problem is solved. You can provide your clients with potentially excellent optimization engines (at their choosing) and you can focus on new features that will generate additional revenues for both you and your clients, all with a much lower level of risk.
Real-time bidding is another solution to this problem since in a real-time bidding environment, the optimization engines of the ad server take secondary priority (if any at all) to the bidder’s targeting and optimization decisions for obvious reason, thus making the ad server’s optimization engine almost redundant.
Of course, in this “perfectly competitive market” in which the ad servers are interoperable and they do not even perform their own optimization, what will they compete over?
The answer is that there is a lot to compete about. The competition will focus around who provides more data, who provides a better performance and less latency, who provides more transparency, who has a better infrastructure for new technologies and thus better optimization, who exposes more media (maybe through interoperability?), who provides better user interfaces to manage the campaigns, who handles more types of media (search, display, WAP, rich media) under one ad server, etc….
Assembly and Fragmentation of Technological Products
The ad servers’ evolutionary process described above is rather similar to the evolutionary process of most products we know today. It begins as a great product with proven economical and financial potential.
Once the product shows solid potential naturally, competition arises and the competition drives the product into developing new and sophisticated features, each with its own niche value requiring niche expertise. These features also evolve in strength and sophistication.
In order for the company to keep its technological leadership and superiority, it is required to diversify its expertise and lose its focus. Most companies and products choose to keep their focus on the main product and use specialist vendors to supply them with the niche features to incorporate into their product.
This process can be illustrated by two types of products: Cellphones and PCs with emphasis on one particular company: Apple
In the late 1980s both IBM and Apple started to produce PCs. IBM based its computers on Microsoft’s DOS and apple created its own OS. The evolution of the PCs went in three directions: IBM insisted on manufacturing most of its own hardware relying on Microsoft for OS. Apple insisted on creating its own hardware and software. Meanwhile cheaper manufacturers started assembling PCs made of various specialized vendors and relying on Microsoft for OS.
In the 1990s both IBM’s and Apple’s market shares in the PC market dropped dramatically, which almost lead to Apple’s bankruptcy, mostly due to its inability to keep up with the technology at competitive prices as both companies insisted on using third party vendors as little as possible and developing and manufacturing most of their technology (hardware and software) by themselves.
The development of the cellular phone had many similarities — from development and manufacturing of all aspect of the product (hardware, OS and applications) to a specialized development and manufacturing of each part of the product — boards, cameras, Bluetooth, OS applications, etc., while the phone manufacturer itself is mostly the assembler and marketer of the final product.
In this case, Apple, with its iPhone, learned from its mistakes (if only partially for the hardware, which is developed by Apple and remains its only Achilles heel) and created a platform for third party software developers to develop many applications for the iPhone, which is one of its strongest success factors.
Like other successful products, the evolution of the ad server’s development should move towards using third party technologies and allowing them to integrate seamlessly into the ad server, making the ad server an ad-serving marketplace where the ad servers talk to each other and trade in impressions, and the actual bidding is done through “bidders” (developed by the ad server or by third parties) who bid based on third party technologies they plug into themselves.
I am curious which ad servers seem to hold the most promise. My company is in the midst of assessing various platforms and I want to make sure that we’re on a technology solution that will allow us to grow as advertiser and customer needs change and morph.
There are many ad servers out there with different characteristics and the right one for you really depends on your specific needs, whether you are an ad network, an agency, an advertiser etc…
In my opinion, AppNexus is the most promising ad server at the moment. They have very sound technology and a very diverse trading possibilities.