Trading desks are primarily of two kinds, one is a principal trading desk and the other is a solution trading desk.
There are three modes of operation; mostly manually operated, heavily relying on process automation, autonomous trading system.
These can be operated under one of seven types of entities; agency holding group, agency, advertiser, publisher, DSP, ad network and pure-play trading desk.
The trading desk can be managed at three levels; completely self-served, partially managed or fully managed.
The trading desk can operate under three models; non-disclosed model, partially disclosed model, or a fully disclosed model.
Systems are of two types; those that consist of a single solution and those that consist of multiple solutions.
Level of customisation are three; off-the-shelf, some customisation and completely custom.
The trading desk can provide three levels of transparency; no transparency, partial transparency or complete transparency.
This way we can understand that there are 6,804 different types of trading desks.
ADDED ON 6th of March 2015: There are two kinds of market access the trading desk can operate on; direct market access and limited market access. Therefore making it 13,608 types in all.
ADDED ON 9th of March 2015: There are two levels of tenancy; single-tenant systems and multi-tenant systems.
Most systems that we see today are multi-tenant systems that aim to offer one solution that has so many features and capabilities that theoretically speaking it could satisfy anyone’s needs. In that kind of approach there is usually a lot of “lowest nominator” type of thinking, which could be dangerous in a rapidly emerging market.
Commentary
It’s important to note that within the ontological study of programmatic trading desks, many other aspects should be analysed and understood. The above 8 are the most significant in terms of relevance to outcomes.
By outcomes here is meant the potential for competitive advantage within one of the 6,804 types of Trading Desks.
For example a completely custom system, while not feasible for most in an emerging market, provides far greater ability to understand and manage the market than an off-the-shelf solution does.
If transparency is a key for any given reason, then that should be a greater focus in evaluating a particular Trading Desk type.
Disclosure and mostly also transparency, should be weighed against performance and guarantees. There should be a point where performance and guarantees are so great, that absolutely no disclosure is needed.
There is a trend in the communication world to outsource things, which has two problems. One is that you’re not learning as much as you could, and the other is that there is always an element of being a middle-man in not trading for yourself.
There is an endless number of various companies that could involve themselves in media trading. Including those that are considered “black hat” practitioners. The seven kinds mentioned above are the primary, each with their own incentives.
It’s very important to understand the incentive models of the various stakeholders, otherwise it is not possible to understand the market. Incentives lead to actions and markets are made up of actions. Therefore understanding incentives is the foundation for understanding a market.
Most of the Trading Desks we see today are mostly manually operated. They have not yet reached a state where the need for massive scale process automation is so great, that they will embrace it. The communication industry is not great with process automation and that is becoming a real problem in dealing with the rapidly growing 60bn per year online advertising fraud pay-out.
What we will see in the near future, is the rise of an army of Trading Desks founded on the principles of “autonomous” as opposed to just process automation. Where as process automation relies on scripting and doing what we already know to do but doing it far more efficiently, autonomy does not rely on scripts to make it’s decisions and is able to do things in ways that are surprising to us.
Finally, there is the point about principal trading vs. solution. What we are seeing now is a lot of companies investing in the solution side of the market. This means that almost all of the legit trading we see in the market, is driven by things like clicks and actions and ad impressions. The principal trader has no such interest, but is only focused on the practice of buying low and selling high.
With the above factors and their various kinds in mind, we can establish some as superior and some as inferior, gaining invaluable insight to the dominant trading models of the future.