Automated trading

The BOP engine is based on a dynamic auction-oriented algorithm supporting programmatic trading. "Programmatic trading" does not necessarily imply full system control of the buying process. A combination of automated algorithms and human decision-making can often lead to more optimal decisions.

BOP already implements a “Human/Programmatic Buying” paradigm, in which the system filters available inventory according to specified audience targeting criteria, provides recommendations based on measures of value such as price and CPM, and upon user initiation, executes both outright buys and fire-and-forget automated bidding with specified maximum willing bids. The human user reviews the system's recommendations and retains decision-making authority on what to buy or bid for, and how much to bid.

The following diagram describes the relative levels of automation and human decisioning in the platform today. Stages 1, 2 and 4 are algorithmic and only stage 3 is human-driven:

This then points the way to fully automated trading, in which even stage 3, i.e., the buying decision, is performed through predictive analytics (AI Buying).

The key hurdle to be crossed when moving to fully automated trading is trust. Buyer organisations must trust the recommendations of the platform as well as the judgement of the analytics engine that initiates trades.