From basketball to big data: the predictive value of semantics
I am a big believer in the fact that the biggest big data revolution of all will be the creation of models that better explain, and maybe even anticipate, reality.
The NBA, according to this article from Grantland, is leveraging big data, analytics and statistics to evaluate a basketball player in the best way possible.
While reading, I started to make similar connections between evaluating the performance of a basketball player and doing the same with a sales rep. Like points scored in basketball, signed purchase orders are a very simple and common way to evaluate performance, albeit on a very black and white scale.
In reality, there are other criteria that are arguably just as important as signed POs. The skill in managing customer objections, shortening the length of the sales cycle and beating out your top competitor are all important obstacles that, even without an immediate signed PO, can have a huge impact down the road.
Just as there is ‘intelligence’ in the details of players’ positions on court that count more than points, there is great value in the variety of rich information we have access to in the business world. Here too, we should use it to come up with a better way to evaluate sales that could in turn be used to improve overall organizational performance.
Communication between sales and prospects could be used to assign values to attributes of the different phases in a sales cycle. Analysis of the communication stream could identify substantial differences in the status of two different sales opportunities in the same phase.
Continuing with the same logic, you could be able to associate an Expected Closing Value starting from an introductory cold call and have this number change during the sales lifecycle based, for example, on the language style and tone used in communication, the frequency of communication, etc. This ECV would drop to next to zero if there isn’t any response to the introductory email, or move up with a positive reply, and so on. In similar way to what is described for basketball, the best sales reps would be the ones who constantly score above the expected value of each phase.
A difference between the sales rep evaluation model and the basketball model is that the value of the score would be much more variable compared to basketball, and so this element should be weighted to reward the sales reps who are also able to be engaged in higher value transactions.
A major potential value of semantic technologies resides in its ability to use text analysis in ways that could revolutionize the way we model reality. With a great deal of communication happening now in written form, an intelligent analysis of this stream should be able to describe events in a way that is both more innovative and deterministic. With the right investment, we could create much more relevant models to describe and evaluate human interaction. Given the impact that a sales rep can have on an organization, it may make it easier to be able to calculate the return of this investment. This doesn’t have to be a complicated model; instead, you can start small and expand over time.
Twitter has recently made their ‘firehose’ available to universities who would like to work on research projects using this ‘communication’ data. It would be great to see some development in these areas because any research would bring us one step closer to the model I describe above. As providers of semantic technologies, we want to help in any way we can.
Author, Luca Scagliarini.