Companies such as Citrix, VMware, and EMC have leveraged predictive analytics to spectacular results. Many smaller agile enterprises have done the same.
But does it translate to the sales pipeline? With the right data — digital body language, behaviors, sales stages, actions and responses, and models based on past quarter performance — it’s possible to make powerful predictions. Ideally, predictive analytics reveal which existing deals are likely to close, when they will close, and what specific actions can be taken to successfully realize the expectations. Follow the script, execute, and keep the salespeople from shooting themselves in the foot. That’s the dream.
But the reality is that the right data is nearly impossible to come by. So while you might be able to generate accurate guesses on a macro level, specific forecasts when it comes to micro-level events, such as individual buyer behavior, are rarely accurate.
Predictive analytics assumes you can map out the sales process like a chess match. If you make this move, I’ll make that move, you’ll respond thusly, and on and on until I win. Yet the sales process is complicated and littered with unstructured and subjective data. Often there are multiple buyers, and buyers tend to give misleading cues. It isn’t as linear and deterministic as a game of chess on a structured board.
There are other challenges. Common definitions are often lacking, and the risk of bad data is high. For example, are the terms describing progress based on buyer behavior-oriented definitions or seller-manipulated definitions? If stage four of the buying process is a proposal submission, a seller could submit a proposal and mark the process in stage four. But if the buyer hasn’t asked for a proposal and isn’t reviewing and responding to one, that data point is useless. Attempting to remedy these problems is extraordinarily difficult.
Markets are constantly in flux. Buyer behaviors perpetually shift. Algorithms must continuously be adjusted. When you start to factor in all of the details that constitute the buyer’s journey, the permutations can get unwieldy. It becomes extremely difficult to turn the information into something actionable. Sales teams rarely have the data, scientists, and bandwidth to make the necessary refinements to generate accurate predictions and successful actions.
Predictive analytics is not good at simply measuring the success probability of a specific deal or providing guidance on how to close specific transactions.
What it can do, though, is predict sales performance in the aggregate. Predictive analytics can give you the capability to model a quarter’s results based on prior performance, wins accumulated thus far, and what is in the pipeline. It can tell you that out of six pending deals, you will probably win three, but it won’t tell you which three you will win.
The essential value of predictive analytics is focus. It can help you discover prospects not already in your database and zero in on those prospects most likely to buy. It can alert you to existing customers likely to buy more and flag those likely to churn. With the right data, predictive analytics can help you generate higher win rates, greater customer retention, faster sales cycles, and lower selling costs.
Like most things, predictive analytics adheres to the 80/20 rule: 80 percent hype, 20 percent reality. That 20 percent, properly leveraged, can generate big gains.
How to Slay Your Number in 2016
Are you going to make your number in 2016?
If you are not sure but would really like to know, turn to page 46 and read our feature titled “How to Make Your Number in 2016.” Here, we summarize the primary findings from SBI’s ninth-annual research project, which captures what the best of the best are doing to exceed their revenue targets.