Predictive analytics is a term that gets thrown about. But what is it? And how does it compare to more traditional methods of analysis? And, more importantly, what are the practical applications for predictive analytics in the world of RevOps. We’ll answer all this throughout this blog and offer examples in 2 videos.
Let’s start with what it is.
We’ll do that, in classic style by talking about something similar, but different, to give it context.
Trend analysis, you’ll be familiar with this, it’s not anything new. But trend analysis looks at describing what has happened and what is happening. This is what a lot of the major business intelligence tools are great at, such as Tableau, Microsoft Power BI. It's been around for a long time and gives companies great insight into the past. E.g What were our total sales for last quarter, what was the conversion rate, regional breakdown, product breakdown etc
Forecasting is the next development of analysis and this predicts the future value of the data by looking at its unique trends. Such as predicting total sales for a company for the upcoming months/quarter based on previous years of closed won data. Interestingly, this is effectively what the Hubspot “AI” forecaster does.
But, this context now allows us to understand predictive analytics. The next step.
Predictive analytics is where we are predicting future behaviour, not just a total number. To put into context using our closed won example, rather than just saying “Your total sales next quarter will be £X”, we can now say, “Your total sales will be £X and these are the specific deals that will and will not close”. The benefit is pretty obvious, now you know, with a roughly 80% degree of certainty, that deals X Y and Z will not happen this quarter. The best tools (of which I’d like to include Infer) go one step further. They tell you WHY those deals won’t close, so you can go make it happen!
So that’s what it is.
Now let’s think about some more ways RevOps can use predictive analytics, in addition to the forecasting/pipeline review mentioned above.
Lead scoring - once the machine learning model understands your business, you can use it to tier leads. Ultimately predicting their future value, likelihood of purchase, deal length. Whatever you want to include. This tiered approach to new leads allows marketing to focus on generating the best ones and then sales to focus time effectively. Lead scoring is of course nothing new, but using real machine learning to do so, rather than a gut-feel, isn’t as commonplace as you might think.
Customer churn - much in the same way as predicting new business sales, you’ll want to predict which customers are going to leave you over the next months/quarters. Feeding this back through to your Customer Success team as a churn score, along with reasons why that specific customer will churn, allows them to get ahead of it.
If you check out the videos you’ll see how predictive analytics works in Infer. Ofcourse, if you’d like to learn more about the platform. Just get in touch with one of the team.
