Sales Forecasting & Predictive Analytics

Sales Forecasting & Predictive Analytics

When you list your top sales challenges, does forecasting make the list? You are not alone. Sales leaders across the United States identified sales forecasting as one of their most vexing challenges. With buyers exerting more control over the sales process, sales leaders are having more trouble with forecasting and pipeline management (Gartner). This specific challenge leads to many sleepless nights for you and your sales ops team.


This blog post shares five attributes you can use to help predict which deals will win and which will lose.


At the start of each quarter your sales team makes a call on which deals to commit for the quarter. You, or your team, roll up these committed deals and make your forecast. Wouldn’t it be nice to have an automated way to determine which deals to commit and which to consider as upside? Here is where predictive analytics come into play.

Predictive analytics can take many forms, but here we are referring to using your CRM data to make a prediction on sale deals. Specifically, we are referring to a supervised learning model. To build these models, you must first identify which attributes to let your model consider. After testing many attributes, across multiple customers, we identified the top five attributes to use in predicting wins and losses.

5 Attributes to Predict Wins

  1. Sales Stage – This referrers to a deal’s current stage in your sales cycle. Typically, the stages are Qualified, Proposed, and Negotiate. As you might expect, a newly qualified deal has a lower probability of winning than a deal that is in the Negotiate stage.
  2. New or Existing Customer – This indicates if the deal is for a net-new account, or if it is for an existing customer. Typically, net-new accounts are harder to close than a cross-sell or up-sell to an existing account.
  3. Dollar Amount – How much you are selling your product for will dictate the probability of it closing. Really large deals tend to have a lower probability of winning. The highest probability of closing are those that fall in your company’s sweet spot. Not too big, and not too small.
  4. Time Since Last Stage Transition – The number of days that have elapsed since a deal last changed sales stage is a key factor in considering if a deal will win. If it is actively being worked, and it changes stages recently it is more likely to win. If it is a zombie deal and has stagnated in the same stage for a long time, it is not likely to win.
  5. Time Until Expected Close Date – This is based on the day of the prediction, and the expected close date from your CRM. The relationship here is not always the same. If you are a long way from the close date, or too close to the close date, then that deal is less likely to win.

These five attributes are important to consider when you are predicting which deals will win and which will lose. If you build a model yourself make sure you include these for a more accurate prediction. At ORM Technologies, we specialize in predictive analytics and optimization for sales and marketing. If you have any questions on this post, or would like help predicting which of your sales deals will win, please email us at info@orm-tech.com.