In most businesses, the sales team will talk about a funnel of opportunities that will mature into closed deals. They may have a waterfall spreadsheet showing the sequence of sales stages and conversion events, and they're almost sure to calculate "pipeline coverage": an indication of how much they have in potential deals to cover their target number for the quarter. Despite the apparent precision of their numbers, sales forecasts al too often don't match reality.

What if the forecasting problem isn't the data, but the model and assumptions they're using? And what does this imply for measurements, reports, and dashboards in your CRM system?

What if the Funnel is a Sieve?

The image of a funnel is nice, because it properly reflects a lot of possible deals being collected (as leads) and implies that a smaller number of deals will actually mature. But a real funnel eventually gets all the "flow" through the thin stem, and that's not at all reflective of reality in sales or marketing. In many B2B situations, only 5% of the stuff at the top of the funnel will mature into deals — the other 95% either go to competitors or, more likely in this economic environment, are lost to "no action."

[ For timely virtualization news and expert advice on strategy, see CIO.com's Virtualization Drilldown section. ]

So a sieve, or a filter, or a net might be a more accurate model to use. What's the difference in how you measure? In a funnel, measurements tend to focus on how long a prospect stayed in the mouth of the funnel, and how frequently you took actions to move them. In a filter, measurements would tend to focus on the match between prospect characteristics/behaviors and your company's ideal profile. You'd want to measure how quickly you were able to identify, disqualify, and reject the bad matches and how well you were able to move the good matches along. This type of measurement fits well with the Sandler sales methodology, among others.

What if the Funnel is a Refinery?

Sales will sometimes speak in terms of a pipeline rather than a funnel. This is good, but let's take it one step further: a distillery or refinery. The input to the refinery is low grade "ore" — leads — and the job of the refinery is to increase the commercial quality of the lead batch to the point a small number of prospects are willing to take a meeting, engage in a serious sales cycle, and sign a contract. This model reflects reality (particularly in terms of the amount of effort it takes to actually make deals happen), and implies that there can be byproducts that can be re-refined (that's your "remarketing database").