Ever noticed how sales managers and sales VPs all get a little weird around forecast time? There’s a bit of edginess in the air; they’re forced into looking at data, and extrapolating some meaning from it; they have to take their conclusions and factor in seasonality, personnel churn and other factors; and they have to present numbers that suggest they’re doing a good job, but not so good that the numbers are unattainable.
Who wouldn’t be somewhat on edge during this process – especially if you had to do it manually by assembling a view of the truth through multiple data sources fed by multiple systems?
It was probably easier in the olden days, when an old-timer in sales leadership would proclaim a forecast number and then everyone would get back to work trying to meet it. It wasn’t really a forecast – it was a goal, chosen somewhat arbitrarily, an assertion that was based on gut feeling and the number from the previous quarter.
But it no longer works that way, especially as sales is no longer the only part of the business dependent on forecast numbers. There are numerous constituents within the business than need a good degree of predictability in the sales process. Depending on how well integrated sales is with the rest of the business, the forecast could affect functions from finance to manufacturing to customer service. The volume of production in a manufacturing setting may be based on the forecast. So might the timing of investments or staffing levels in different parts of the organization. Being able to trust the forecast allows others in the organization to make important decisions with confidence.
So, the goal based on someone’s gut isn’t going to cut it anymore – especially since there are so many data sources now available to sales managers. There’s the basic financial information – total sales, average deal size, average time to close a deal, and so on. If you’re using a compensation management solution, you can take this data and divide it down to an individual sales rep-level. If you’re in a subscription business, how soon do contracts get signed and terms of service begin?
Then there’s trend data. How do sales fluctuate based on seasonality? How much is closing in the last week of each quarter?
With that data available, guessing is out of the question. But there are still the “intangibles:” how will changes to compensation impact results? Are your sales team’s assessments of the pipeline accurate? If we change selling strategies, how could it affect the number of deals we close? What will the impact on our business be if we implement new technology, like CPQ or CLM? How will more discounting – or less discounting – impact deals closed and the top-line revenue?
These things introduce uncertainty and conjecture back into the mix. This is why forecasting is still something of a dark art: data can help nudge it into closer alignment with reality, but only so much. That’s why almost 4 in 10 respondents to a survey CallidusCloud conducted late last year said that a forecast with less than 80 percent accuracy was acceptable.
However, that degree of resignation to a somewhat accurate forecast will hopefully start to change soon. Predictive analytics are being applied to sales forecasting to take the data known about these “intangibles” and pair it with the harder data to provide a predictable range of outcomes. CallidusCloud demonstrated this at the C3 2015 event; the requirement for success is to have a platform (like the Predictive Sales Performance Platform) to collect the pertinent data and perform the analysis that does not rely on manual importation of data from disparate systems. The C3 demonstration included an examination of real customer data for a 12-month period; the first nine months of performance was analyzed and then the next three months’ performance was projected, then compared to what actually occurred. In every case, the reality fell inside the range of predicted outcomes.
This approach will not take all the guesswork out of forecasting – for example, without data on new sales staff it’s not easy to forecast their individual performance for a quarter or two. But in many cases, using technology to make sense of the intangibles and building them into the mathematical process of creating a data-driven forecast will lead to more accurate forecasts that are more attainable and more defensible, and will help the entire organization function as a more productive and profitable whole.