Sales forecasts are often one part guesswork, one part analysis, and three parts wishful thinking.
And yet, sales forecasts are critical business tools used to make big decisions on everything from hiring to budgets to expansion plans. Basing pivotal business decisions on a wish-cast is insanity — yet companies do it every day. NASDAQ and NYSE are littered with stock dives from “surprising” revenue results. Even the most respected companies in the Fortune 20 sometimes get their forecasts dead wrong.
As New Yorker journalist James Surowiecki points out, experts in any given field “are prone to making daring and confident forecasts, even at the risk of being wrong, because when they’re right the rewards are immense.” But when you’re wrong about a sales forecast and then use that “data” to make critical business decisions, the results can be disastrous.
There is a better way: Use software to move sales forecasting out of the land of wishful thinking and into the measured, metrics-driven world of big data. Sales forecasting algorithms can serve as a pair of unbiased eyes analyzing every aspect of your forecast — like a business consultant, but without the fees, ramp time, and 30 slides of endless abstraction. Algorithms can score your numbers, leads, and opportunities and compare them to petabytes of industry sales data to make the “call” on a forecast.
NOTE: VentureBeat’s upcoming GrowthBeat event — August 5-6 in San Francisco — is exploring the data, apps, and science of successful marketing. Get the scoop here, and grab your tickets now to save $200!
Software-driven sales forecasting facilitates a new type of culture within a company, one where action, facts, and clarity become the tenets of your business, instead of wishful thinking.
Here are three reasons your sales department needs to embrace data science now.
Reason No. 1: Improve clarity in sales communications
Sales people speak a different language — and few outside of sales seem to understand what they’re talking about. Other teams in the office often struggle to interpret what sales people say into actionable strategies to improve business outcomes.
Case in point: “It will close tomorrow” could mean “the deal will close tomorrow” or “it will probably never close.” This is why a vice president of sales hires “his” guys because he trusts them, understands them, knows their language, and can translate their language back into the C-suite.
A data-driven approach, using methods like opportunity scoring and forecasts generated by machine learning, takes sales out of the realm of sales-speak and into the land of measurable, metrics-driven business culture. A high, medium, or low score is a something everyone in the office can understand and use to make the right business decisions in their departments.
Reason No. 2: Reduce complexity in forecasting
Algorithms don’t try to think outside the box. They just crunch data and spit out numbers. But sales managers often do try to think outside the box when making forecasts. They are human, so they usually consider too much or not the right kind of information when making predictions and thus come to non-valid conclusions. Complexity more often than not reduces validity.
Several studies (described in Daniel Kahneman’s book “Thinking Fast and Slow”) have shown humans will make inaccurate predictions even when told what the correct outcome is before making a decision. They feel they can overrule a proven finding by analyzing additional information about a case, but this additional research is actually anecdotal, not empirical, and does nothing to change the final outcome.
Reason No. 3: Rein in inconsistency
Not only does over-analysis and complex thinking get in the way of creating accurate sales forecasts; so, too, does inconsistency. People, unlike computers, are inconsistent. Their thoughts, feelings, and judgments change all the time in response to internal and external circumstances.
Companies depend on people to come up with accurate sales forecasts, but people are by nature inconsistent in how they approach data collection and analysis. One day a sales rep may be feeling excited and happy and will look at data and see a wildly positive forecast. The next day, feeling negative or tired, he may see an entirely different nuance. (Sales forecasts can even change depending on the time of day they’re created; fueled up on coffee, sales reps tend to create more positive forecasts in the morning, compared to more metered forecasts in the afternoon.)
Algorithms don’t have feelings, so they approach data analysis and forecasting with complete consistency. When you’re looking for an accurate sales forecast, a computer will always deliver the more reliable report over its human counterpart.
Of course, until now, applying data science to sales forecasting didn’t make much sense, because companies didn’t have the data, expertise, or software needed to infuse data science into the sales cycle, namely into weekly pipeline or forecast meetings. On top of that, there was a lack of will and, to a certain extent, fear and disdain toward using software to forecast sales. Many professionals still argue sales is an art, not a science.
But the availability of cloud-based sales applications leveraging data science have changed all that. To get accurate sales forecasts, there’s no longer any need to hire high-priced data scientists, no more lengthy processes to gather up company and industry-wide data, and no more questions about how to interpret the results.
Software will never replace the magic a great salesperson brings to closing a deal, but then again, it will never present wish-casts just to make executives happy.
Michael Howard is chief executive of C9 and who brings over 20 years of experience in the analytics industry, including at companies such as Greenplum/EMC, Ingrian Networks, Oracle, Outerbay and Veritas.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.