Alan Greenspan, former head of America’s Federal Reserve, once famously confessed: “We really can’t forecast all that well, and yet we pretend that we can.”
If the world’s top economists can’t get it right, then it’s hardly surprising that forecasting is a problem in business too. Forecasting is at its most challenging in sales, where month after month, sales reps churn out their revenue forecasts for management teams that have little or no faith in their accuracy. CSO Insight’s 2016 Sales Performance Optimization Study shows that the average win rate of forecast sales deals is only 45.8%. Flipping a coin would provide better odds!
The problem with sales forecasting lies in a combination of human bias in pipeline reporting and spreadsheet models that are not sufficiently sophisticated to take in all the possible factors that might be driving sales deals. Inaccurate revenue forecasts can potentially jeopardize a company’s share price, endanger its cash flow, and inflate inventory costs, so businesses are increasingly turning to a new technology—predictive analytics—for help. Predictive analytics is revolutionizing sales forecasting by replacing the limitations of human inference and bias with models based on machine-learning algorithms.
How predictive analytics improves sales forecasting
Sales forecasting with predictive analytics starts with the combining of internal customer data such as win/loss ratios, delay factors, close rates, and completeness of the sales process, with external data that indicate a customer’s propensity to buy (these data points could be as diverse as company revenue, executive changes, and social media activity). Forecasting algorithms then use machine learning to look for patterns in these large volumes of data, in ways and speeds not humanly possible. The relationships spotted in the data are then used to score each deal in the pipeline and predict its likely revenue with levels of accuracy reported to be as high as 82%.
This approach differs from using business intelligence tools and spreadsheets for forecasting: These traditional approaches rely on the human brain to infer correlations between the different factors (historical data) and the outcome (sales). However, the brain does not have sufficient power to spot links between thousands of variables and nonlinear relationships.
Sales forecasting with algorithms: in practice
While some have claimed that predictive analytics will replace sales reps altogether in the forecasting process, in practice this is not the case. At a global software company, sales managers compare forecasts by sales reps with those output from algorithms and discuss variances with their reps, who then take a closer look at their pipeline and fine-tune their estimates. When analyzing actual sales scored against those forecast, opportunities that didn’t close are also examined to find new factors driving closure that are then digitized and added to the forecasting algorithm for increased accuracy.
A Swiss chemical company uses predictive analytics to enhance the accuracy of its sales forecasting by complementing experience-based decisions with model-based forecasting. Through this process of forecast validation, the company has been able to reduce inventory levels and costs while increasing product availability, delivery capabilities, and customer satisfaction.
Predictive analytics is also becoming a key tool in sales enablement by letting salespeople know how and when to communicate with prospects based on algorithms that leverage every imaginable variable that impacts a customer’s decision to buy. A leading IT networking company is using predictive analytics in this way to help account mangers determine which customers to call on and which products to promote, based on deep insights into what their customers are likely to buy.
The bigger picture
In an earlier blog, Predictive Monthly: How A Little Bit Of Wizardry Can Transform Your Bottom Line, I explained that the key benefits of predictive analytics include not just objective and accurate predictions but also automated decision-making and the uncovering of new business opportunities. It’s for these reasons that predictive analytics is being adopted not just in sales but across many lines of business, including marketing, operations, HR, and finance. With the recent exponential increase in computer processing power and the rise of Big Data, predictive analytics is seen by many organizations as an ROI decision instead of a cost, because of the incredible value it can release from existing data and infrastructure.
Now that algorithms, machine learning and Big Data can support sales forecasting, excuses are running out for getting it wrong!
To learn more about adding predictive to your sales analytics strategy read my earlier blog, Predictive Monthly: The Role of Predictive in Your Analytics Strategy.
For 5 traits that lead to the successful use of algorithms, read the Digitalist blog, Algorithms: The New Means of Production.