When supported by the right technology, sales-team insights will raise forecast accuracy and improve enterprise planning.
Sales is a high-demand job that requires an unwavering attention to prospect detail, value propositioning, and the development of trust. Having to step out of that focus every month to comb data, draw inferences, and make consequential guesstimates about the future consumes time and adds stress. Moreover, sales people are trained to sell, not interpret historical data, extrapolate trends and patterns, or build forecast models for product launches – as would a statistician or a professional forecaster. Yet across industries, senior managers rely on Sales to develop forecasts that the rest of the company plans around. This approach is not only functionally inefficient, but distracting to the very people keeping the lights on.
The solution is to put in place the tools that will enhance the sales team’s ability to raise forecast accuracy, but with minimal effort. Yes, this will require investment, most logically in either dedicated forecasting professionals or in an enterprise-grade forecasting and planning tool set. The idea is the same.
Sales teams have invaluable information to share and that knowledge is a key input to the forecasting process. But putting them in charge of preparing forecasts from scratch forces them to make difficult judgments and can amplify the inherent biases of the role. Instead, statistical modeling should form the basis of forecasting, with sales team knowledge and insight an invaluable complement.
Indeed, without intel from Sales, companies would draw plans based solely on history, effectively blind to real-time threats and opportunities. But again, it has to start with a statistical modeling of the trends and patterns that are otherwise impossible to detect. Let’s back up here a bit and look at two ways of forecasting sales.
The Traditional Sales Forecasting Process
In a typical sales-driven approach, a salesperson is responsible for forecasting demand for his product, customers, region, etc. Often using a spreadsheet, he takes last month’s, last quarter’s or last year’s performance as a starting value and adds or subtracts a percentage, based on judgment. Or maybe he adds up the current firm orders from his customers and adds a percentage on top of that. He then runs the numbers and presents a “forecast” that gets rolled up into higher-level forecasts by sales management. The company uses these forecasts and roll-ups to make decisions about supply planning, financial planning, and more.
The problem with this process is that it’s based purely on judgment, rather than on a combination of statistical modeling and judgment. Software models are far superior at identifying trends and patterns and calculating objective baseline forecasts. Let ‘em have at it. People are better at improving these baselines with foreknowledge and insight. So equip them with the best baseline projections. And let them focus on selling products rather than pulling together forecasts, a process that is time consuming and prone to human error and bias.
SiriusDecisions, a sales and marketing advisory, estimates that sales reps spend an average of 2.5 hours per week on forecasting. Yet about four out of five sales organizations do not forecast within 10% of actuals.
In sum, it just makes more sense to give the sales team the best-possible starting point and let them make knowledge-based adjustments to improve accuracy. This saves time, plays to their strengths, and increases accuracy.
The salesperson’s struggle to forecast new products illustrates how good it might be to let computers do what they do best. New products have little or no historical data, so the salesperson can’t rely on the product’s past to predict its future. She is limited to typing in a gut-feel estimate, and this is not the best place to start when critical decisions about production, stocking, and hiring are on the line.
Fortunately, sophisticated tools can derive solid forecasts for new products using the historical data of similar products as proxies. Likewise, the sales history of a discontinued product can help forecast sales of its replacement – and model the transition, phase-out, and any cannibalization. These examples are just the tip of the iceberg.
Now Then – The Technology Enhanced Forecasting Process
Business forecasting and planning software is designed to capture trends and patterns in historical data and extrapolate an objective, baseline prediction of what is to come. That’s the best starting point you can possibly achieve. And it’s a grand favor to sales people. With an objective baseline in hand, they can quickly layer in customer and market knowledge. They know about rising trends and sudden market shifts happening right now. They have boots on the ground scouting and reporting other critical market movements, such as the new competitor, the unprecedented spike in a customer’s upcoming orders, or the ad campaign they’ve got planned for the next quarter. All of these events are fundamental to predicting sales volumes, but none show up in historical data. That’s why combining history with the foresight of people is such a powerful synergy.
So, Sales is critical to the demand forecasting process, but not seminal. Sales teams can and should be in charge of making prudent, knowledge-based adjustments to baseline forecasts.
To collect data on sudden demand changes from your sales force, it’s best to keep the process as simple as possible. A complicated process can be a barrier to adoption. Straightforward inputs about expected changes are best. The baseline forecast can then adjust accordingly. Careful though, sales people can be tempted to input changes beyond what they know, such as hunches and preferences. This taints the data-input process with judgment. Lesson? Stick to what you know, start with quantitative methods, and apply market knowledge prudently to fine tune results. Remember that the stakes are high. When the sales forecast is off, everything else is off too.