The hardest forecast is that of the top line because we have the least control of it, whereas costs are mostly well defined by the revenue. But as the top-line is the most elusive to predict with certainty, pressure is brought on the organization to show forecasts that the executives “want”. As such, businesses largely have no unbiased means of forecasting and no unbiased means of assessing the probability of the forecasts.
Budgets take on average 6 weeks and are out of date when done, and finance teams spend weeks after each quarter explaining why the budget wasn’t met. This creates the time consuming and expensive need for people and process to constantly monitor, explain variances to, and change the forecasts to meet the outcomes.
Without data scientists or ML, AI-Enabled One-Touch forecast can automate the entire forecast, budgeting, and planning processes with unbiased statistical forecasts. These forecasts can then be run through managers for their adjustment based on their knowledge, not guess, of the future. The blend of man-and-machine yields higher forecast accuracy in the shortest time.
Further, these forecasts can be run through a built-in Monte Carlo Simulation for unbiased probabilistic risk assessment. Planning is now done to a range of tolerance vs. a point specific number, thus reducing the need for more forecasts or the explanation of variances to a point value.
Applications include Service Level Agreements, dynamic min/max inventory management, ABC inventory classification, budgeting, LRP, S&OP...
Even with AI-Enabled forecasts that provide an unbiased baseline the managers can change, we still need a way to determine if the change is statistically reasonable, to assure the change was made from knowledge vs. guess.
Also, better planning comes from triangulating on signals that relate to the future but finding those external leading indicators of cost or demand are very hard to come by with spreadsheets, and if found they’re often only high-level surrogates that have spotty applicability when used to predict the details of the business.
Not knowing what the leading indicators are, means business is open to being blind-sided. In essence it’s blind to taking advantage of upcoming opportunities or mitigating risks that could derail revenue. At the end of the day, diminished planning ultimately affects valuation, and companies that don’t plan well, on average, suffer a 27% reduction in their PE ratio vs. companies that do plan well.
AI-Enabled forecasts as adjusted by managers can be are compared to a further AI-Enabled statistical forecast based on a selectable external indicator; e.g. unemployment, consumer sentiment, gas price, etc. This external indicator forecast is correlated to measure the strength and direction of the lead, then compared to the manager adjusted forecast to assess if the adjusted forecast is reasonable. These two-mints-in-one analytics both finds lead indicators and assesses if the adjusted forecast is within the bounds of being statistically reasonable. All this is built-in and done without need of data scientist or IT support.
Applications include budgeting, demand planning, pricing, promotion, LRP, S&OP...
Budgets are static, no sooner completed than out-of-date. But we all march to the plan, then monthly and quarterly explain why we’re off plan . . . a wasteful consumption of time and money!
For the average business, hundreds of hours are consumed in budget variance explanations. Time that would otherwise be used to better manage the business.
First, use the built-in statistical drivers to establish those drivers that “drive” the business. For example, the budgeted revenue drives the number of Sales Reps needed and the manufacturing costs.
Now, when the budget was made, there were assumptions to the revenue generated and from that revenue the costs required. If there were more Sales Reps in the first quarter then the budget assumed, then sales will be higher along with the manufacturing costs . . . so it looks like the costs have exceeded budget.
In this simplified example, we can “Flex” the budget to account for the driver of Sales Reps. When this is considered, the budgeted expenses are adjusted to reflect the added driver so a normalized comparison can be made, which often shows that expenses are in-line with revenue.
Beyond Flex budgeting, applications of statistical drivers include demand planning, LRP, S&OP...