If you didn’t know, no forecast is 100% accurate. At their core, forecasts are the best guess of what will happen, an educated guess. But it’s still a guess of human behavior which is difficult under the best of circumstances.
And yes, the accuracy of a forecast can improve with time. A myriad of models, formulas and theories can be applied to drive improvements. But what happens when you don’t have the time because the delta between forecast and actual is too large and/or the results impact business today?
First, keep working on the forecast. Forecasts drive multiple aspects of business, so work to improve the accuracy by:
- Reviewing accuracy monthly – Set up a monthly review to compare the forecast to what happened.
- Figure out where the miss occurred – Analyze the formulas/data/assumptions included in the original forecast to determine where the misalignment occurred.
- Adjust the forecast where needed – If you identify formulas/data/assumptions are the cause, adjust where needed.
In the meantime, here are three things you can do right now to mitigate the impact of forecast accuracy challenges:
- Use historical demand – Include a review of 12- or 3-month demand. Ideally, you would pick one or the other to reduce complexity and the possibility of someone picking whatever number they like.
- Use inventory analysis to categorize – Analyze your inventory, assign volume and volatility categories to each item. Determine your high volume items using (A.B,C analysis) and your high volatility (X,Y,Z analysis).
- Create a service level matrix based on categorization. – Create a matrix based on business needs for coverage. For example:
- If you have a high volume, low volatility item (A,X) you may want a high service level, and set it up with a smaller safety stock with consistent deliveries.
- If you have a low volume, high volatility item (C,Z) you may want a lower service level, and set it as on demand only or carry a set amount and only replace as needed.
Questions? Feel free to reach out.
Clarity For Business 