The basic thinking behind the concept of customer forecasting is simple: your inventory replenishment planning will be more sensitive and responsive to changes in demand if you factor in your customers changes in demand.
Sure, that all sounds great, but let’s do some math to see
how quickly this approach can fall apart.
Say you have 10,000 products and you sell product out of 20 locations. Due to longer lead times, you keep a running forecast that looks ahead to the next 12 months. For each of those products, you have 10 customers. That means you need to calculate 10,000 products times 20 locations times 12 months times 10 customers…
that’s 24 million forecasts that need to be checked each month!
That’s not just unreasonable; that’s basically impossible, unless you have an army of workers doing the calculations for you. A process that requires that much labor will ruin the profitability you were hoping to reap from your JIT process.
To be fair, most inventory-based businesses have fewer locations and fewer products than the previous hypothetical. So let’s say your business is a bit smaller and you have one warehouse with a few hundred products that sell to a few primary customers.
Here’s what that might look like:
If you have 100 customers buying a product that sells 20 units a month, then each customer’s forecast is so small that you’ll have too much noise to combine these customer forecasts and create a usable total forecast. When you add up hundreds of “noisy” forecasts with the more reliable forecasts, you just end up with a jumbled mess. It’s better to leave that kind of calculation to a statistical engine and carry just a bit of of buffer stock.
It should be noted that if you add up forecasts from customers that are slightly inflated, you could end up with a combined buffer stock that is far beyond your intended buffer stock level. You should also consider the demand of those customers who don’t provide you with forecasts. The net result is a forecast that’s barely comprehensible and does little to help with replenishment planning.
Let’s go one step further on this quest for useful customer forecasts and ask some questions about the systems needed to make this happen.
Will you collect forecast spreadsheets from each customer? How are you going to combine all that data into a single set so you can build your forecast? Do you have someone who can follow up with your customers if something is amiss, and how will that affect your forecast in the meantime? As you can see, there are more questions than answers here. The odds of this working are simply not in your favor.
What you need is a statistical forecast that takes all the above factors into account without a tedious manual process. Consider how useful it could be if all the data that’s held in your ERP system could be fed into a data-based forecast.
If you’re already an Acumatica user, you know that you’ve got a powerful tool at your disposal for sorting through this kind of data. However, additional help may be required to help build dynamic inventory forecasts to use in replenishment, meaning you’ll want to skip the manual process of compiling individual customer forecasts. That’s the far easier way to prevent stock-outs and keep excess stock to a minimum.
That's why Acumatica offers add-on tools that work with the data you already have in your ERP to build more accurate forecasts, like NETSTOCK.
Inventory replenishment shouldn’t be a guessing game, and it shouldn’t require a massive team spending endless hours collecting individual customer forecasts on outdated spreadsheets. If you employ the right tools for your needs, then you’ll achieve a balance in your inventory that fits with your customers’ needs based on their history with you. That’s the key to inventory optimization and maximizing your profitability in your warehouse.