Inventory forecasting asks a short question: what should I order this week?<\/strong><\/p>\n Demand prediction asks a longer one: what is going to happen over the next few months, and what should we do about it?<\/strong><\/p>\n The horizon changes what the answer is for. A reorder proposal feeds a purchase order. A demand forecast feeds capacity planning, cash flow, hiring, supplier negotiation and what you decide to promote.<\/p>\n It also changes how confident you can be. Accuracy falls sharply the further out you look<\/strong>, and any forecast presented without that caveat is being oversold.<\/p>\n Five decisions that need a view beyond next week.<\/p>\n Production capacity.<\/strong> Do we need another shift, another machine, more subcontracting?<\/p>\n Cash flow.<\/strong> Big stock purchases have to be paid for. Knowing three months ahead changes how you plan.<\/p>\n Supplier commitments.<\/strong> Volume pricing needs a volume commitment. Better forecasts mean better negotiating positions.<\/p>\n Staffing.<\/strong> Seasonal businesses hire ahead of demand, not during it.<\/p>\n What to promote.<\/strong> Pushing a product you cannot supply is worse than not promoting it.<\/p>\n FIGURE 1: WHAT A DEMAND FORECAST IS FOR<\/p>\n Capacity and staffing<\/p>\n Cash and commitments<\/p>\n What to promote<\/p>\n Beyond raw sales history:<\/p>\n The best forecasts also take in things from outside your data: planned promotions, known customer commitments, a product being discontinued. Odoo can hold this context; it cannot know it unless somebody enters it.<\/strong><\/p>\n Three points that determine whether a forecast helps or misleads.<\/p>\n Next month is a reasonable estimate. Six months out is a direction. Two years out is a conversation, not a number.<\/p>\n Use forecasts at the horizon they support. Treating a twelve-month figure with the same confidence as a one-month figure is how planning goes wrong.<\/p>\n “We expect 400 units” implies a precision that does not exist. “Between 340 and 470” is more useful, because it tells you how much room to leave.<\/p>\n If your system only gives point estimates, mentally add a range<\/strong> \u2014 and plan for the lower end of it when the cost of over-committing is high.<\/p>\n Every forecast assumes tomorrow resembles yesterday. That is usually true and occasionally very wrong.<\/p>\n A new competitor, a price change, a lost major customer, a regulatory shift \u2014 none of these are in your history, and all of them break the forecast.<\/p>\n FIGURE 2: THREE RULES FOR READING A FORECAST<\/p>\n Match the horizon to the decision<\/p>\n Prefer a range to a point<\/p>\n Remember what it assumes<\/p>\n Works well:<\/strong> established products with years of history, recognisable seasonality, a stable market, and enough volume that patterns are visible above the noise.<\/p>\n Struggles:<\/strong> new products, a few large unpredictable orders, project-based businesses, markets that have just changed structurally.<\/p>\n A business selling ten thousand units a month across a stable range<\/strong> gets genuinely useful forecasts.<\/p>\n A business winning four large contracts a year<\/strong> does not. The pattern it would need to learn does not exist, and a forecast will be produced anyway.<\/p>\n Knowing which you are is the most important thing in this article.<\/p>\n Worth its own mention, because it catches people repeatedly.<\/p>\n If you ran a promotion last March, your history shows a spike in March. The forecast may repeat that spike next March \u2014 even if you have no promotion planned.<\/p>\n Equally, if you plan a promotion this June, the forecast will not know unless you tell it.<\/p>\n Tag promotional periods in your data.<\/strong> Otherwise the model learns that certain months are naturally busy when really they were artificially busy, and it will keep expecting a lift you no longer create.<\/p>\n Forecast at the level you decide at.<\/strong> If you plan capacity by product family, forecast by product family. Individual product forecasts are noisier and you probably do not act on them at that level anyway.<\/p>\n Review monthly, not quarterly.<\/strong> Compare last month’s forecast against what happened. That comparison, done regularly, is how the forecast \u2014 and your judgement about it \u2014 improves.<\/p>\n Keep human overrides visible.<\/strong> When someone adjusts a forecast, record why. In six months you will want to know whether the override or the model was right.<\/p>\n Track your error.<\/strong> How far off were you, on average? A forecast you have never measured is a number you should not be planning on.<\/p>\n FIGURE 3: FORECASTING THAT HELPS AND FORECASTING THAT MISLEADS<\/p>\n Used well<\/p>\n Used badly<\/p>\n Predict something unprecedented.<\/strong> A pandemic, a competitor collapsing, a regulation change. By definition these are not in the history.<\/p>\n Replace judgement.<\/strong> Your sales team knows a major customer is about to change supplier. The model does not.<\/p>\n Be accurate on new products.<\/strong> No history, no pattern. It will produce a number and the number is decoration.<\/p>\n Work on bad data.<\/strong> Sales history with wrong dates, missing records or unrecorded returns produces a confident wrong forecast.<\/p>\n Demand prediction is for decisions that need weeks or months of notice<\/strong> \u2014 capacity, cash, commitments, hiring.<\/p>\n It works where you have years of history, stable patterns and enough volume for the signal to show. It does not work on new products, lumpy demand, or a market that just changed.<\/p>\n Read it as a range, not a number. Match the horizon to the decision. Measure your error every month.<\/strong> And keep the people who know what is coming in the conversation, because the forecast only knows what already happened.<\/p>\n Planning capacity or cash months ahead and working from a guess?<\/p>\n Get in touch. We will look at whether your demand pattern is one that can actually be forecast \u2014 and tell you straight if it is not.<\/p>\n<\/div>","protected":false},"excerpt":{"rendered":" A different question from reordering Inventory forecasting asks a short question: what should I order this week? Demand prediction asks a longer one: what is going to happen over the next few months, and what should we do about it? The horizon changes what the answer is for. A reorder proposal feeds a purchase order. […]<\/p>\n","protected":false},"author":3,"featured_media":9187,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[26],"tags":[],"class_list":["post-9186","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai"],"yoast_head":"\nWhat it is used for<\/h2>\n
\n
\n
\n
What it uses<\/h2>\n
\n
Reading the number honestly<\/h2>\n
Accuracy decays with distance<\/h3>\n
A range is more honest than a point<\/h3>\n
It predicts the past continuing<\/h3>\n
\n
\n
\n
Where it works and where it does not<\/h2>\n
The promotion trap<\/h2>\n
Using it in practice<\/h2>\n
\n
\n
What it will not do<\/h2>\n
The short version<\/h2>\n