This is what one large Demantra customer lamented to me a few years ago, in reference to the poor statistical forecasts coming out of Demantra just a short time before the intended go-live of their major demand planning project. But within a few days of some initial basic “engine tuning” steps being performed, the Project Team were all smiles.
Are you tolerating poor statistical forecasts being produced by your Demantra system? Maybe you don’t realize how flexible and powerful Demantra’s statistical engine is. Maybe you accepted the out-of-the-box configuration, or you spent a minimal effort on engine tuning before go-live just to meet your timeline or budget. There are many important configurable elements of the Demantra engine – such as causal factors, various engine “models” (various types of regression, Holt, Crostons, etc), the forecast tree, dozens of engine parameters – all of which can add or detract from attaining your forecast accuracy goals.
Here are some examples (industry/challenges) of tuning exercises AVATA has performed recently. Noted here are the industry, the business/modelling challenges, the nature of the engagement, and some results for our clients.
Contact us to find out more about how we might help you ensure your statistical forecast is the best it can be.
written by Andrew Calder, Director of Consulting, AVATA Australia