A maturity ladder from spreadsheet to ML — and an honest account of its limits — for beer and non-alcoholic demand. The full track, in order:
-
Forecasting Malt and Hop Prices for Smarter Procurement
How forecasting malt and hop prices guides procurement timing and contract sizing for breweries, and why hedging manages the residual risk forecasts can't remove.
-
AI for Taproom Footfall and Staffing Forecasts
How AI forecasts taproom footfall from weather, day and events to plan staffing, stock and food prep, and why spiky events and small samples limit it.
-
Demand Forecasting and What-If Analysis in Tableau
Build a demand forecast and what-if dashboard in Tableau using built-in forecasting, parameters and TabPy — and know where the basic model stops.
-
AI for Vineyard Yield Forecasting
How AI forecasts vineyard yield from bunch counts, berry weight, NDVI remote sensing, and weather — to plan tanks, barrels, and sales — and why vintage weather still breaks the model.
-
The Beverage Demand Forecasting Maturity Model: From Spreadsheet to ML
A practical demand forecasting maturity model for breweries and beverage brands — from spreadsheets to machine learning, with honest tradeoffs.
-
Forecasting With No History: The Non-Alcoholic Beer Problem
New product forecasting for non-alcoholic beer with no sales history — analogue methods, Bayesian priors, and practical launch-stage approaches.
-
Driver-Based Forecasting: Beyond the Annual Budget
Driver-based forecasting replaces static brewery budgets with dynamic models tied to real volume, mix, and cost drivers — including for NA beer.
-
Seasonality and Weather: Beer's Most Predictable Swings
How weather and seasonality drive beer demand forecasting — decomposition methods, temperature indices, and what changes for non-alcoholic beer.
-
Promotional Lift: Separating Real Demand from Discount Noise
How to measure and forecast promotional lift in beer — separating baseline demand from discount-driven volume spikes and avoiding the forward-buy trap.
-
Hierarchical Forecasting: Reconciling SKU, Brand, and Total Volume
How hierarchical forecasting reconciles demand signals across SKU, brand, and portfolio levels — and why misalignment between levels destroys planning credibility.
-
Cannibalization: Does Non-Alcoholic Beer Eat Your Lager's Sales?
A framework for measuring cannibalization between non-alcoholic beer and conventional lager — and how to separate cannibalised volume from genuine category expansion.
-
Forecast Accuracy Metrics That Matter (and the Vanity Ones to Drop)
Which forecast accuracy metrics actually drive better beverage planning decisions — MAPE, WMAPE, bias, CFE — and which metrics flatter but mislead.
-
The Honest Limits of Sales Forecasting: When to Trust the Human
The real limits of demand forecasting in beverages — what models cannot do, when human judgment outperforms algorithms, and how to design a hybrid process.