Demand forecasting for energy
Consumption forecasts drive purchasing, balancing and capacity decisions, so forecast error goes straight to cost.
▶ The solution
I build a demand forecast on your own sales data and put inventory, purchasing and production planning on a measurable basis instead of experience alone.
▶ Use case
Hourly or daily consumption is forecast from temperature, the calendar and past consumption patterns, and forecast error is tracked continuously.
▶ Data you need
Hourly or daily consumption history, weather data and the public holiday calendar.
The industry's data landscape
In energy, data comes from meters, field equipment, maintenance logs, billing and customer systems. Time-series data is abundant and often high-frequency. The real challenge is turning it into a form the people making decisions can use.
Demand forecasting uses past sales, seasons, promotions and other effects to predict how much will sell in the coming period. The value is not in the forecast itself but in how it connects to planning: how much stock to hold, when to order, how to plan production.
The math behind it
Safety stock = z × √(L × σ_d² + d² × σ_L²)
- z: factor for the target service level (about 1.65 for 95%)
- L, σ_L: average lead time and its standard deviation
- d, σ_d: average demand per period and its standard deviation
How I work
- Week 1
Data and the current process.
Sales, stock and promotion data are reviewed together with today's planning process. The accuracy of the current forecasting method is measured, giving a baseline to compare against.
- Weeks 2–3
Model and product groups.
Products are grouped by sales behaviour and a fitting method is chosen for each group. Forecasts are compared with the current method over the same period.
- Weeks 4–5
Connecting to planning.
The forecast, together with its error, is turned into stock and order suggestions, in the format the planning team uses, refreshed on a schedule.
- Final week
Handover.
How the model works, when it needs updating and how to monitor results is documented and handed to your team.
▶ Questions and answers
Demand forecasting for energy — what data do you need?
Hourly or daily consumption history, weather data and the public holiday calendar.
When is this the wrong choice?
You don't have enough history. For a new product range or a new company, a statistical forecast won't be reliable. The data needs to build up first. Your sales depend on a few large contracts. When demand hinges on a handful of customers' decisions, working closely with sales beats a model.
How does this work start?
With a short intro call. I listen to what you need and tell you plainly whether your data is enough for the job and what the first step should be. I reply to every message within 24 hours.
▶ Concepts
Linked to Wikipedia and Wikidata.
▶ Sources
- [1]Wikipedia, Safety stock, 2026-10. en.wikipedia.org/wiki/Safety_stock
- [2]Wikipedia, Mean absolute percentage error, 2026-10. en.wikipedia.org/wiki/Mean_absolute_percentage_error
▶ Intro call
Let's talk about your situation.
In a 30-minute call I listen to what you need and tell you plainly how I can help.
Book an intro call▶ In other industries