What this category does
Weather drives electricity and gas demand, and it drives renewable supply. Providers in this category sell forecasts and analytics to trading desks, utilities and large enterprises that need to anticipate those swings. The products include temperature and wind forecasts tuned to demand, load forecasts for a region or a portfolio, and ensemble data that shows the range of possible outcomes.
The buyers are specialists, and the products reflect that. A desk may want raw model output delivered the moment it is published. An enterprise operations team may prefer a dashboard and a daily briefing with a meteorologist on call.
Because the buyers are specialists, expect the vendor to ask detailed questions about your process. Answer them fully, since the quality of that conversation is a good preview of the support you will receive later.
How buyers evaluate
Desks tend to evaluate on timing and consistency. They ask how soon after a model run the data arrives, whether it arrives at the same time every day, and what happens on the days it does not. A forecast that is slightly better but arrives late can be worth less than one that arrives in time to trade on.
Many buyers compare providers against the forecasts they already use, over a period that includes the demand peaks that matter to them. For load forecasts, they check errors at peak hours separately from the daily average. Trials of one to three months are common, and a full winter or summer season is better.
RFP criteria usually cover data latency, ensemble availability, historical depth for back-testing, the regions and weather stations covered, support hours, the redistribution rights for internal users, and price per seat or per feed.
Trading desks often keep more than one provider, because a second forecast is a cross-check and a fallback. If you plan to do the same, ask whether the contract permits two sources and how much overlap in underlying models you are paying for. Two feeds built on the same upstream model give less independent information than they appear to.
Questions to ask a vendor
- When is each forecast published, and what is your record of delivering it on time?
- Do you provide ensemble members, and how many?
- Can I back-test with the same data I will receive live, including past forecasts, not only observations?
- How do you tune temperature or load forecasts to my region or portfolio?
- What support is available during market hours and during severe weather?
- How many people in my organization may use the data under one contract?
- What is the notice period if you change a model or a delivery format?
How to read the capability facts
Read the forecast types for the horizons you trade, from nowcast to subseasonal, and the delivery field for the format your systems take. The pricing model tells you whether the vendor quotes per feed, per seat or by negotiation, which shapes how you budget a trial.
A profile cannot tell you how much of the data you can back-test with. That is a question for the vendor, and a good answer comes with a sample.
Also consider what happens at the edges of the contract. Ask how the vendor handles a request for a regional adjustment, and how quickly it responds when a forecast looks wrong to your analysts.
Where to go next
Decide which decisions the forecast feeds before you compare vendors, because timing and horizon requirements narrow the list quickly.
Browse the Energy Trading & Enterprise Forecasting providers, filter them by region, parameter and pricing, or compare several side by side. The full list of categories is the place to start if your need spans more than one.