Most people think that the key to a successful weather company is providing timely and accurate forecasts. While that’s certainly essential, it’s only part of the equation. Going forward, the success of weather companies requires they do much more than furnish clients with forecasts of precipitation, winds and temperatures. They must also help forecast the impact of weather on a client’s business operations.
An excellent example of “impact forecasting” is what DTN does for utility companies. In the past, we would have given utility clients advance notice of the arrival of a significant weather event, such as a severe thunderstorm complex or a large-scale windstorm. We’d explain that winds would switch to the northwest and gust at over 60 miles per hour for up to six hours. It would then be up to our client to determine what to do with this weather information — whether they should make arrangements to have on-call staff at the ready to repair downed power lines, etc.
Now, however, through the use of Storm Impact Analytics, we can provide estimates for the amount and extent of damage such an event might cause. For example, we can tell a customer that the upcoming storm will cause from 2,000 to 2,500 outages that may last six hours. This kind of information allows them to prepare for high-impact events in a deliberate, purposeful way.
Moving from forecasting weather to forecasting its likely impact requires a collaborative effort with our clients, a review of historical weather data and historical impacts, and the use of machine learning. Machine learning creates statistical models that reflect the relationship between weather events and the occurrence of significant effects, such as power outages. The ability for machine learning to “learn” how a specific utility is affected by weather is essential because every utility’s network is different due to designs, age, maintenance practices and other factors. Ultimately, this can help a utility company better prepare for forecasted storms and their impact, ensuring a faster, more efficient response to power restoration.
As weather companies increasingly adopt an “impact analytic” perspective, they’ll identify more ways to help more clients in more industries. In addition to utility companies, for example, businesses that count on the expected delivery of parts and supplies can be impacted by weather events. Weather-based analytics, produced from a range of probabilistic forecasts, may suggest that a company can expect delivery of parts to be delayed by two days when they usually receive them every six hours. Thus, when a high-impact weather event threatens to disrupt transportation, companies may want to consider alternative ways of sourcing their parts.
The main point here is that entrepreneurial-minded weather companies need to do things that truly add value. Selling weather forecasts, while still a vital service, has essentially become a commodity. Collaborative consulting that emphasizes understanding how a company is impacted by weather and helps them determine the best course of action in light of probabilistic forecasts, takes the enterprise to a new level.