Horseshoes, Hand Grenades and Weather Forecasts

When close enough is good enough

Image source: Gemini AI

"Figures often beguile me, particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: There are three kinds of lies: lies, damned lies, and statistics.” — Mark Twain

As part of my initial foray into substack, I’ve been cataloging future topic ideas. As it turns out, I have a lot to say, and my backlog of unfinished newsletter inventory is growing daily.

Weather forecasting accuracy is one of them, and a couple of recent Wall Street Journal articles, including this one, motivated me to flesh out my thoughts on the topic.

Particularly as it relates to retail sales and marketing.

Here’s the full article that set me off on this dispatch with a gifted link so you can bypass the paywall if you aren’t a WSJ subscriber:

Why Can’t the Weather Apps on Our Phone Get it Right

You know the drill: Before packing for a trip or leaving for work, you check your phone’s weather app. And still you end up stuck in rain that wasn’t in the forecast. What gives?

There are driverless taxis, human-sounding chatbots and smart rings that can predict pregnancy.

Yet apps can’t seem to get the weather right.

Fair enough.

Weather forecasts are sometimes noticeably wrong, and weather apps, which human weather forecasters almost universally disparage, tend to amplify the “wrongness” to millions of people (who then amplify it on social media), creating a doomsday loop of wise-guy commentary directed at weather forecasting in general.

How many of us in the field have heard this tired cliche: “I wish I had a job that paid me for being wrong.”

If only.

Since I became a certified weather forecaster at Vandenberg AFB, California, in 1983—you know, before electricity—I've been asked about weather forecast accuracy … many (many!) times.

It’s almost universally one of the first questions weather forecasters get when introduced as practitioners, and it’s also a nearly impossible question to answer without sounding glib or hand-wavy.

When close enough is good enough …

The reality is that the forecasts the vast majority of us use to plan our daily lives are remarkably accurate, and with the advent of new sensing systems and AI, they are only getting better.

Today, a five-day weather forecast is as accurate as a one-day forecast was back in 1980.

The two-day forecast for heavy rainfall is now as good as the same-day forecast was back in the mid-1990s.

Flawed predictions about the path of hurricanes are about half as likely as they were just a few decades ago.

Back in 1990, forecasters could only provide a relatively accurate prediction of weather seven days in advance. Now they can make relatively accurate forecasts ten days in advance.

They may be one of the many things we take for granted in the modern world, but more accurate weather forecasts — and our ability to access them anytime on our smartphones — have tremendous value for our economy.

Source: NPR / Planet Money

Of course, there will be slight differences, and some forecasts will be better in different seasons and areas of the country. Still, the forecasts available on our phones or on TV today are more than suitable for meeting the needs of most consumers.

For advertisers …

The magic of consumer-facing apps is in their presentation, not their precision. Granted, many claim to be more accurate than their competitors, but that’s more about the science of marketing than the science of weather forecasting.

Most Americans (at least 60%) check the weather forecast every day, sometimes multiple times a day, and they do so with a distinct planning mindset. The vast majority take the forecast at face value because, while it might be materially wrong on occasion, it is right far more often than not.

Because of that, the weather forecast provides an incredibly valuable data point for measuring and predicting what people want and need, more than almost any other external driver of behavior—including the actual observed weather.

While it might sound like heresy coming from someone in the weather forecasting business for forty years (yikes!), the net/net is this: whether or not a forecast is dead-on accurate doesn’t matter.

It’s not the observed weather that moves the needle for consumers; it’s the (reasonably accurate) forecast, the presentation, and the reach that matters.

The video below is by far my favorite example. You can watch it here and via the image below.

Increasingly accurate forecasts will positively impact our lives and the economy, and I’m excited to see what the future holds.

For advertisers, though, the future is now as it relates to using weather forecasts to create, time, and target uber-relevant and hyper-local ad messages.

The bottom line is that if you’re a retailer or marketer (or both) representing a seasonal service or product category and you haven’t incorporated a weather strategy into your planning and execution … you might be doing it wrong.