Hi there. I'm Adam Robbings — founder of Reuben’s Brews, a multi-location brewery based in Seattle. Like a lot of hospitality businesses, we’ve spent years making staffing schedules, preparing food, and making inventory decisions based on a mix of gut feel, guesswork, and (occassionally) checking weather apps. Some days we nailed it. Some days we threw out thousands of dollars in food. Some days we just couldn’t keep up with the crowd.
Over the last year, the industry has definitely gotten tougher. I was hearing a lot of explanations for the good, bad and indifferent being due to the weather - but based on gut feel. It had became clear: weather impacts foot traffic more than almost anything else — and no one was showing us that clearly. With a degree in Economics, and three finance qualifications, I love a challenge and this was one that gripped me.
And so I built it myself.
What you see on this site is a forecasting model that estimates how weather impacts hospitality traffic, based on real data from multiple hospitality operations.
I got the idea of the random forest model from an insight a friend said when I met him for lunch. He runs a very successful brewpub, and he said: "When the temperature breaks 70 everyone just comes out." This made a lightbulb go off in my head. We weren't looking at a linear relationship between weather and foot traffic - we were looking at a series of decision criteria, specific to the weather itself, day of week, and season. Our model has over 110 paths the weather can uniquely fall into.
The core engine uses a machine learning approach — specifically, a random forest model — to weigh multiple weather variables (not just temperature and rain, but combinations, sequences, and extremes). The output? A single number that estimates expected change in foot traffic for any given day.
Right now, we’re publishing Seattle’s forecast publicly, so others can benefit from what we’re learning.
This isn’t a theory. It’s built on:
Millions of dollars in real-world transaction data, over multiple locations
Hundreds of weather history data points
Live sales vs weather comparisons
We’ve already seen actionable patterns:
A certain type of sunny Saturday will regularly drive traffic up 20% or more.
The same temperature on a rainy Thursday might drop traffic by 15%.
And big event weekends like Oktoberfest or St. Patrick’s Day? Total wildcard - unless you know what to expect.
We’re not giving out the full playbook - you know your business better than us. We don't have all the answers. But what we’re publishing here is grounded in operational reality, not speculation.
We’re just getting started. Future iterations of the model will:
Expand the historical data window for deeper patterns
Integrate location-specific event data (think game days, beer festivals, parades)
Shift from focusing purely on transaction volume to actual guest counts, especially as businesses move toward shared tabs or fewer purchases per visit
We’re building this to help the industry make better, faster decisions — without needing a data team or a BI platform.
We’re testing custom models for individual businesses, using their own data to generate tailored forecasts. If that sounds like something you’d use - even in a city outside Seattle - please join the waitlist.