In 2019, McDonald's acquired a Tel Aviv-based technology company called Dynamic Yield for around $300 million. Not for the website — for the drive-thru. Within a year, personalized menu boards were running at more than 9,500 locations across the US. The company later sold the platform to Mastercard but continues to use it to this day. This story is more than an anecdote. It's a blueprint for how to roll out personalization the right way.
Why McDonald's Bought Dynamic Yield
In March 2019, McDonald's announced an acquisition that caught almost everyone off guard. The company purchased Dynamic Yield, a personalization technology firm, for around $300 million. It was McDonald's largest acquisition in twenty years. At the time, commentators puzzled over what a fast-food chain would want with an AI company from Tel Aviv.
The answer was more straightforward than most expected. McDonald's didn't have a technology problem or an image project on its hands. The company had identified a very specific business problem. A large share of US revenue runs through the drive-thru. Digital menu boards were already in place there — technically modern, yet showing every customer the same thing. The same display at seven in the morning as at four in the afternoon. The same in rain as in a heatwave. The same for a regular as for a family passing through.
Every one of those ordering situations is different. The display was not. Closing exactly that gap was what Dynamic Yield was brought in to do.
What Dynamic Yield Actually Does at the Drive-Thru
The idea is easy to explain. The menu boards adapt their content to the situation — taking into account the time of day, the weather, current restaurant traffic, and which items are in high demand. A second mechanism layers on top of that. As a customer orders, the board suggests relevant additions to their current selection. Someone choosing a combo meal might see a coffee alongside it. Someone ordering on a hot day might see a cold drink instead.
What's notable is how little data this required. No customer profiles, no purchase history, no app sign-in. Time of day, weather, and footfall are signals every restaurant already has. McDonald's didn't start with the most complex version of the system — it started with the simplest. That is the first lesson of this story, and probably the most important one. Many companies put personalization off because they believe they first need a perfect data foundation. The largest real-world case in food service proves the opposite. It launched with signals that had been available for years.
From Pilot to Rollout: Dynamic Yield Across 9,500 Locations
The pace is also worth a closer look. The acquisition was announced in March 2019. By summer of that same year, personalized boards were running at more than 700 US locations, and the company reported higher average order values. On that basis, the decision was made to expand to 8,000 restaurants. By 2020, more than 9,500 US drive-thrus had been equipped.
The sequence is worth sitting with. First, a pilot at a few hundred locations. Then results. Then the rollout decision. The large-scale investment didn't come from a strategy document — it came from measured outcomes. That sounds obvious. In practice, it often works exactly the other way around: the full solution for everyone gets approved first, then it gets built, and measurement happens at the end when nothing can be changed anymore.
One point deserves honesty. McDonald's never published a specific figure for how much revenue the personalized boards generated. The company spoke of higher average order values and then acted on that. When you equip thousands of additional locations after a pilot, you've clearly received your answer internally. But no publicly verified percentage exists — and anyone who quotes one has made it up.
Why McDonald's Sold Dynamic Yield and Still Uses It
2022 brought the second surprise in this story. McDonald's sold Dynamic Yield to Mastercard — and has continued using the technology ever since, with plans to extend it to additional ordering channels and further markets.
At first glance, that seems contradictory. On second look, it's the logical next step. McDonald's had spent three years proving the technology worked in its own business. But a burger company is not a software company. Continuing to develop a platform used by banks, retailers, and travel providers is a different business from running restaurants. So the company handed the platform to an owner who could scale it — and remained a customer.
For every other company, there's a reassuring message in that. You don't need to acquire a technology firm to benefit from its technology. The platform that McDonald's once owned exclusively is now available as software for any company. The entry price is no longer a $300 million acquisition — it's a project.
Three Lessons from McDonald's Use of Dynamic Yield
What can you take from this story if you're not running tens of thousands of restaurants? Three things, from our perspective.
The first lesson is about where to start. McDonald's began with the channel that drives the most revenue. Not the website, not a prestige project — the drive-thru. Translated to your business, the question is: which of your channels has the highest volume but still shows every customer the same thing today? That's your drive-thru.
The second lesson is about data. Getting started didn't require a perfect customer database — just three simple signals. Nearly every company already has signals like these: time of day, location, device, traffic source, shopping cart contents. Companies that wait until the big data platform is ready usually wait too long. Companies that start with what they have learn immediately.
The third lesson is about sequence. Pilot, measure, roll out. Not the other way around. That protects against the most expensive kind of digital project — the large ones built on assumptions. We've written about how to build that measurement culture in detail in the article Stop Going with Your Gut.
"We're Not McDonald's" — the Objection That Always Comes Up
At this point in any conversation, we almost always hear the same line: we're not McDonald's. That's true — and it's not a counterargument. The underlying mechanics don't depend on scale. A mid-sized online retailer has the same time-of-day patterns, returning customers, and shopping carts as a global corporation. It even has an advantage: decisions can be made faster, and a pilot doesn't require sign-off from three layers of management.
What mid-sized companies genuinely tend to lack is the habit of seeing ordering situations as something that can be changed. The menu board, the homepage, the category page, the newsletter — these are treated as fixed. But they're anything but. They're exactly the levers that personalization works with.
What This Playbook Means for Your Personalization
If a company with tens of thousands of restaurants can personalize its ordering channels within eighteen months, the excuse of complexity starts to wear thin. The real question isn't whether personalization works in your industry. McDonald's answered that for an industry few would have expected it from. The question is which of your touchpoints is still showing every customer the same thing today — and what that's costing you every single day.
We show what a first step looks like without a million-dollar investment on our Dynamic Yield page. You'll find further analysis in our Dynamic Yield Insights.