Technology

Your drive-thru isn’t broken; it’s invisible

Mike MacLennan, co-founder and co-CEO of Arc, a voice AI platform that helps restaurant operators measure and optimize the drive-thru, explains how to upsell at the drive-thru.

Photo: Arc

September 17, 2026 by Michael MacLennan — Co-CEO & Co-Founder, Arc

Summarize with

The first time I sat near a fast-food drive-thru with a notebook and no food order, a group of elderly women knocked on my window. They thought I was panhandling.

I wasn't. I was doing market research.

I spent months in drive-thru lines across the country, simply listening. Hundreds of orders. Dozens of locations. Different brands. Different markets. What I heard changed how I think about the drive-thru.

For most quick-service restaurants, the drive-thru generates roughly 70% of revenue. Yet it has long been one of the least measurable parts of the business. Operators could track mobile orders and loyalty activity in remarkable detail, but the conversations happening at the order post remained largely invisible. There was no transcript and no practical way to understand at scale what guests were experiencing.

What I kept hearing was this:

"What did you say?"

"Can you repeat that?"

"Is that all?"

Crew members weren't being rude. They were working under intense pressure, measured on speed and expected to move cars as quickly as possible. The result was a rushed, transactional experience repeated hundreds of times every day, with little visibility into where conversations were breaking down.

One listening study illustrated just how common those breakdowns had become. Across hundreds of drive-thru orders, crew members made dozens of unprompted apologies. They weren't apologizing for poor service. They were apologizing because a menu board was outdated, an order was difficult to hear or a guest was confused about a promotion.

Each apology seemed like an isolated moment. Viewed together, they revealed a pattern restaurant leadership otherwise had little chance of seeing.

There's something about Mary

Somewhere in the middle of all that listening, I met Mary.

She worked at a Burger King and a KFC across the street, and regulars sought her out at both locations. She knew customers by name. She remembered their orders. She was warm, unhurried and genuinely delightful. She was the exception, not the rule.

What struck me wasn't just how good she was. It was that everything that made her exceptional disappeared the moment her shift ended. The institutional knowledge she carried existed entirely in her head. There was no way to identify her most effective habits, teach them to others or measure their impact.

Almost every restaurant has a Mary. The challenge is that most brands can't explain exactly what makes these employees exceptional. Without visibility into customer conversations, the behaviors that create loyal guests remain personal strengths instead of organizational knowledge.

Listening to hundreds of drive-thru conversations revealed something else. It also challenged one of the industry's most common assumptions about upselling.

The upsell myth

One insight kept surfacing. Less than 10% of orders included a meaningful upsell or menu suggestion. And when upsells did happen, they were almost always the same generic offer tacked on at the end of the order, like, "Do you want to add an apple pie?"

I call these Hail Mary upsells because they're delivered at the last possible moment with little context or personalization. Unlike the Mary I mentioned earlier, they rarely delight anyone. They convert at just 4% to 5%.

That isn't a crew problem. Crew members are measured primarily on speed, not the quality of their recommendations. After hearing "no" repeatedly, many simply stop trying.

The biggest opportunity is making recommendations that feel relevant. When suggestions are delivered at the right point in the conversation and connected naturally to what a guest has already ordered, conversion rates can reach 20% to 40%.

One restaurant operator believed a dessert recommendation delivered at the end of every order represented the strongest upsell opportunity. The numbers told a different story. The recommendation converted at just 5%, a classic Hail Mary. When it moved earlier in the conversation and connected naturally to what the guest had already ordered, conversion increased dramatically.

The operator had spent two decades in the industry and trusted his instincts. The data gave him a way to test them.

That distinction matters. Data doesn't replace experience. It reveals which instincts hold up across thousands of transactions and which ones deserve another look.

What listening revealed

Consider an iconic Southern comfort food brand with hundreds of locations. Before making any changes to its drive-thru experience, leadership wanted to understand what was already happening at the order post.

The ordering experience was warmer and more personal than many QSR brands. But listening uncovered missed opportunities to make relevant suggestions and recurring breakdowns that were easy to dismiss individually but significant at scale. Those insights informed changes to the drive-thru experience, including AI-powered ordering designed to make recommendations more relevant and timely.

In less than four weeks, average ticket sizes were up 3%, without any measurable impact on line speed. Guest satisfaction scores actually improved.

The same exercise can benefit other brands with already strong drive-thru operations. Speed and guest satisfaction scores tell operators whether an experience is performing well. They do not necessarily explain why it works or where incremental improvements may be hiding.

A restaurant hitting its target menu and window times for drive-thru orders may have little interest in overhauling its operation. But listening can still reveal which menu items generate questions, where ordering slows down and where there may be opportunities to suggest add-ons that guests actually want..

The lesson applies regardless of a brand's starting point. Underperforming drive-thrus can identify recurring friction. Strong operators can better understand what they are already doing well and where small adjustments could generate meaningful incremental revenue. In both cases, intuition becomes more useful when it can be examined against what guests and crew members are actually saying.

Start by listening

The drive-thru deserves the same level of operational rigor that restaurants apply everywhere else. Operators invest heavily in optimizing their apps, loyalty programs and third-party delivery channels because those experiences generate rich, measurable data. Yet the drive-thru, often a restaurant's highest-revenue channel, has largely run on instinct and institutional memory.

Today, operators have more ways than ever to study the conversations taking place at the order post. They can identify recurring patterns, test changes and understand the guest experience with greater precision.

Years ago, I sat in drive-thru lanes with a notebook because there wasn't another way to hear what customers were experiencing.

Today, there is.

My advice is simple: Before you change anything, listen.

What you hear might surprise you.

Technology

About Michael MacLennan

Prior to Arc, I led the Hardware team at Square, where I deployed millions of payments and POS products that powered everything from the farmers market to the Super Bowl. Prior to Square, I led private equity and product teams at Bain.

Connect with Michael: