By Anton Denisov
At a large fast-food chain, customers order in three ways: in the app, at a self-service kiosk, or at a counter staffed by a person. I was responsible for the first two, along with pricing and the loyalty program, and I reported on them consistently each quarter. The metric that opened every management review was the share of sales coming through digital channels. It went up year after year, and management treated that rise as the main result of the transformation.
People stick to the way they are used to ordering, so a rising digital share means a lot of them have changed their habits. The number counts users, and it says nothing about whether those orders cost less to serve, whether the same customers buy more often, or whether that demand would have arrived at the counter anyway. Those three answers determine whether digital growth improved the business.
What it takes to move customers into the app
Moving a large share of order volume into a digital channel raises problems that have little to do with technology. Customers give up a habit they are comfortable with, and when an app crashes during an order, it sends them back to the counter, and most of them do not try it again. The app has to connect to payments, the kitchen, and the delivery fleet, and customers notice within a day if prices or promotions differ between the app and the counter. Restaurant teams work to a different set of steps, and someone has to train them. Any one of these can stall the move into digital.
The benchmark a company chooses at this stage shapes what it ultimately builds. Customers open a restaurant app once or twice a month, and it has to feel just as quick and effortless as the apps people use every day if it is going to become a habit. We rebuilt the ordering journey to that standard and used recommendation models to personalize the upsell for each customer. The app became the leading product in its category in its market, ahead of the delivery aggregators.
That is the work behind a rising digital share, and once adoption reaches this scale, leadership needs a more demanding set of questions.
Four types of digital growth
Growth through a digital channel reaches the business in four ways, each showing up in a different place on the P&L.
| Type of growth | Where it comes from |
| Efficiency | An existing order handled at a lower cost to serve |
| Behavioral | A customer buying more, or buying differently |
| Relationship | Direct contact with the customer and the data behind it |
| Market | A customer or a moment in the day the business could not reach before |
Efficiency growth occurs when a customer who would have bought anyway buys through a channel that costs the company less. The order goes through a kiosk instead of a cashier, so it no longer takes anyone’s time behind the counter. Revenue stays where it was, and cost to serve falls. Queues get shorter, the restaurant gets more people through lunch, fewer orders come back wrong, and the customer now has a name attached to the transaction. Companies write this off as movement rather than growth, which underestimates the impact on margins.
Behavioral growth happens when the digital product changes what people buy. The average check at our kiosks was 20 percent higher than at the counter, partly because the interface offered extras and partly because there was nobody else in the transaction. Customers add a dessert more freely when no one is watching them make the decision.
Relationship growth happens when the company gets a direct line to the customer, along with first-party data and the ability to reach that customer without paying an intermediary. The value shows up in moments the channel metrics never capture. One UK lettings group that grows by acquiring agencies uses its ownership of client communications to spot landlords going quiet or turning terse, and intervenes before the complaint is ever filed. Churn through those integrations runs at a fraction of what the market loses in a typical acquisition.
Market growth occurs when the channel generates demand that did not exist before. Delivery is the clearest case, because the customer who was not going to make the trip that evening orders anyway.
The difference shows up when demand shifts for reasons the company had nothing to do with. During the pandemic, for example, orders moved into delivery and mobile ordering on their own, and companies that recorded that period as behavioral or market growth built the next year’s plan on a base that was never theirs.
Every channel needs a defined role
An aggregator, an owned app, a kiosk, and a staffed counter do different jobs, and judging them by the same standard leads to bad investment decisions.
An aggregator buys reach. It puts the brand in front of customers that the business cannot reach on its own, charges a commission for that, and maintains the relationship with the customer. An owned app costs far more to build but returns loyalty, first-party data, personalization, and control over how offers work. The case for a kiosk lies in conversion, average check, and channel margin within restaurants that the company already pays rent on. The counter handles interactions that require a person, including complaints and anything emotionally charged.
Each channel should carry a defined job and a revenue target, and the two must match. A kiosk, measured on average check and channel margin, is managed differently from an app, measured on repeat orders and first-party data. Without that, investment follows whichever channel reports the strongest quarter, which is usually the one running the deepest promotions.
Incentives should grow with the channel
Discounts make sense early. They lower the barrier to a first digital order, pull customers out of the old journey, build the user base to a size where personalization works, and produce the behavioral data that targeting later depends on.
Promotions designed for the early stage often outlive it. A discount that once bought a new habit ends up paying for orders that would have happened anyway, and rising volume through the channel hides the cost because the reporting shows growth. Our app reached that point, and a large share of its usage was driven by promotions rather than by the product. We tied the incentives to a commercial outcome instead of cutting them.
As a channel matures, the promotional strategy has to mature with it. That means segments rather than blanket discounts, different rewards for different customers, control groups, measuring what people do after an offer ends, and incremental contribution margin as the test of whether an offer worked.
Loyalty shows you who is buying before it explains why
A loyalty program does something discounts do not: it tells the company who is buying. Close to 40 percent of our checks came with a customer attached, and once four of ten receipts carry a name, the company can track who comes back, how often, and what they add to the order the second time.
Economics comes next. Customers in our program showed a lifetime value twice that of unidentified customers. The number needs careful reading, because the customers who join a loyalty program first are usually the ones who were already buying most often. Some of that gap belongs to the program, and some of it was there before anyone signed up.
Separating the two requires a specific set of checks, from frequency relative to a baseline to average contribution after accounting for rewards costs, and what customers do once the rewards stop.
Four groups of measures
A channel needs four groups of measures, each answering a different question. Together, they cover whether enough customers moved into the channel, what it does to the customer, what it does to the P&L, and what it is worth to the company in five years.
Adoption shows whether the channel has reached scale: digital revenue share, active users, transaction volume, frequency, and share of identified transactions.
Customer value covers whether the experience improved: conversion, repeat rate, retention, basket size, speed, and satisfaction.
Economic quality is what the channel does to the business model: contribution margin, cost to serve, promotion cost, incremental revenue and frequency, cannibalization, and payback period.
Strategic value is what the channel is worth over the years: ownership of the customer relationship, first-party data, ability to personalize, exposure to platform dependency, scalability, and room to build new commercial mechanics on top.
The four groups will often point in different directions. Orders through the app can go up every quarter, while the margin on each one goes down, and the growth number alone will never show that.
The questions a board should ask
The board can ask these seven questions during the review that sets the channel’s budget. The answers show whether the channel has earned that budget or whether the budget has simply carried over from last year.
- What kind of value is each channel expected to create, and does its target reflect that?
- Is the channel creating new demand, changing behavior, or making existing demand cheaper to serve?
- How much digital growth survives when incentives are reduced?
- What does the channel do to the contribution margin and cost to serve?
- Does it strengthen or weaken the company’s ownership of the customer relationship?
- Are product teams accountable for both economics and adoption?
- Are we running a mature channel on the metrics we used at launch?
Digital adoption demonstrates that customers have accepted the channel. The company now has to give each channel its own target, built from cost to serve, behavior, the customer relationship, and margin, alongside its revenue number.
