Most Asian restaurant owners can tell you exactly how many new customers walked in last month. Far fewer can tell you how many of last year's customers never came back — or why. That second number is usually the bigger one, and it's the one nobody tracks.
Industry-wide, repeat guests drive roughly 60% of restaurant revenue, based on an analysis of more than 100 million guest records by online ordering platform Olo. Toast's Regulars Report 2026, drawn from Toast's own platform transaction data, found that just 7% of guests can account for up to half of a restaurant's order volume. The gap between restaurants that keep those guests and restaurants that don't is not a marketing problem. It's a systems problem — the same kind of systems problem that shows up in labor costs and table turnover, except this one shows up quietly, on a delay, in a number most restaurants never calculate: how many of the people who ate here in January are still eating here in July.
This white paper is about that number, and about what actually moves it: how a restaurant tracks repeat visits in the first place, what happens to the customer relationship when most orders arrive through a third-party delivery app, and what changes when a restaurant puts a loyalty program, a CRM, and targeted marketing on top of its point-of-sale system instead of leaving each of those as a separate, disconnected tool. None of this requires a restaurant to abandon delivery apps or spend heavily on advertising. It requires knowing which guests are regulars, reaching them directly, and giving them a reason tied to the restaurant itself — not a coupon that trains them to shop around.
Every figure in this paper is attributed to a named source: Olo's guest-data analysis, Toast's Regulars Report 2026, Resy's multi-year reservation data, Restroworks' restaurant retention research, OPA!'s delivery-commission analysis, Harvard Business Review's research on acquisition versus retention costs, and Chowbus's own disclosed platform data. Where a number doesn't have a source attached to it, it isn't in this paper.
Chapter 1 looks at why repeat visit rate is the metric most restaurants should be tracking and mostly aren't. Chapter 2 covers what happens to that number when a growing share of orders comes through third-party delivery apps. Chapter 3 covers what a real loyalty and CRM system changes about repeat visit rate, with data from restaurants that have implemented one. Chapter 4 covers how automated, data-driven marketing — SMS, email, review management — turns the customer list a restaurant already has into actual return visits. The conclusion lays out a 90-day plan an owner-operator can start this month.
Ask a restaurant owner how business is doing and the answer is almost always about sales, covers, or foot traffic — numbers a POS system already surfaces on a daily dashboard. Ask the same owner what percentage of last month's diners came back this month, and most don't know. It isn't tracked because it isn't obvious where the number would come from, not because it doesn't matter. It matters more than almost anything else in the P&L.
According to research from restaurant technology firm Restroworks, industry-wide restaurant customer retention averages around 55%, well below the roughly 75% average retention rate reported across industries broadly. Within food service, format matters: quick-service restaurants retain closer to 71% of customers, fast-casual concepts around 68%, casual dining around 64%, and fine dining closer to 51%. The same research found that roughly 70% of first-time diners never return to a restaurant at all — meaning for most independent restaurants, the large majority of the marketing and word-of-mouth effort that brought someone in the door produces exactly one visit and nothing after it.
The revenue math behind this is not subtle. Restroworks' analysis puts repeat customers at 65–80% of total restaurant sales, with new customers contributing only 20–30%. Existing customers also spend more per visit — Restroworks found repeat guests spend an average of 67% more per order than first-time visitors, since they already trust the menu, know what they like, and are more open to trying higher-margin items a server or kiosk suggests. Toast's Regulars Report 2026 found the same pattern from a different angle: regulars are 34% more likely to spend more per check and 80% more likely to try a new menu item than a first-time guest, simply because comfort with a restaurant removes the hesitation that keeps new guests ordering the safest thing on the menu.
Put those two data points together and the implication is direct: a restaurant that improves its repeat visit rate by even a few percentage points is not chasing a marketing vanity metric. It's moving the line item that already represents the majority of its revenue.
The reason this number goes untracked is structural, not a failure of attention. A restaurant's POS records a sale. It usually does not record whether the person paying is the same person who paid three weeks ago, because most POS systems have no reliable way to link one transaction to the next — especially when a meaningful share of orders arrive through third-party delivery apps that never share the customer's identity with the restaurant at all. Without a shared customer ID across dine-in, pickup, online ordering, and delivery, "repeat visit rate" isn't a number a restaurant can pull. It's an estimate at best, and most owners don't even attempt the estimate.
This is the same disconnected-systems problem that shows up elsewhere in restaurant operations — a manager checking one dashboard for sales, another for labor, a third for online reviews — except retention data is harder to notice missing, because there's no single moment where its absence causes a visible failure. A kitchen running out of an ingredient is obvious within the hour. A slow erosion in repeat visit rate over two quarters is not obvious until a slow season turns unexpectedly bad, and even then it's easy to blame the season instead of the churn that had already been building underneath it.
Restaurants that do track this number tend to set a working target of 30–40% repeat visit rate as a healthy baseline, per Restroworks' research, with quick-service and fast-casual formats reasonably aiming higher given faster order cycles and lower average check sizes. The goal isn't a specific number in isolation — it's knowing the number at all, tracking it monthly the same way sales and labor cost percentage are tracked, and treating a downward trend as seriously as a downward trend in covers. A restaurant that can see its repeat visit rate slipping from 34% to 29% over two quarters has three months of runway to respond. A restaurant that can't see the number at all finds out only when a slow season turns into a genuinely bad one.
The economics in this chapter apply to every restaurant format, but they land with particular weight on Asian restaurants, for a structural reason: a large share of the format is built around repeat, communal occasions rather than one-off transactions. A hot pot restaurant's busiest table on a Friday night is often a group of six or eight returning together, not six individual first-time guests who happened to arrive at once. A bubble tea shop near a college campus or office park depends on the same handful of nearby regulars ordering multiple times a week, not a constant stream of new discovery traffic. A family-run Chinese, Korean, or Vietnamese restaurant frequently serves multiple generations of the same families over years, sometimes decades — a relationship a generic retention statistic doesn't fully capture, because the loyalty already exists culturally. The gap is that this loyalty rarely gets captured systematically: the restaurant knows Mrs. Chen's family orders the same three dishes every visit, but that knowledge lives in a server's memory, not in a system that can proactively reach her family when they haven't come in for six weeks, or recognize a new customer of the same background who deserves the same kind of attention.
A restaurant doesn't need a full CRM system to start estimating repeat visit rate — it needs a way to link one transaction to the next. The simplest starting point is phone number: any restaurant that already collects a phone number for online ordering, waitlist, or reservations can pull a month of orders and count how many phone numbers appear more than once. It won't be perfectly accurate — group orders, shared family accounts, and cash-paying dine-in guests won't show up cleanly — but it turns an unknown into a rough, trackable number, and a rough number tracked consistently month over month is more useful than a precise number nobody ever calculates. A restaurant already running a loyalty program has a cleaner version of the same signal available immediately: loyalty scan rate at checkout, and the percentage of enrolled guests who scan again within 30 or 60 days, function as a close proxy for repeat visit rate without any additional setup.
Most conversations about third-party delivery apps focus on the commission — DoorDash, Uber Eats, and Grubhub each take a cut that ranges roughly from 15% to 30% per order, according to delivery-cost analysis published by restaurant technology firm OPA!. That commission is real and it matters. But the commission is the visible cost. The customer relationship a restaurant loses on every one of those orders is the cost that doesn't show up on the monthly statement, and for a restaurant trying to build repeat business, it may be the larger one.
When a guest orders through a third-party delivery app, the platform — not the restaurant — collects the name, email, phone number, delivery address, order history, and ordering frequency behind that transaction. OPA!'s analysis of a representative 50-location, 20-orders-per-location-per-day operation found the restaurant loses roughly 360,000 customer profiles per month to the platform, or over 10 million records across three years, none of which the restaurant ever owns or can use directly. As the same analysis states plainly: the major delivery platforms do not share first-party customer data — names, emails, phone numbers, or order history — with the restaurants fulfilling those orders.
The practical effect is that a restaurant can run a third-party delivery channel profitably in isolation and still have no idea who half its actual customers are. It cannot text them about a new menu item. It cannot enroll them in a loyalty program. It cannot even tell whether the person who ordered pad thai last Tuesday is a first-time customer or someone who has ordered from the restaurant every week for a year, because the platform sits between the restaurant and the answer to that question — permanently, unless the guest also orders directly at some point.
There's a second-order effect that's less discussed but arguably more damaging. OPA!'s research notes that third-party platforms use the order data they collect to train recommendation algorithms that surface competing restaurants to the same customers — meaning a restaurant's own order history can end up promoting a nearby competitor to a guest who already ordered from them once. Subscription programs like DashPass, Uber One, and Grubhub+ compound this: they use a customer's order history to lock that customer into loyalty with the platform itself rather than any single restaurant, so a discount that looks like customer appreciation is functioning, structurally, as a tool that keeps the guest platform-loyal instead of restaurant-loyal.
This is why a restaurant's own delivery-commission percentage understates the real cost of over-reliance on third-party apps. The commission is a fee for a transaction. The lost customer relationship is a fee on every future transaction that guest might otherwise have made directly — at full price, without a platform intermediary, and with the restaurant able to invite them back.
OPA!'s analysis found that restaurants which maintain direct, first-party ordering channels see 15–22% higher average order values on those direct orders compared to third-party platform sales — a gap driven largely by the personalization and loyalty incentives a restaurant can only offer when it actually knows who it's serving. That's the case for pairing third-party delivery with a direct online ordering channel and a branded ordering app rather than treating delivery marketplaces as the only online channel: every order that comes in through a channel the restaurant controls is an order where the restaurant, not the platform, owns the resulting customer relationship. Chowbus's third-party delivery integration is built around the same principle from a different angle — pulling delivery-platform orders into the same system that handles dine-in and direct pickup, so a restaurant can at least see delivery volume alongside everything else instead of managing it in a separate app on a separate tablet, even where full customer-data sharing isn't possible.
None of this argues for dropping delivery marketplaces — for many restaurants they're a meaningful source of first-time trial, and the goal covered in Chapter 1 is turning trial into repeat visits, wherever the first visit came from. The point is narrower: a restaurant that treats delivery apps as its only channel for a large share of orders is, by definition, giving up the ability to build a direct relationship with a large share of its customers. The restaurants closing the retention gap are the ones actively working to move repeat guests — the same 7% who can represent up to half of order volume, per Toast's Regulars Report — onto channels where the restaurant can see them, reach them, and reward them directly.
OPA!'s analysis notes that restaurants commonly raise delivery-app menu prices by roughly 15–20% over dine-in pricing to offset commission costs — a workaround that keeps the restaurant's margin intact on paper but creates a second, subtler retention problem: a guest who orders once through a delivery app at marked-up pricing, then later notices the same dish costs less when ordered directly or in person, has a reason to feel the restaurant wasn't straightforward with them. The same analysis found 67% of diners say they'd prefer to order directly when pricing is equal, which suggests the preference for direct ordering already exists — most guests simply don't know a direct channel is available, or don't find it as easy to use as the delivery app already on their phone. This is a solvable awareness problem, not a demand problem, and it's a large part of why restaurants that promote their own direct ordering page on receipts, table signage, and social media tend to shift order mix toward direct channels over time rather than assuming guests will find it on their own.
"Loyalty program" often means a stamp card or a generic points app bolted onto a POS as an afterthought. The data on what a properly integrated loyalty and CRM system does to repeat visit rate is different enough from that image that it's worth being specific about what "properly integrated" means and what it changes.
Toast's Loyalty Impact Analysis, covering a sample period from mid-January through mid-April 2026, found that guests enrolled in a restaurant's loyalty program return at roughly 4 times the baseline rate — moving from around a 7% return rate for non-members to close to 30% for enrolled members. The same analysis found loyalty members retain at 2 times the rate of brand-new customers and 1.5 times the rate of the restaurant's general guest base. Restroworks' broader retention research, drawing on a wider set of restaurant loyalty programs, found members visit approximately 20% more frequently and spend approximately 20% more per visit than non-members, with mobile-based loyalty programs specifically associated with spending increases as high as 60%.
These numbers matter more than they might first appear, because they compound with the revenue concentration covered in Chapter 1. If repeat guests already represent 60–80% of a restaurant's revenue, and a loyalty program measurably increases how often those guests return and how much they spend per visit, the loyalty program isn't a side feature — it's acting directly on the majority of the restaurant's revenue base, not a marginal slice of it.
Restroworks' research found 57% of restaurants have implemented some form of loyalty or rewards program, and separately found that 57% of diners say they would spend more at a restaurant if a loyalty program existed — meaning roughly half of restaurants without one are potentially leaving spend on the table that guests themselves say they'd be willing to give. Quick-service loyalty enrollment specifically rose from about 36% of guests in 2021 to nearly 42% more recently, per the same research, suggesting the category is still in a growth phase rather than a mature, saturated one.
The practical failure point for most restaurant loyalty programs isn't the incentive structure — it's that enrollment happens manually, at the counter, dependent on a cashier remembering to ask during a rush, using a sign-up process slow enough that most guests decline. A loyalty and CRM system built into the POS removes that dependency: enrollment happens automatically at checkout, tied to a phone number or card already in use, with no separate app download or account creation required from the guest. Chowbus's loyalty and CRM platform is built on exactly this logic — the loyalty system reads directly from the same transaction data the POS is already recording, so a restaurant isn't running loyalty as a separate system that requires separate reconciliation, and isn't dependent on staff remembering to pitch it during a dinner rush.
A loyalty program's second function — separate from the direct spend increase — is that it's the mechanism that actually generates the customer contact data Chapter 4 depends on. A restaurant cannot send a targeted SMS to guests who haven't ordered in 45 days if it has no record of who those guests are or when they last visited. The loyalty enrollment step is what converts an anonymous transaction into a contact record the restaurant can act on later — which is also why loyalty enrollment rate, not just loyalty redemption rate, deserves to be tracked as its own number. A program with strong redemption but weak enrollment is only reaching the guests who were already regulars; a program built into checkout with high enrollment is capturing new customers into the same system from their very first visit, which is the difference between reacting to churn after it happens and having the contact information needed to prevent it. Chowbus's consumer-facing ordering app extends this further by giving enrolled guests a direct channel back to the restaurant that doesn't route through a third-party marketplace at all.
A loyalty and CRM system only works if the guest actually understands the offer, and for Asian restaurants serving communities where English is a second language for a meaningful share of regulars, a loyalty message sent only in English is a message a real percentage of the customer base will not fully act on, even if they open it. This is a retention gap layered on top of the retention gap: not just failing to reach a lapsed guest, but reaching them in a way they can't fully use. A system that supports enrollment and messaging in English, Chinese, Japanese, Korean, and Spanish addresses this directly rather than treating it as a nice-to-have translation layer bolted on afterward — it's the difference between a loyalty program that technically exists and one that an immigrant-owned restaurant's actual regulars can use the way it was designed to be used.
Collecting customer data through a loyalty program only pays off if a restaurant does something with it. This chapter covers the three channels — targeted messaging, promotions tied to actual behavior, and review management — that convert a contact list into measurable repeat visits, and why automation matters more here than in almost any other part of restaurant marketing.
SMS marketing data compiled by messaging platform Salesmessage shows text messages achieve roughly a 98% open rate compared to about 22% for email, with click-through rates on links sent by SMS running around 45% versus under 5% for email. For a restaurant with a lapsed-guest list — customers who haven't ordered in 30, 45, or 60 days — the difference between a channel that gets opened nearly every time and one that gets ignored four times out of five is the difference between a win-back campaign that works and one that quietly fails without anyone noticing, because a low-open-rate channel doesn't generate a visible complaint; it just generates silence.
The mechanism that makes this practical at the restaurant level is automation triggered by the same data a POS and CRM already hold: a guest who hasn't ordered in 45 days gets a message automatically, without a manager needing to remember to pull a list and send it manually. Chowbus's SMS marketing tools work directly off the same transaction data the loyalty system captures — a guest's last-visit date, order history, and enrollment status all feed the same trigger logic, so a lapsed-guest message and a birthday offer and a new-menu-item announcement can all run automatically off one shared customer record instead of three disconnected marketing tools each maintaining its own incomplete list.
A blanket 20%-off promotion sent to every customer on a list, regardless of when they last visited, spends the discount on regulars who would have ordered anyway and undercharges the exact group least likely to come back on their own. A promotion engine that segments by actual behavior — a stronger offer for someone who hasn't visited in two months, a lighter or no incentive for someone who ordered three days ago — spends the same marketing dollar more precisely on the guests actually at risk of churning. Chowbus's promotions engine is built around this segmentation logic specifically, targeting offers by visit recency and order history rather than sending the same discount to an entire list regardless of where each guest actually sits in the retention curve described in Chapter 1.
Review response is usually filed under reputation management, but it functions as a retention tool too. Research from local-search firm BrightLocal, cited in Olo's guest-retention analysis, found consumers are 41% more likely to use a business that responds to all of its reviews compared to one that responds to none — and a guest who left a mediocre review and then received a genuine, specific response is a guest a restaurant has a real chance of winning back, provided someone actually sees the review and responds to it before too much time passes. Chowbus's review management tools consolidate reviews across platforms into one place a manager can actually keep up with, which matters mechanically as much as it matters as a courtesy — a review sitting unanswered for three weeks on a platform nobody checks isn't a retention opportunity, it's a missed one.
Blanket promotions built for a typical fast-casual guest often don't map cleanly onto all-you-can-eat and hot pot formats, where the transaction is a group booking rather than an individual order, and the more useful lever is often a group-size or off-peak incentive rather than a percentage discount on food that's already priced as a fixed-rate experience. A promotion system that can target by party size, day-of-week, and time slot — filling a Tuesday 5:00 pm seating instead of discounting an already-full Saturday night — gives AYCE and hot pot operators a retention lever that fits the format instead of forcing a generic discount rule built for a different kind of restaurant onto a business model where the check size per person is largely fixed regardless of what's ordered.
The common thread across SMS, promotions, and review management is that each one only works as well as the customer data feeding it, and that data only exists because of the loyalty and CRM foundation covered in Chapter 3. A restaurant running four separate marketing tools — one for email, one for SMS, one for reviews, one for promotions — each with its own login and its own partial customer list, is running four smaller, less accurate versions of what one integrated system does with the complete picture. This is the same disconnected-systems cost that shows up in labor and table turnover, applied to marketing: the tools exist, but they don't talk to each other, so none of them sees the full customer relationship, and the restaurant ends up spending marketing effort on guesswork instead of on the actual visit history sitting in its own POS.
Closing a retention gap doesn't require a new product launch or a large marketing budget. It requires knowing the number, capturing the customer data that's currently falling through the cracks between disconnected systems, and using it — in that order. Here's what that looks like across the next 90 days.
Start by finding out what the current repeat visit rate actually is, even as a rough estimate, using whatever combination of POS transaction history, loyalty enrollment records, and phone-number matching is available. If there's no loyalty program in place, this is the month to launch one — not as a separate app, but built into checkout so enrollment happens automatically rather than depending on a cashier remembering to ask. The goal for month one isn't a dramatic increase in repeat visits; it's simply having, for the first time, a real number to track and a system in place to keep capturing customer contact data on every transaction going forward.
With a month of enrollment data in hand, segment the guest list by visit recency: guests who haven't ordered in 30-plus days, guests who visit regularly but haven't been offered anything, and guests who just made their first visit and haven't been invited back. Set up automated SMS messaging for the lapsed-guest segment and a targeted, behavior-based promotion rather than a blanket discount for the whole list. This is also the point to audit review response — pull every review from the past 90 days across platforms and respond to the ones still sitting unanswered, then set a standing process so new reviews get a response within 48 hours going forward.
By day 90, there should be enough data to see whether repeat visit rate moved and which segment responded best — often the lapsed-guest SMS campaign, since it's targeting guests who were already customers rather than trying to create demand from nothing. Double down on whichever channel performed best, cut whichever one didn't, and make repeat visit rate a number reviewed monthly alongside sales and labor cost percentage rather than a one-time project that quietly stops once the initial campaign ends. The restaurants that hold onto retention gains are the ones that keep watching the number, not the ones that ran one good campaign and moved on.
None of this requires the operational overhaul described in Chowbus's companion white paper on operational efficiency, though the two compound well together: a restaurant running faster table turns and lower order-error rates is also a restaurant giving guests a better reason to come back, and a restaurant that knows which guests are regulars can prioritize service and attention accordingly during a rush. Retention and efficiency aren't separate projects competing for the same hours — they're the same underlying system, viewed from two different angles.
Q1: What is customer retention rate for a restaurant, and what counts as a good one?
Customer retention rate is the percentage of guests who order again within a set window, typically measured monthly or quarterly. Industry-wide, restaurant retention averages around 55%, according to research from Restroworks, with quick-service restaurants closer to 71% and fine dining closer to 51%. Most restaurants treat a 30-40% repeat visit rate as a healthy working target, though the more important step for most owners is simply starting to measure the number at all, since the majority currently don't track it in any form.
Q2: How can an Asian restaurant increase repeat visits without discounting every order?
The most effective lever isn't a blanket discount — it's reaching guests who are already customers with a reason to come back before they lapse entirely. That starts with capturing contact information through loyalty enrollment at checkout, then using visit-recency data to send a targeted offer only to guests who haven't ordered in 30-plus days rather than discounting orders from regulars who would have returned anyway. Chowbus's loyalty and promotions tools are built to segment this way automatically off existing POS transaction data.
Q3: Is a restaurant loyalty program more effective than relying on third-party delivery promotions for retention?
Yes, for the specific goal of retention. Third-party delivery apps typically don't share customer contact information with the restaurant, so a promotion run through a delivery platform builds loyalty to the platform, not the restaurant. A restaurant's own loyalty program, by contrast, captures the customer's information directly, which is what enables win-back messaging, personalized offers, and repeat-visit tracking later. Toast's Regulars Report 2026 found enrolled loyalty members return at roughly 4 times the rate of non-members, a gap a delivery-platform promotion alone cannot close.
Q4: How much does a restaurant loyalty and CRM system cost compared to what it saves?
Costs vary by provider and restaurant size, but the more useful comparison is against what's already being lost: research cited by Harvard Business Review puts the cost of acquiring a new customer at 5 to 25 times the cost of retaining an existing one, and repeat guests already represent 60-80% of most restaurants' revenue. A loyalty and CRM system built into the existing POS, rather than run as a separate subscription with its own hardware and reconciliation, avoids adding a new standalone cost on top of a system the restaurant is already paying for.
Q5: My restaurant gets most of its orders through DoorDash and Uber Eats — can I still build repeat-customer relationships?
Yes, though it takes a deliberate push toward direct channels rather than assuming it happens automatically. Practical steps include promoting a direct online ordering link on receipts, packaging, and social media; offering a modest incentive for a guest's first direct order; and using in-store signage to move dine-in and pickup guests toward loyalty enrollment, since those are the visits where a restaurant already has full access to the customer relationship a delivery-app order would otherwise obscure.
Q6: What's the first step to improving customer retention this month?
Establish a baseline. Pull whatever transaction history is available — phone numbers, loyalty scans, or online ordering accounts — and get even a rough estimate of current repeat visit rate. If there's no loyalty program in place, launching one built into checkout, rather than as a separate app, is the single highest-leverage step, because it's the mechanism that generates the customer contact data every other retention tactic in this paper depends on.
%20(1)%20(1).png)
