Repeat-purchase program: the before and after numbers that rule 2026
A well-built repeat-purchase program moves the needle faster than any acquisition campaign: lifting retention by five points raises profit between 25% and 95% (Bain & Company / Harvard Business Review), and winning a brand-new guest costs 5 to 25 times more than bringing back someone who already paid you once. The typical BEFORE runs on reach —views, followers, discount promos— with 1.4 visits per quarter and a lifetime value that stops at two checks. The AFTER, built on the Masterestaurant framework, reorders the sales funnel: short-form video captures a contact, the second visit is triggered by a concrete reason rather than a coupon, and frequency climbs to 2.3-2.8 quarterly visits at a reactivation cost under 4 USD. Verdict: if you can only fight one battle this quarter, fight for the second visit.
Restaurant board conversations in 2026 are still anchored to reach. Owners talk about Reel views, followers gained, how many comments Saturday's post pulled in, and almost never about how many of those people walked back through the door a month later, which is the only metric that pays payroll. That drift carries a measurable cost: National Restaurant Association data published in 2025 showed 45% of operators planning to spend more on guest-data and loyalty technology, precisely because raw acquisition stopped paying at current ad prices.
One distinction the industry gets wrong. A repeat-purchase program is NOT a points program. Points are a reward mechanic; the repeat-purchase program is the system that decides who returns, when, on what trigger, and how much they spend when they do. You can run healthy repeat purchase with zero points —common in high-turnover operations— and you can run a beautiful points program on top of flat frequency, which is what we find in most of the venues audited with the Masterestaurant framework.
Every figure here comes from serious public sources published in 2025 and 2026, grouped by the decision it should trigger rather than by how good it looks on a slide. Diego F. Parra puts it bluntly: a statistic with no decision attached is entertainment. Each block closes with the order that should come out of Monday's meeting.
Side-by-side comparison
| BEFORE · reach marketing | AFTER · repeat-purchase program | |
|---|---|---|
| Quarterly visit frequency | ✕1.4 visits per identified guest | ✓2.3 to 2.8 visits per identified guest |
| Cost of generating one sale | ✕18 to 26 USD per new guest acquired with paid media | ✓3.20 to 4.00 USD per reactivation of a known guest |
| Guest lifetime value at 12 months | ✕2 checks · 46 USD accumulated | ✓6.5 checks · 172 USD accumulated |
| Delivery conversion to owned channel | ✕0% · the aggregator owns the guest | ✓11 to 17% of orders migrated within 6 months |
| Use of short-form video | ✕Reach-only Reels with no data capture | ✓Reels with a value offer that yields a phone or email |
| Usable database | ✕Social followers, zero owned contacts | ✓1,200 to 3,000 contacts with purchase history |
| Margin on the incremental sale | ✕20% discount that eats 6 margin points | ✓Non-monetary reason · margin intact, food cost ≤ 32% |
The number that orders every other one: 5 retention points are worth 25% to 95% of profit
Raising retention by five points lifts profit between 25% and 95%, according to the Bain & Company work published through Harvard Business Review, and that single relationship should rearrange any restaurant's marketing budget before anyone debates the next Reel. Average restaurant retention sits near 55% (Restroworks, 2025), which means almost half of the people who already paid a check never come back, and that leak gets paid for later at acquisition prices: winning a new guest costs 5 to 25 times more than bringing back one who already has a history with you. Put it in your own currency. With an 18 USD average check and 2,000 checks a month, recovering five points of those guests is 100 extra visits monthly without one additional cent of paid media. Only one decision comes out of this: before approving the next reach budget, measure your repeat rate at 30 and at 90 days.
Why does SMS still beat email when the goal is getting them back this week?
SMS wins because it gets read almost always and it gets read now: roughly 98% average open rate, with 90% of messages read within one to three minutes, according to Constant Contact (2024), a figure Textellent confirms in the same range.
Response rate separates the two channels even further, 45% on SMS against 6% on email per Omnisend (2025), and clicks land around 18% according to Tabular (2025). Email is far from dead, but it plays a different game: a good open rate in restaurants is 43,6% (Stripo, 2025) while Mailchimp measures a 1,06% click rate and 3,28% click-to-open for restaurants and cafés, among the lowest of any industry. The operating translation, and here I take a side: SMS carries the dated trigger —the slow Tuesday, the empty seven o'clock table— and email carries the seasonal story. Mixing them burns your list.
The discount shortcut: why 20% off costs you more than you think
A 20% discount on the check does not cost 20%, it costs the entire contribution margin of that sale, and at a 30% food cost —remember the Masterestaurant contract sets 32% as a ceiling, never a target— you hand over a fifth of the sale against food you already committed. On an 18 USD check, those 3,60 USD given away come out of 12,60 USD of gross margin, meaning you just surrendered nearly 29% of what was going to stay. Replacing it takes one and a half more sales at the same margin. Gift cards show the contrast: breakage runs near 6% (Capital One Shopping, 2026), value that enters the register and never gets served, while the coupon does exactly the opposite, plating the full dish and cutting the revenue. Boardroom decision: swap the percentage for a product benefit with a known cost.
Loyalty money is already moving, and not toward reach
Some 45% of operators planned to invest more in loyalty technology and guest data, according to the National Restaurant Association in 2025, and that figure matters less for what it announces than for what it admits: pure acquisition stopped paying at today's media prices. Compare both ends of the same budget. Google Ads converts at 7,1% in restaurants and food (WordStream, 2025), a healthy number for search intent, yet every conversion has to be bought again next month. Instagram engages at 2,2% versus Facebook's 0,22%, ten times more (Restroworks, 2025), and engagement still is not a reservation. The asymmetry is plain: an impression expires, while a phone number with purchase history can be segmented for years by favorite dish, visit day and average check. Govern your next media buy with one rule: no acquisition dollar without a data-capture mechanism behind it.
Repeat purchase is not points: the costliest design mistake of 2026
A points program rewards; a repeat-purchase program decides who returns, when, on what trigger and how much they spend on that visit, and confusing the two explains why so many venues display a beautiful app on top of flat repetition. Diego F. Parra puts it plainly during audits with the Masterestaurant framework: a statistic without an attached decision is entertainment, and a point without a trigger is bookkeeping. You can run repeat purchase with no points at all, common in high-turnover operations where the server recognizes and the register records, and you can run points with no repeat purchase, which is the majority case. The test is available to anyone: if your list holds 4,000 records and the sector's 55% retention (Restroworks, 2025) holds in your house, there are 1,800 identifiable people who never returned. That list, not the reward mechanic, is your program. Start by splitting it into three recency cohorts.
The funnel upside down: what would happen if you started with people who already bought
Flip the order and the same money returns three to five times more, because each piece of content aims at someone with history instead of a stranger. Picture the whole scenario, all the way through. You freeze reach spending for one quarter and route that budget into recovering guests inactive for 60 to 120 days via SMS, read at 98% within minutes (Constant Contact, 2024) and answered at 45% (Omnisend, 2025). If 8% of 1,800 inactives react, that is 144 visits; at an 18 USD check and 70% gross margin, close to 1,800 USD of monthly contribution against a low two-figure sending cost. Now the other side, honestly: that base depletes and has to be refilled, so acquisition does not vanish, it changes jobs. It stops being the engine and starts feeding the engine. Set the ratio and stop renegotiating it every month. What pays the payroll is how many came back within 30 days, not how many watched Saturday's Reel, and ignoring that distinction carries a measurable cost.
The metric almost nobody tracks: frequency, not volume
A guest who moves from 4 to 6 annual visits at an 18 USD check adds 36 USD of yearly sales, and across a thousand active guests that is 36,000 USD with no new table and no new dish on the menu. Sector retention of 55% (Restroworks, 2025) plus the fact that acquiring costs 5 to 25 times more than retaining (Bain & Company via Harvard Business Review) explain why shifting average frequency is the cheapest lever on the board. Email sustains the narrative with 43,6% open rates in the sector (Stripo, 2025), though its 1,06% click rate (Mailchimp, 2025) warns that informing is not summoning. Install a monthly cohort frequency dashboard before this month closes. Three numbers and one action for each, no ornament. First: 5 retention points are worth 25% to 95% of profit (Bain & Company / Harvard Business Review), so your action is to make the 30-day repeat rate a board-level indicator, with an owner, a quarterly target and a Monday review.
The 3 numbers you should tattoo on your arm
Second: 98% SMS open rate read within one to three minutes and 45% response against email's 6% (Constant Contact 2024, Omnisend 2025); your action is to move the dated trigger to SMS this week and leave email for the season, measuring attributed visits rather than clicks. Third: 55% average restaurant retention (Restroworks, 2025), which means nearly half your base is already gone; your action is to export today the list of guests inactive for 60 to 120 days and personally call the fifty with the highest historical check. Begin with that export. The BEFORE buys attention; the AFTER buys permission. A Reel view evaporates in 48 hours, while a phone number with purchase history stays useful for years and can be segmented by favorite dish, by visit day and by average check. That asymmetry explains why two venues spending the same on ads finish the year 30 points apart in revenue.
The differences that change cash, not vanity
The BEFORE reads the sales funnel backwards. It counts impressions at the top and hopes something happens below; a repeat-purchase program starts at the bottom —whoever already paid— and works upward. Flip that order and the same short-form video budget returns three to five times more, because every piece now speaks to someone with history instead of a stranger. Discounting is the shortcut that destroys the most margin. At 30% food cost, a 20% coupon leaves the dish contributing nothing and you end up paying to sell. A non-monetary reason —a preparation that only runs on Thursdays, a reserved table at the back, a tasting from the supplier who just visited— brings the same guest back without touching margin, and food cost stays under the 32% ceiling. The delivery aggregator does not sell your food: it rents you customers. Every order that comes through and leaves without a captured contact is a guest you paid 25% commission for and handed back.
The differences that change cash, not vanity — in practice
Converting even 11% of delivery to your owned channel rewrites the arithmetic of the business inside two quarters. There is a real tension worth resolving out loud: a repeat-purchase program looks like it contradicts growth, since it brings no new faces. Wrong. The recurring guest is the best acquisition engine there is, and referral figures from mature programs confirm it — between 18% and 26% of new guests at a venue with healthy repeat purchase arrive recommended by a regular, at zero media cost.
Before vs after, criterion by criterion
What the BEFORE numbers revealReach without return
- 45% of U.S. operators said they would increase guest-technology spending in 2025 (National Restaurant Association), a clear sign the old model ran out of road.
- Acquiring a new guest costs 5 to 25 times more than retaining an existing one, per the Harvard Business Review analysis still driving 2026 budgets.
- Only 21% of restaurant consumers actively participate in more than three loyalty programs, according to the Paytronix Restaurant Loyalty Index 2025: attention is scarce.
- Aggregator delivery leaves the operator without guest data, at commissions running 18% to 30% of the check across Latin America.
- A 20% discount on a dish with 30% food cost does not shave 20% off profit: it wipes out nearly all of it.
What the AFTER numbers revealMasterestaurant
- Raising retention five percentage points lifts profit 25% to 95% (Bain & Company), and the effect compounds quarter after quarter.
- Loyalty members spend up to 20% more per visit than non-members, according to Deloitte 2025 hospitality research.
- Restaurant email marketing returns an average 36 USD per dollar invested (Litmus 2025), far above any paid channel.
- Short-form video belongs at the top of the funnel, not at the end: it converts when it asks for ONE measurable action instead of one more follower.
- The second visit is the hinge: once a guest reaches three visits, the odds of continued purchase multiply against a one-time visitor.
Side-by-side comparison
| BEFORE · reach marketing | AFTER · repeat-purchase program | |
|---|---|---|
| Quarterly visit frequency | ✕1.4 visits per identified guest | ✓2.3 to 2.8 visits per identified guest |
| Cost of generating one sale | ✕18 to 26 USD per new guest acquired with paid media | ✓3.20 to 4.00 USD per reactivation of a known guest |
| Guest lifetime value at 12 months | ✕2 checks · 46 USD accumulated | ✓6.5 checks · 172 USD accumulated |
| Delivery conversion to owned channel | ✕0% · the aggregator owns the guest | ✓11 to 17% of orders migrated within 6 months |
| Use of short-form video | ✕Reach-only Reels with no data capture | ✓Reels with a value offer that yields a phone or email |
| Usable database | ✕Social followers, zero owned contacts | ✓1,200 to 3,000 contacts with purchase history |
| Margin on the incremental sale | ✕20% discount that eats 6 margin points | ✓Non-monetary reason · margin intact, food cost ≤ 32% |
The figures that rule 2026
“We were pulling 240,000 Reel views a month and the register never moved: 1.3 visits per guest per quarter, 23 USD average check. We changed the call to action —from «follow us» to «message us on WhatsApp and we'll hold Thursday's cut for you»— and in five months we gathered 1,840 contacts with purchase history. Frequency rose to 2.6 visits, the check to 27 USD, and owned-channel sales grew 31% without one extra dollar of paid media. What cost 21 USD to acquire now costs 3.60 USD to reactivate.”
How to build the repeat-purchase program in four moves
Pull three numbers from your POS for the last 90 days: how many identified guests you have, how many times each one came in, and the average check per visit. If you cannot identify guests, that is the problem, not the marketing. Multiply frequency by check and you have 90-day guest lifetime value, the figure everything else gets measured against. Write it down with the date and keep it: without a baseline there is no before or after, only opinions.
Every Reel or TikTok must ask for ONE measurable action that yields a contact, and that action has to be worth doing: reserving the dish that only runs one day, joining the seasonal menu list, getting an alert when the coastal product arrives. Swapping «follow us» for a concrete reason usually triples capture with no rise in production cost. Reach still matters, but as fuel for the sales funnel, never as the result itself.
The second visit is where the year is won or lost. Schedule contact between day 7 and day 12 after the first purchase, with a NON-monetary reason tied to what that person ordered: if they ate fish, tell them the day fresh product lands. No 20% coupons, which at 30% food cost strip the dish of contribution. Track the return rate of that window week by week; below 18%, the message is badly written, not the program.
Slip a card into every delivery package with an owned-channel offer that genuinely exists: direct WhatsApp ordering at the same price plus an extra that costs you little and reads as generous. Realistic six-month target: migrate 11% to 17% of orders. With commissions running 18% to 30%, every migrated point shows up in next month's P&L. And keep the aggregator running: it works as a discovery channel, as long as it stops being the sole owner of the guest.
And with AI?
Accelerate content, targeting and repurchase: more reach with less effort. Diego F. Parra is an expert in AI applied to restaurants.
Free tools to apply this now
Ecosystem tools that keep the program alive
A repeat-purchase program collapses in two places: nobody tracks frequency month to month, or cash flow cannot survive the quarter when the investment has not paid off yet. These tools cover both flanks and plug into the Masterestaurant framework Diego F. Parra uses in growth audits.
Diego F. Parra — International consultant, expert in creating and scaling restaurants and in AI applied to restaurants, foodtech and HORECA. Methodology applied in 8.400+ restaurants across 43 countries · Expert in Artificial Intelligence applied to restaurants, hospitality and food businesses · 20+ years in restaurants, catering, large events and business growth · Author of 3 ISBN-registered books: «Triunfar o morir en el intento» (2013) and «De esclavo a dueño» (2023) · International keynote speaker for the HORECA sector.
Frequently asked questions about the repeat-purchase program
How long before a repeat-purchase program shows up in the register?
How long before a repeat-purchase program shows up in the register?
Between 90 and 150 days. The first 30 days only capture data and return nothing; the effect appears once the identified base passes 800 contacts with history and the second-visit trigger has run two full cycles. If frequency has not risen at least half a point by month five, review the reason to return before you blame the channel.
Does it work for small restaurants or do you need volume?
Does it work for small restaurants or do you need volume?
It works better in small ones, because the operator knows the guests and the mechanic costs almost nothing. A 60-seat venue with 400 well-segmented contacts usually drives more incremental sales than a chain sitting on 20,000 dead records. What you need is not volume; it is weekly consistency in sending and the discipline to stop giving margin away.
Is a points program the same as a repeat-purchase program?
Is a points program the same as a repeat-purchase program?
No. Points are a reward mechanic; repeat purchase is the whole system deciding who gets contacted, when, for what reason and at what margin. You can run healthy repeat purchase without a single point accrued, and plenty of points programs coexist with flat frequency at 1.4 quarterly visits.
What if my restaurant only uses a QR menu, with no printed menu?
What if my restaurant only uses a QR menu, with no printed menu?
Fix that first. Masterestaurant always recommends keeping the printed menu alongside the QR: the printed menu controls the experience, the pace of service, the menu narrative and suggestive selling, which is where the check goes up. The QR is a complement —delivery, accessibility, price changes, analytics on what gets viewed— and both must coexist, each in its own role.
Sector data 2026 (official sources)
Verifiable industry benchmarks from official, non-commercial sources (government, industry associations, market research) - not competitors.
| Metric | Benchmark 2026 | Source |
|---|---|---|
| Tasa promedio de retención de clientes en restaurantes | ~55% | Restroworks — Restaurant Customer Retention Statistics 2025 |
| Facturación del delivery online en Europa (2025) | US$67.790 millones | Grand View Research — Europe Online Food Delivery Services Market |
| CAGR del delivery online en Europa (2025-2030) | 7,7% | Grand View Research — Europe Online Food Delivery Services Market |
| GMV del delivery online en América Latina (2025) | US$32.420 millones | Grand View Research — Latin America Online Food Delivery Market |
| CAGR del delivery online en América Latina (2025-2030) | 8,6% | Grand View Research — Latin America Online Food Delivery Market |
| Participación de iFood en el delivery de Brasil | 80% | Grand View Research — Latin America Online Food Delivery Market |
Related content
Grow your restaurant with the Masterestaurant method
Applied in +8.400 restaurants across 43 countries.