Customer loyalty: the before and after of paying for the same guests every month

Measure frequency and LTV before you touch a single piece of content

Your first deliverable is not a Reel, it is a sheet with three numbers: how often your average guest eats with you each year, how much they leave per visit and what it costs to bring them in. That Peruvian restaurant in Bogotá billed well and still bled cash because its acquisition cost had climbed from 4.10 to 9.80 dollars in fourteen months while the check moved two points. Pull the last twelve months from your POS, match by phone or email, and work out average frequency, average check and LTV, which is those two multiplied by the years the guest stays. Sector retention sits near 55% according to Restroworks, and annual repeat purchase in full-service barely reaches 26% according to the National Restaurant Association: if you are below that, traffic is not your problem. It is done when those four numbers are written and dated, and you can repeat the calculation next month without asking anyone how.

Split off the recurring 20% and make it the core of your funnel

Sort your customer base by visit count and mark the top quintile: that 20% usually explains between half and two thirds of an independent restaurant's cash, and it is the only audience worth talking to daily. With the list in front of you, label three stretches —reach, repeat, referral— and give each its own channel, because the numbers look nothing alike. SMS opens around 98% and gets read within one to three minutes according to Constant Contact, with 18% clicks according to Tabular and 45% response against email's 6% according to Omnisend; restaurant email opens 43.6% according to Stripo yet converts at 1.06% clicks according to Mailchimp. Deliverable: a file holding the recurring quintile's phone numbers, verified consent and the content promise they will receive every week. Without that segmentation, everything you send afterwards is expensive paid media dressed up as loyalty. Retention content has one concrete job: to occupy space in your customer's memory while they eat somewhere else.

Design the creative pillar that holds the 27 days your guest stays away

A Reel of the cook boning the leg at six in the morning does not sell that dish, it installs a reason to come back, and that reason outlives the algorithm for three months while the two-for-one dies the day you switch it off. Pick four pillars and no more: kitchen process, the people on your team, product origin and the backstage of a full service. Instagram delivers 2.2% engagement against Facebook's 0.22% according to Restroworks, so that is where the muscle goes. At Masterestaurant we run this as a calendar of twelve monthly pieces with one control figure per stretch, never as loose inspiration. The deliverable is a four-pillar script with two filmable examples each and the team member responsible for filming them, by name, not by job title. Paid media rents attention that gets pricier every quarter; an owned base of phone numbers and emails cannot be taken away from you by any platform.

Turn the follower into owned data before your reach goes dark

Install capture at three physical points —a table QR, wifi with registration and the printed check— and promise something a guest genuinely wants: the dressing recipe, the seasonal dish alert two days ahead of everyone else, the Friday table held. No automatic discounts. A discount buys one visit; belonging buys the Monday conversation at the office table, which is advertising you never paid for. Track weekly capture against guests served: under 15% the system feeds nothing. And if Meta halves your organic reach tomorrow, what is left? The list is left, or nothing is. Deliverable: a base growing with sign-up date, source channel and written consent. Three monthly SMS contacts plus one email is where most independent restaurants sustain repeat visits without burning the base. Sequence it this way: one message of pure value with no offer, one carrying menu news and one inviting with a specific hour, because content arrives when the guest opens the app and is choosing where to eat on Friday, not when it suits you to send it.

Build the contact cadence and cap the frequency

With 98% SMS open rates according to Textellent, the risk is not that they miss you, it is that they hear from you too much and unsubscribe. Watch the weekly opt-out rate and pull back above 2%. I got this wrong for years, pushing promotion into every message, and the base eroded month by month until it stopped answering. It is done when the cadence is scheduled with dates and an owner, and a written rule states what gets cancelled if opt-outs spike. The stamp card rewards visit ten and teaches your customer to wait; the content system rewards from visit one and teaches them to belong. Replace the stamp with access: a closed tasting for thirty people, the new dish tried before launch, the chef's table once a month. It costs less than a discount and it does not erode margin. If you use gift cards, know that roughly 6% of the value is never redeemed according to Capital One Shopping, income that exists but must not become your model.

Close the loop with rewards for belonging, not for transaction number ten

The control figure is simple: frequency of the recurring quintile before and after six months. A returning guest costs between five and seven times less than a new one, and that gap is exactly what pays for your content team. Deliverable: two access-based reward formats, each with capacity, cost per guest and a date on the calendar. The mistake I run into again and again is measuring retention content with the yardstick of acquisition: the owner demands sales attributed to a process Reel and, finding none, kills the one lever that was lowering cost per guest. Second comes handing filming to an agency that has never walked into the kitchen, so the material looks pretty and reads foreign. Third is speaking to the whole base as if it were one person, when the recurring quintile and the one-visit customer need different messages. Fourth is loading an offer onto every contact, and here Google Ads converts 7.1% in restaurants according to WordStream while restaurant email clicks 1.06% according to Mailchimp: those are channels with opposite trades, not rivals.

The four mistakes that sink this and how to dodge them

Fix it by assigning one figure per stretch, one internal owner with a phone number and a calendar nobody can improvise on the day. Check seven things and sign the date. One, the four starting numbers exist —frequency, check, LTV and acquisition cost— with a monthly cut. Two, the recurring quintile is identified by name and phone with consent. Three, four content pillars are running at twelve pieces a month with a named owner. Four, base capture clears 15% of guests served every week. Five, the cadence of three SMS plus one email is scheduled and opt-outs stay under 2%. Six, the access reward has capacity, cost per guest and a date. Seven, by month six the recurring quintile's frequency went up, and that movement is the only thing validating everything above it. If any line is missing, do not build more content: finish that line first. Open your POS today and pull visits per customer for the last twelve months.

Five differences that decide whether the guest returns

A discount competes against the restaurant next door; content competes against a guest being bored on their couch, and there you play alone. That is why one well-made process Reel sustains repeat visits for months while the 2-for-1 dies the day you switch it off. The punch card rewards transaction number ten; the content system rewards belonging from visit number one, and belonging is what a guest recounts at the office table on Monday. Your monthly email lands when it suits you to send it; content lands when the guest opens the app, which is precisely the moment they are choosing where to eat on Friday. Paid media rents attention and reprices every quarter; an owned base of followers and subscribers is an asset that appreciates, and it gets valued as such when the business sells. Discounting drops average check and margin at once, a double hit almost nobody models; the non-monetary benefit —a table held on a Friday, the first cut of the ossobuco— costs cents and reads as status.