Restaurant social media promotion: myth vs reality, measured against margin

Chapter 1 — What actually moves cash when a restaurant promotes on social?

Cash moves through the capture circuit, not through reach: the post creates discovery, but the money lands on your own destination, where the guest leaves an order or a booking.

Deloitte Digital documents +14.1% revenue at brands with the strongest social strategy and +9.9% direct B2C revenue in 2024; the gap between those two numbers is precisely the toll the platform charges when you have nowhere to land the demand. Run the margin arithmetic: with the sector operating between 3% and 9% net profit (Statista), one badly assigned marketing dollar eats a slice of EBITDA that no view count gives back. I work this with owners at Masterestaurant against a single line, attributable orders through owned channels, and I have seen beautiful dashboards coexist with sixty orders in ninety days. The lever is not better posting. It is the funnel that catches what the post wakes up. Finding a place to eat stopped being a search-engine act and became a social one, with numbers that no longer allow argument.

Chapter 2 — Discovery moved house, and the outside evidence confirms it

The 2024 MGH survey records 58% of respondents visiting a venue after seeing it on TikTok, against 38% in 2022: twenty points of displacement in two years, a pace no traditional channel has matched. Among younger diners the dependency is structural, since TouchBistro Diner Trends 2025 puts at 67% those who settle their choice by looking at social platforms, and Restroworks 2025 places at 41% those using TikTok as a restaurant search tool. Add the size of the market in play: US$473.49 billion projected for United States online delivery in 2026 (Statista Market Forecast, 2026). That said, none of those figures reaches your bank account on its own. They arrive when a circuit exists to collect them. Retention returns several times what a new customer costs to win, and budget allocation in most operations runs exactly backwards. Restroworks 2025 measures 67% higher spend per order from the returning guest versus the first-timer; Tabular 2025 calculates US$4.20 of revenue for every booking-confirmation SMS sent, with 97% of those messages read inside the first fifteen minutes.

Chapter 3 — Repeat business returns more per dollar than acquisition, and almost nobody funds it

Compare that against the US$202 an average influencer collaboration costs according to Collabstr's 2025 Influencer Marketing Report: that spend needs nearly fifty retention messages to break even, and retention does not expire the next day. The National Restaurant Association reports in 2025 that 75% of QSR brands running loyalty saw more traffic. My recommendation is uncomfortable but simple: shift a third of the reach budget toward contact capture and repeat business. One additional Yelp star lifts revenue between 5% and 9%, according to Michael Luca's study at Harvard Business School, and that elasticity turns review management into a finance-level decision rather than a community manager's chore. Exposure to public judgment is total: Restroworks 2024 puts at 92% the diners who read reviews before choosing, BrightLocal 2024 measures 71% reading specifically on Google, and TouchBistro 2025 records 55% of the youngest cohort reading opinions inside Instagram.

Chapter 4 — Reputation is a financial asset, not an inbox of incidents

Replying is not courtesy either: Toast 2024 finds 43% consider it very important that the brand answers. Run the counterfactual against your own books. If you bill 1.4 million a year and climb from 3.8 to 4.8 stars, Luca's range projects between US$70,000 and US$126,000 extra annually, and a twenty-minute daily reply protocol gets you there. Discovery, Reputation, Conversion and Repeat Business each need a separate budget line inside the 3% to 6% of sales that Toast documents as marketing spend for an established venue in 2025. The reason to split them is accounting before it is creative: when everything falls into one bag labelled social media, money migrates toward the most visible component, which is reach, and starves the one returning 67% higher spend per recurring ticket (Restroworks, 2025). On 1.4 million in sales, 4% is US$56,000 a year; split unevenly, I would put 40% into conversion and repeat business, 30% into discovery, 20% into reputation and 10% into content production.

Chapter 5 — Four components with their own budget: the Masterestaurant framework

I got this wrong for years, recommending a single budget for operational simplicity. The single bag always ends up funding impressions, because they are the metric that reaches the owner's desk first. Business size changes the entire recommendation, and treating the bands alike is the costliest allocation error I meet in advisory work. Below US$500,000 a year, 5% is US$25,000: no influencer is worth it there, and the return sits in a spotless Google listing and replies to reviews, with the 5% to 9% per-star elasticity Michael Luca measured at Harvard Business School. Between US$500,000 and a million the first capture system appears, because Tabular's 2025 figure of US$4.20 per SMS already justifies an owned contact base. Past the million mark, with US$42,000 to US$84,000 of budget under Toast's 2025 band of 3% to 6%, the conversation shifts to channel attribution.

Chapter 6 — Every revenue band plays a different game

And above five million, the operation needs a dedicated analyst, not an agency on retainer. Above five million in sales, the celebrity-chef venue or the large-format themed concept stops buying reach and starts managing reputational risk at scale, with a cost structure no independent-operator manual describes. The logic inverts because volume amplifies both sides: one star lost on that base, at the 5% to 9% elasticity from Michael Luca's work at Harvard Business School, means between US$250,000 and US$450,000 a year, a figure that funds an in-house team of three without debate. The US$202 average collaboration Collabstr reports in 2025 stops being a useful reference in this segment, where an activation with established talent gets negotiated with contract, exclusivity and a crisis clause. And yet the 92% who read reviews before deciding (Restroworks, 2024) applies just the same to the neighbourhood bistro and to the ten-venue group.

Chapter 7 — Ninety days, three KPIs and a board that understands the number

Install attribution before producing one more piece of content, because without it the next ninety days generate activity without evidence. Days 1 to 30: an owned destination with a redemption code, phone capture on every order, and a daily review-reply protocol, backed by the 43% who value a public answer (Toast, 2024). Days 31 to 60: confirmation and reactivation SMS, whose US$4.20-per-message performance Tabular measures in 2025. Days 61 to 90: only then, investment in discovery against the 58% post-TikTok visit rate MGH measured in 2024. The board-level indicators are three: attributable orders through owned channels, contribution margin per dollar invested, and ninety-day repeat rate. Stress the plan against input inflation of 5%, 12% and 20% before you approve it. FIRST fracture, the metric. Operators steer by audience while the register runs on attributable orders; the 58% who visited a restaurant after seeing it on TikTok (MGH Survey, 2024) never shows up on your dashboard without an owned destination collecting it.

Chapter 8 — Four fractures between the myth and the register

Spending, the second fracture, enters through the wrong door. Acquisition costs more and returns less than retention, since returning guests leave 67% more per order (Restroworks, 2025), yet the repeat-purchase budget is usually whatever survived the month. THIRD: reputation gets managed as an incident rather than a financial asset. Michael Luca's work at Harvard Business School is blunt here — one extra Yelp star lifts revenue between 5% and 9% — and still the 43% of diners who value a public response (Toast, 2024) get attention whenever somebody has a spare moment. The fourth one costs most: the loop never closes toward the owned channel. With 202 dollars of average collaboration spend (Collabstr, 2025) and a global delivery market projected at 1.51 trillion dollars for 2026 (Statista Market Forecast, 2026), handing the destination to an aggregator is a margin decision, not a convenience.