WhatsApp marketing and broadcast lists: the numbers that decide it for you

The channel opens four times more than email, and that changes the editorial decision

A WhatsApp broadcast list gets read between 75% and 98% of the time, while email open rates in the restaurant industry sit at 21,3% according to Mailchimp's sector benchmarks, and that four-to-one gap decides where the week's effort goes. Meta reports more than 2 billion monthly active users, with penetration above 90% of smartphone users in Brazil, Mexico, Colombia and Spain. A neighborhood restaurant in Medellín closes the month with 1.180 contacts saved on the business phone and zero revenue attributable to them: the tool isn't missing, the decision about what that list is for is. If the channel opens four times more and still doesn't bill, the problem stopped being technical. It's what gets said, to whom and how often. Bringing back an existing customer costs five to seven times less than acquiring a new one, a calculation Harvard Business Review has circulated for years, which makes a permission-based list the cheapest marketing asset a restaurant owns.

Repeat purchase, not cold acquisition: the number that rules is the cost of bringing someone back

With sector net margins between 3% and 9% according to Statista, every ad dollar you don't spend re-buying the same audience stays whole at the bottom. Run the math on that list of 800 contacts: if 20% buys one extra time per quarter at a 45 USD ticket, that's 7.200 USD in additional quarterly revenue without purchasing a single impression. The honest comparison against paid reach isn't reach versus reach. It's attributed cash against attributed cash, and there the list wins almost every time. The WhatsApp Business broadcast list, the free app, only reaches people who have your number saved in their contacts, caps at 256 recipients per list, and measures nothing beyond blue ticks. The WhatsApp Business API charges per business-initiated conversation, requires approved templates, and in exchange delivers real segmentation and attribution down to the sale.

Broadcast lists and the API are not the same thing, and confusing them produces absurd budgets

An owner with 1.200 contacts and one monthly message lives perfectly well on the free app split into five lists; a six-location group that wants to trigger by purchase behavior needs the API, and its cost per conversation gets compared against the 4,20 USD an average reservation confirmation SMS generates according to Tabular. Mixing both tools into one plan produces quotes nobody sustains past the third month. A phone number captured during a reservation does NOT authorize advertising, and that distinction decides the useful life of the entire channel. With explicit, dated opt-in, monthly unsubscribes stay below 2%; without it, between 8% and 14% of the base blocks or mutes the number within the first six weeks, and a number blocked by hundreds of people drags delivery problems across the whole account, including the reservation confirmations that actually matter. Turn it around: if you lost 12% of 1.180 contacts tomorrow, you wouldn't lose 142 people, you'd lose the ability to reach the remaining 88% at the same delivery rate.

Dated permission separates a channel that lasts years from one that dies in six weeks

That's why the consent record, with date and origin, belongs in the same table as the phone number. Not in a separate note. The correct unit of measure is identified revenue per active contact, not how many received the message, and switching metrics reorders everything else. You calculate it by dividing revenue identified through the campaign code or link by contacts who opened or replied in the last 90 days. With 800 active contacts and 7.200 USD attributed per quarter, the indicator lands at 9 USD per contact per quarter, and that number now compares against any paid channel. Diego F. Parra insists at Masterestaurant on anchoring the list to the location's average ticket from the very first send, because a list without a tracking code isn't a marketing channel, it's an expensive phone book. 97% of SMS messages get read within 15 minutes according to Tabular, and WhatsApp's immediacy is comparable: if the message lands at eleven on a Friday morning, the buying window already closed.

How to read these numbers in YOUR operation: small, mid-size and group?

The three scenarios share neither tool nor cadence, and forcing one plan on all of them is the costliest mistake on this list. Small location, under 500 contacts:

free app, four lists segmented by visit frequency, two sends per month, target of 4 to 6 USD in attributed revenue per active contact per quarter. Mid-size, 500 to 3.000 contacts across two or three sites: the free app still works, but with lists per site and a tracking code per campaign, weekly cadence in season, target of 8 to 12 USD per contact. Group of four locations or more, above 3.000 contacts: API, approved templates, behavior-triggered sends and attribution down to the POS. If your base is 900 contacts and someone is quoting you an API integration, ask first for three months of the free version with measurement: 75% of QSR brands with a loyalty program reported more traffic in 2025 according to the National Restaurant Association, and none of them started with software.

Where these benchmarks come from and what you can NOT ask of them?

The figures in this analysis come from three different origins and it's worth saying so plainly. Channel-behavior numbers come from public reports with open methodology:

Mailchimp by sector for the 21,3% email open rate, Tabular for the 4,20 USD per reservation SMS and the 97% read rate within 15 minutes, Statista for the 3% to 9% sector net margin. WhatsApp adoption figures are Meta's own statements, not independent audits. And the 75% to 98% open range consists of aggregated industry observations, not a controlled trial: it varies by country, by send hour and by how old the base is. Use them as an order of magnitude to size a decision, never as a forecast of your next campaign. Your own three-month data, with a tracking code, is worth more than any borrowed benchmark. A well-run list sends less and bills more, and that paradox resolves once you look at which message earns a reply.

What pays isn't the promotion, it's the message the customer was already waiting for?

Telling 210 people who ordered last year's seasonal dish that it's back converts above 15%; a generic discount blasted to the entire base converts below 2% and costs you unsubscribe points.

Social influence confirms it from another angle: 58% visited a restaurant after seeing it on TikTok, up from 38% in 2022, according to MGH's 2024 survey, and brands with the strongest social strategy saw 14,1% higher revenue according to Deloitte Digital. The same logic runs here, with prior permission. This week, before scheduling any send, split your base into four lists by last dish ordered and send one single thing to the smallest list. The first gap is PERMISSION, and everything else follows from it. A number captured on a reservation does not authorize advertising; a dated opt-in does, and that distinction separates a channel that pays for years from one that dies in two months.

Where the real gap opens?

With explicit consent, monthly opt-out stays under 2%; without it, 8 to 14% of the base blocks or mutes within six weeks, and a number blocked by hundreds of people starts having delivery problems across the entire account.

The second is the UNIT OF MEASURE. Traditional practice reports reach — how many received it — because that is all the free app shows. We measure attributed sales per active contact, calculated by dividing revenue identified through the send's code or link by the number of contacts who engaged in the last ninety days. That figure stings at first, since it almost always lands lower than the owner assumed, and that is precisely why it works. Third comes FORMAT. A restaurant that only knows how to send discounts trains its base to wait for discounts, so full-price sales collapse between sends. Rotating formats — a story about the beef supplier, a dated Thursday reservation, the new menu with two dishes, and yes, one offer with an expiry — holds open rates without burning margin.

Where the real gap opens — in practice?

I got this wrong for years: I treated brand content in a direct channel as filler, until we saw segments fed only promotions drop nearly thirty points of open rate in six months.

Fourth is DESTINATION. Sending a guest to the delivery marketplace from your own WhatsApp means paying 18 to 30 points of commission on an order you had already won. The correct link goes to the owned channel, with an updated digital menu, and the physical menu waiting if the guest chooses the dining room, because the QR settles price and availability while the menu in hand governs service pace and suggestive selling. The fifth barely gets discussed: OWNERSHIP of the asset. If the base lives on the community manager's personal phone, you don't have a channel, you have a dependency. Diego F. Parra raises this in every Masterestaurant diagnostic: the day that person leaves, the list leaves, and rebuilding a thousand consented contacts costs months of operation.