The 98% WhatsApp open rate has no study behind it

Four numbers run the follow-up business. WhatsApp opens at 98%. 78% of customers buy from the first business that answers. Closing takes eight touches. 80% of sales need five follow-ups. Trace any of them to a study and you land in a loop of blogs citing blogs. Meta publishes no open or response rates for business messaging at all, so the 98% has nothing underneath it. What Meta does publish is preference: 73.3% of consumers say they prefer messaging when they deal with a business (Meta WhatsApp 2026 Consumer Report, 2025, from 11,056 consumers across 22 markets).

The number that does survive an audit is better than the myths anyway. Across more than 246 million campaign sends and over 27,000 brands, broadcast messages converted at 0.12% while messages triggered by customer behavior converted at 0.78% — roughly 6.5 times — earning $0.75 per message against $0.15 (Omnisend, 2025). Those are US figures for SMS. The mechanism transfers: a message that arrives because a person did something beats one that arrives because it is Tuesday.

A five-minute rule is a competitive position, not a best practice

The strongest evidence on response speed is old and still unbeaten. An audit of 2,241 US companies and 1.25 million sales leads found that 37% responded to a web inquiry within an hour, 24% took longer than a day, and 23% never responded at all. Contacting a lead inside the first hour made a firm nearly seven times more likely to qualify it than waiting one more hour, and more than sixty times more likely than waiting a full day (Harvard Business Review, 2011).

Read the 23% again. Not answered slowly. Never answered. The companion study everyone quotes for the five-minute ratios is 2007-era data co-produced by a vendor selling response software, so treat those exact multipliers as illustration rather than measurement (MIT/InsideSales.com, 2007 data). Two cleaner substitutes: XANT reported eight times the conversion for replies inside five minutes versus six minutes or more, and Drift found only 7% of business software companies answered within five minutes (Expertise.ai, 2026). Drift measured software firms, not local businesses. The shape holds: almost nobody keeps a five-minute rule, which is what makes keeping one worth something.

So write one sentence and put it on the wall. Every new inquiry gets a first response within five minutes during opening hours, and within fifteen minutes of opening otherwise. Then name the person on the clock for each block of the day. A rule without a name is a wish.

Disclose the AI, then keep it on logistics

Somebody has to answer at 9pm. The best evidence on letting software do it is a randomized field experiment: 6,255 customers, six conditions, at a lender making outbound loan-renewal calls. The undisclosed chatbot produced a 23.7% purchase rate — statistically indistinguishable from proficient human agents at 25.1%, and about four times better than inexperienced workers at 4.9%. Disclosing the bot as AI before the conversation collapsed that rate to 4.8%, a drop of 79.7% (Marketing Science, 2019).

The honest conclusion is not to hide the bot. It is that AI can match a good human on outcome, that disclosure carries a measured cost, and that the cost is smaller than being caught. Disclosing after the conversation reduced the penalty substantially, customers with prior experience of AI apps were far less sensitive, and the experiment ran in 2019, before large language models were ordinary. Disclose in the first message, keep the assistant on logistics — confirm, two qualifying questions, offer times, hand off — and never let it discuss price, diagnosis or legal position.

Reminders work, and the phone call has the firmer evidence

A booked appointment that does not arrive costs more than a lost inquiry, because you held the slot. A systematic review and meta-analysis covering 12 studies from nine countries found that reminders of all types improved attendance, with a pooled risk ratio of 1.11 across 8,236 participants — about an 11% relative improvement. Broken out, telephone reminders returned RR 1.11 at P=0.002, statistically significant, while SMS reminders returned RR 1.14 (95% CI 0.99–1.31) at P=0.07 with very high heterogeneity — a trend, not a result (Journal of Hospital Management and Health Policy, 2024).

That inverts the usual advice. "SMS cuts no-shows by 38%" is not a number the pooled research supports. Call for first-time and high-value bookings, message the routine repeats, use two channels, and make confirming a single tap.

Picture a two-van plumbing company

Picture a two-van plumbing company in a mid-sized city. Calls arrive while both technicians are under a sink, the phone rings out, the caller leaves no voicemail and does not call back. The famous "62% of calls to small businesses go unanswered" traces to a vendor in 2016 (Numa, 411 Locals 2016) — dated vendor research that illustrates a mechanism rather than measuring your market. The owner does not need the statistic. He needs the missed call to produce a message.

Five stages, built in this order. Capture: phone, WhatsApp, web form and Instagram messages land in one inbox with one named owner, and a missed call fires a text back inside 60 seconds — "We're on a job. Reply with your postcode and the problem and we'll book you today." Respond: the five-minute rule during hours, a disclosed assistant outside them. Follow up: the quote goes out itemized at 24 hours with the earliest slot already held, then at 72 hours something useful — a photographed comparable job. Show up: confirmation, a 48-hour reminder, a call for first-time bookings. Keep: 24 hours after the job, the review request and the annual maintenance plan in one message.

Those last two steps carry the book's strongest evidence. Purchase likelihood for a product with five reviews is 270% greater than for one with none, with a larger lift on higher-priced items (380%) than lower-priced ones (190%) — reviews matter more as the ticket rises, exactly the service-business case. Likelihood peaks between 4.0 and 4.7 stars and declines toward 5.0, because a perfect score reads as fake (Medill Spiegel Research Center, 2017). And across 5,181 referred and 4,633 non-referred bank customers, the referred were worth at least 16% more over six years and roughly 18% less likely to defect at any given time, with a EUR 25 reward generating a EUR 40 lifetime-value difference (Journal of Marketing, 2011). The referral ask is not the last item on a checklist; it is the highest-yield message that company sends.

What's in the playbook

• The five-stage method — Capture, Respond, Follow up, Show up, Keep — in build order, each with its own check
• 14 industry playbooks: hotels, restaurants, health, beauty, general SMB, law, real estate, retail and gyms, automotive, education, events, financial services, home services, wine
• A 90-day plan in twelve rows, each with what you build and how you know it worked
• Scripts to paste and edit: missed-call text-back, the three-step follow-up, the reminder set, the review-to-referral ask
• Nine prompts for Claude, ChatGPT or Gemini, plus the AI first-responder rules and hand-off triggers
• The weekly funnel sheet: six numbers defined, with where each comes from
• Compliance on one page: TCPA revocation inside 10 business days since April 2025, WhatsApp's opt-in and commerce rules, Chile's Ley 21.719 from 1 December 2026
• 28 numbered sources with every figure linked, the unsourced stats refused by name, and the English and Spanish PDF in every purchase

None of this needs a developer and most of it needs a Saturday. Start with the free audit: write down your last ten inquiries, the channel each arrived on and the time you replied. If you cannot reconstruct that list, that is your first finding. From Click to Client is on the playbooks page at netwebmedia.com/playbooks, with the Spanish edition in the same purchase.

Questions owners ask

Is the 98% WhatsApp open rate real?

No. Meta publishes no open or response rates for business messaging, and no study behind the figure exists. What Meta does publish is preference: 73.3% of consumers say they prefer messaging when dealing with a business (Meta WhatsApp 2026 Consumer Report, 2025). Ask any vendor quoting 98% to show you the study.

Do automated messages really beat broadcasts?

In US SMS data covering over 246 million campaign sends and 27,000-plus brands, broadcasts converted at 0.12% and behavior-triggered automations at 0.78% — about 6.5 times — earning $0.75 per message against $0.15 (Omnisend, 2025). Trigger on what the customer did.

Should I remind people by text or by phone?

Both, but call the ones that matter. A meta-analysis found telephone reminders significant at RR 1.11, P=0.002, while SMS returned RR 1.14 at P=0.07 with very high heterogeneity — a trend, not a result (Journal of Hospital Management and Health Policy, 2024). Call first-time and high-value bookings.

← Back to all articles