“85% never call back”: autopsy of a phantom statistic

Take any statistic circulating about missed calls and trace it back. You’ll find a blog post citing another blog post, citing a vendor’s page, citing… nothing more. The chain ends in thin air. That’s the hallmark of an orphan statistic: a number that looks credible because it’s repeated everywhere, but that never had a verifiable starting point.

A primary source is the original study: a named institute, a described method (who was surveyed, how many people, when, how), a clear scope. If you can’t answer “who measured this, on whom, and when?”, you’re not holding data. You’re holding a well-dressed rumour.

For the famous “85% never call back”, this test fails. To our knowledge, there is no public, dated, methodologically described study establishing this figure for the French market. Saying so plainly isn’t an admission of weakness: it’s exactly what you’d expect from someone who respects their reader.

A viral statistic combines three ingredients. It’s round and sharp (85, not 83.4). It confirms an intuition everyone already has (“yes, it’s true, I don’t call back either”). And it serves a sales narrative: the bigger the loss seems, the more indispensable the solution appears.

Each time a figure is copied, its cautionary quotation marks fade a little more. An “around”, a “according to some sources”, a “in certain sectors” evaporate with each retelling. By the end of the chain, the approximation has become an etched-in-stone truth. It’s a game of telephone: the final message is crisp and false.

• Be wary of a figure without a year: behaviour from 2014 isn’t behaviour from 2024.
• Be wary of a percentage without a scope: “people” means nothing; which people, which sector, which country?
• Be wary of a figure that suits the person citing it a little too well: that’s rarely a coincidence.

There is solid data on the business phone, and it doesn’t need inflating to be telling. According to the ARCEP market observatory, voice traffic from French businesses represented 31.2 billion minutes in 2024. The channel is slowly declining (−3% over the year, and a steady drop for more than ten years), but the order of magnitude remains massive: professional voice in France is still measured in tens of billions of minutes per year.

ARCEP also notes that 68% of business lines are now on VoIP, versus 32% still on legacy PSTN. In other words, the infrastructure is modernising while the channel slowly ages. A sober, defensible conclusion: the phone remains a front-line point of contact, and every unhandled call has real value there. You don’t need an “85%” to justify that a missed call counts.

A national statistic, even a true one, doesn’t describe your business. The right instinct isn’t to hunt for the perfect figure to copy, it’s to measure your own. It’s free, it’s fast, and it’s the only number you can defend without blushing.

The method comes down to four questions, tracked over two or three weeks to smooth out fluctuations.

• How many calls actually come in? Your carrier or switchboard gives you inbound volume, answered and unanswered.
• When are the missed ones? Clustered at lunchtime, in the evening, on weekends? The timing tells you what to fix.
• What is a call worth to you? Take your average ticket or the value of a won customer: that’s your benchmark, not another sector’s.
• How many callbacks succeed? Call back your missed callers for two weeks and count those that close. You get your real rate, your own.

Refusing to cite a figure you can’t prove is also a commercial stance. A provider who tells you “here’s a public data point, here’s its limit, and here’s how to measure your own case” treats you as an adult. The one waving an anonymous percentage is hoping you won’t check.

This standard applies well beyond the phone. A useful frame to keep in mind: according to the 2024 Observatoire des Services Clients (BVA Xsight), 89% of French people think AI will be used in customer service within less than ten years, but only one in three genuinely wants it. That figure is attributed, dated, named. So it can be cited, and it’s a reminder: buy-in can’t be decreed with spectacular stats, it’s earned through transparency.

Tinos is a voice agent that answers in your place, 24/7, qualifies the call, books the appointment and notifies you by SMS or email, then logs it in your tool. But it shouldn’t be an “85%” that makes up your mind: it should be your own measurement. Count your missed calls, their timing and the value of a customer to you, then compare that real cost to a monthly subscription. If the maths doesn’t add up for your business, don’t take our word for it — and that’s exactly the kind of honesty we stand for.

• ARCEP — Electronic communications markets observatory in France 2024 (final results)
• BVA Xsight — 2024 Customer Service Observatory