Your Next Ad Is a Voice Note, and It's Doing a Very Good Job of Sounding Human

There is a particular kind of British dread that arrives with a notification from a retailer you bought socks from in 2023. You know it will say something like “We miss you!” in the tone of a colleague who has been asked to clear their desk. So when a mid-sized homeware brand started sending 42-second voice notes to customers last month, the surprise wasn’t that it happened. It was that it worked.

The message wasn’t a robot reading a discount code. It was a woman with a soft Northern accent, slightly rushed, apologising that a delivery had been delayed. “Honestly, it’s the courier, not us,” she said, before offering to bump the order to next-day. It sounded like a voicemail from a friend who works in logistics. That, roughly, is the pitch now landing in UK CRM teams: stop writing, start speaking.

The 42-second sweet spot

Voice notes have been the default reply format on WhatsApp for years, but their migration into marketing is a 2026 story. Agencies are now producing scripted audio that mimics the rhythm of a real voice note: a small stumble, a half-laugh, a “sorry, one sec” that is entirely written in advance. Some are voiced by actors. Some are cloned from consenting staff. Some are synthetic, and the brief usually includes the phrase “make it sound like they’re in a car”.

The length matters. Anything under 20 seconds feels like a hostage message. Anything over a minute invites the skip. The 35-to-45-second window is where the format feels casual enough to belong in a chat thread, but structured enough to carry a call to action. Marks & Spencer could send you a polite note about a returns policy. A DTC mattress brand might send you a slightly chaotic update about a restock. Both are testing the same psychological trick: a human voice is harder to ignore than a paragraph.

Why intimacy is the new targeting

Email is a business letter. A push notification is a nudge. A voice note is, in theory, a person. That gap is where British brands are currently spending their energy. The format borrows the social contract of a mate checking in — and in the UK, where emotional directness is often wrapped in understatement, a spoken “just letting you know” can land more softly than text.

There’s an obvious risk. A brand that sends a voice note every Tuesday is no longer a mate. It’s a colleague who keeps leaving voicemails. Several CRM leads I spoke to for this piece were careful to frame the format as a one-off: an apology, a restock, a delivery update. The kind of thing that genuinely benefits from tone. “Sorry we missed you” is a sentence that works better spoken, especially when the alternative is another red badge on a phone screen.

The measurement problem

The industry hasn’t solved attribution. A voice note can be delivered, opened, played, paused, replayed and ignored, and none of those signals map neatly onto a click. Some platforms report “completion rate” — the percentage of listeners who reach the end — but that number tells you almost nothing about whether the message changed a decision. Did the customer reorder because the voice note was charming, or because the courier actually turned up?

Agencies are currently relying on soft metrics: reply rates to follow-up SMS, redemption on a spoken discount code, and the kind of qualitative feedback that doesn’t fit in a dashboard. One retailer reported a spike in customer service replies that began with “Hi, thanks for the voice note” — which is nice, but not a media plan. The measurement gap is the reason voice notes remain an experiment rather than a channel. Until someone builds a proper attribution model, the format will live in the same drawer as QR codes on coffee cups: promising, popular, and slightly unaccountable.

Disclosure, consent and the ASA

The regulatory picture is less murky than the analytics. The Advertising Standards Authority has been clear that marketing communications must be identifiable as such. A voice note from a brand is still an ad, even if it sounds like a friend. The disclosure doesn’t have to be a klaxon — a spoken “this is a message from [brand]” at the start is usually enough — but the context has to be obvious. If a synthetic voice is used, the ASA’s existing rules on misleadingness apply. If a real actor is voicing a message, the listener shouldn’t be left thinking it’s a named customer service agent.

Consent is the other hurdle. Sending a voice note by WhatsApp requires opt-in, and the ICO’s guidance on electronic marketing hasn’t changed just because the medium feels informal. Brands that treat a voice note as a cheeky workaround for email fatigue are likely to discover that the regulator is less charmed than the customer.

Sorry we missed you

The most interesting use case isn’t selling. It’s service. British retailers have spent years trying to make delivery updates feel less like a threat. A voice note that says “your parcel is with the neighbour, I’m sorry, I know it’s raining” is not just warmer than a push notification. It’s closer to how people actually apologise in this country: briefly, slightly awkwardly, and without over-explaining.

That might be the format’s real limit. Voice notes work when they replace a moment that already feels transactional. They stop working when they become a habit. The brands getting it right are treating the medium like a phone call from a shop assistant, not a new inventory of content. The ones getting it wrong are already planning a weekly series.

For now, the 42-second voice note is Britain’s most interesting ad unit precisely because it doesn’t feel like one. The moment it does, it’ll go the way of the newsletter pop-up: muted, deleted, and quietly resented.