Dead Air Isn't Neutral: Rethinking the Call Queue at a Small UK Firm

There's a moment every caller knows. The greeting lands, someone says your call matters, and then — nothing useful. Music, a repeated apology, a promise that someone will be "with you shortly". Ten minutes later the caller is on a competitor's website. Nobody in your business decided that should happen. The queue just did what it was set to do when the engineer handed over the logins.

That's the thing about call queues: the default is still a design decision, even when you never made it. And at small-firm scale, where one abandoned call can be a five-figure job, it's worth an hour of scrutiny.

Start with what the caller actually hears, in sequence

Before touching a setting, ring your own main number from a mobile at three different times of day and record it. Listen back with a notepad. You're looking for the full sequence: how long it rings before anything answers, what the first words are, whether you're told your position, whether the music changes or loops awkwardly, and what happens if you wait it out — voicemail, a ring-out, a dead tone.

Most SMEs discover two things. First, that the announcement is longer than anyone realised and says almost nothing. Second, that the voicemail at the end is a cheerful message recorded in 2019 by someone who has since left.

Cut the announcement down to information

"Your call is important to us" is filler. Callers have heard it so often it now reads as a warning sign. What they want in the first fifteen seconds is orientation: am I in the right place, how long is this likely to take, and what are my alternatives?

A useful announcement does three jobs. It confirms the company name so the caller knows they haven't been misdirected. It gives a realistic wait estimate — more on that below. And, if the caller is ringing near closing time, it says so and offers an email address or a callback instead of letting them wait twenty minutes for a voicemail.

Two sentences is usually enough. Anything longer and people start talking over it to whoever is in the room with them.

Estimated wait times should be honest, not flattering

If your system offers an estimated wait announcement, resist the temptation to configure it optimistically. A promise of "two minutes" that turns into eleven is worse than an honest "around ten minutes" that turns out to be eight. Under-promising is the oldest trick in service design and it still works because it's the only one that reliably exceeds expectations.

If your platform can't calculate a live estimate, don't invent one. A recorded "we're busier than usual this morning" is more credible than a number that's plainly wrong.

Hold music is a licensing question as much as a taste one

Here's the bit people forget. Playing recorded music to callers in the UK is a public performance, and it needs a PPL PRS licence — even in a queue, even for a small firm, even if it's just a Spotify track streamed from a desk phone (which, incidentally, is a technical mess anyway). The cheapest route for most SMEs is a royalty-free library built for telephone systems: short loops, no lyrics, no sudden key changes.

Silence is the worst option. A caller with no audio assumes the line has dropped and hangs up. If licensing is a hassle, a simple spoken message repeating on a loop — opening hours, website, callback instructions — is licence-free and far more useful than music anyway.

The callback option only works if you honour it

Callback features are wonderful on paper and corrosive when abused. The rule is simple: if you offer a callback, ring back within the window you promised. Not the same afternoon, not "when we get to it". If you can't staff that promise, switch the feature off — an unreturned callback is a worse experience than a long wait, because the caller has now given you their number and been let down twice.

Offer it early in the queue, not after eight minutes. And make sure the message says roughly when you'll ring — "within the next hour" or "before we close" — so the caller can decide whether to hold or go about their day.

Abandoned calls are a staffing signal, not a telecoms fault

Every decent phone platform will show you abandoned-call data: how many callers hung up before being answered, and at what point. Ignore the headline number and look at the shape of it. Do most abandons happen in the first thirty seconds? That's an announcement problem — people are hanging up before they've been given a reason to stay. Do they cluster at 9:15am and 2pm? That's a rota problem. Are the same numbers calling back three times? That's an unresolved-issue problem, and no queue setting will fix it.

Small changes move this data faster than people expect. Shifting one person's lunch break by half an hour to cover the post-9am spike is free. Adding a second pair of hands to the phones between ten and eleven is often cheaper than the job you lose by missing the call.

A one-week test your team can actually run

Monday: ring your own line three times and record it. Tuesday: time the announcement and rewrite it to two sentences. Wednesday: check your hold music licence or swap it for a spoken loop. Thursday: turn callback on, test it internally, then ring in as a customer and see how long it takes to be called back. Friday: pull the abandon data and pick one thing to change. The following Monday, listen again and compare.

None of that requires a new contract, a hardware refresh, or a meeting longer than twenty minutes. It just requires someone to sit on the other side of the line and decide, deliberately, what happens after the greeting.

Because the hold button isn't a strategy. It's a pause — and the caller gets to decide what fills it.