"Press 1 for sales" — done properly.

Menu design: where auto attendants are won and lost

The difference between an auto attendant that helps and one that infuriates is almost entirely in the design, and the rules are short. Three options is the ceiling for most businesses — by option four, callers have forgotten option one. Put the most-chosen destination first, because a majority of callers press the first plausible thing they hear. Say the department before the number ("for sales, press one") — reversed, callers must hold the number in memory while parsing what it was for. And keep the greeting under ten seconds. Every second of preamble is multiplied across every call you'll ever receive; "thank you for calling" is polite, a paragraph about your company values is a hold-music review waiting to happen.

The details that separate professionals from amateurs: always give the zero-out. Callers who don't fit your menu need "or press zero to speak to someone", and it should genuinely ring a human — a zero option that loops back to the menu is how businesses end up in one-star reviews. Behave differently out of hours: an attendant offering "press one for sales" at 11pm, followed by ringing an empty desk, tells callers nobody thought about them. Time conditions should hand evenings to voicemail, an AI receptionist, or an honest closed message with your hours. And if callers might ever need you urgently — tenants, patients, contract clients — emergency routing behind a menu option beats an emergency mobile number taped to someone's monitor.

Now the honest part: some businesses shouldn't use one at all. If you're five people and every caller wants "whoever picks up", a menu is theatre — a hunt group ringing everyone at once serves callers faster and sounds smaller in a good way. The attendant earns its place when callers genuinely split into distinct destinations — sales versus support, bookings versus enquiries, English versus another language — or when you need calls sorted before they hit busy people. If a provider proposes an IVR tree to a four-person firm, they're selling features, not solving problems. On our quotes, the recommendation follows the call pattern, and we've talked plenty of customers out of menus they didn't need.