The best hours to call, measured on four thousand calls

We measured roughly four thousand outbound calls across our own teams and looked at one thing: what an hour of dialling actually returns depending on when it happens. The morning produced 1.85 times the result of the afternoon. Not slightly better. Nearly double. No feature in any phone system moves a number like that, which is why this article is about the schedule and not about the software.

Why the ratio is so large

Two effects compound. Pick-up rates are higher earlier, because calendars fill as the day goes on and people are in back-to-back meetings by the afternoon. And the conversations that do happen early are longer, because the person you reached has not yet spent their patience on six other things. A higher pick-up rate multiplied by a better conversation is how you get to 1.85.

It also means the standard reflex, which is to add dialling hours at the end of the day because that is where the empty calendar space is, buys you the worst hours on the clock. The team ends up working longer for a flat result and concludes that calling does not work.

What to do with the number

Protect the morning. Not as a preference, as a rule: no internal meetings, no pipeline reviews, no training before midday. Those things feel productive and they are being paid for out of the most valuable hours the team has. Move them to the afternoon, into the slot where a call is worth roughly half as much.

Then use the afternoon for the work that supports the morning: research, follow-up emails, LinkedIn touches, CRM cleanup. The afternoon is not wasted, it is the wrong hour to spend a fresh list on.

Measure your own version of it

Our 1.85 is ours. It comes from our markets, our segments and our time zone, and yours will differ. What should not differ is that you have the number at all. Most sales teams cannot tell you what an hour of dialling returns at 9am against 4pm, which means their schedule is set by habit rather than by evidence.

The measurement is not complicated: every call already carries a timestamp and an outcome. Group by hour, count conversations per hundred dials and outcomes per conversation, and look at the shape. A few hundred calls is already enough to see it. A phone system that syncs outcomes to the CRM automatically makes this a query rather than a project, which is the practical reason to care about that feature.

What the phone system contributes

Three things that survive this analysis. A local number per rep, because a number that looks local is answered more often and can be called back. Recording and transcription, because the conversation is worth more than the outcome field someone types afterwards. And automatic write-back to the CRM, because the alternative is a rep choosing between making the next call and documenting the last one, and they will always choose the next call.

At 32 dollars per user per month billed yearly for the tier that includes CRM sync, AI summaries and call tags, the arithmetic is straightforward against a rep salary. The tier above, at 100 dollars, buys AI agents, and that is a different decision with a different justification.

The part nobody automates

The follow-up on a call nobody answered. Most pipelines leak there: a dial that does not connect produces nothing, and nobody goes back to it. A voicemail plus a message on another channel, sent the same day, is the cheapest recovery available, and it is a rule you set once rather than a decision a rep makes forty times a day.

What this does not fix

A bad list. Calling the wrong people earlier in the day just gets you rejected sooner. The hour multiplies whatever the list and the pitch already produce, which is why this measurement belongs at the end of the optimisation list and not at the start, even though it is the easiest one to act on.