
"Best time to cold call" studies disagree because they do not measure the same thing. Some count pickups, some count conversations, some count meetings, on different countries, personas and number types. Our own benchmark of 1,000,000 French B2B cold calls shows the problem inside a single dataset: connect rate is flat from Monday to Friday, while meeting rate moves 1.7x between the best and worst day. Both statements are true. They answer different questions.
Take the day of the week. If you rank days by connect rate, the share of dials where someone picks up, our data gives no winner. The rate is stable through the week.
If you rank days by meeting rate, the share of conversations that end with an accepted meeting, there is a clear slope. Meetings are booked mostly early in the week and booking collapses on Friday, from 4.26% on the best day to 2.55% on the worst.
A study built on the first metric would conclude that the day does not matter. A study built on the second would name a best day. Neither would be wrong. Published studies name Tuesday, Wednesday or Thursday depending on the year and the author, and this is a large part of the reason.
There are at least four possible numerators: a pickup, a conversation of a minimum length, a positive outcome, a meeting booked. In our data a call that ends in a meeting lasts 254 seconds on average and a call that does not lasts 81 seconds. A study that only counts conversations above a duration threshold is already selecting for good calls, and its "best hour" drifts toward the hour where prospects have time to talk.
Connect rate per dial, per valid number or per contact reached are three different rates. Wrong numbers and switchboards do not behave like direct lines. Our hourly rates are computed on dials, and the meeting rates on productive conversations, with calls without a pickup excluded. Before comparing two studies, check what sits under the fraction.
This is the big one. In the benchmark, sales and leadership personas connect above 15% while HR and IT connect under 9%. Founders pick up more around 10am, marketing around 4pm, HR around 1pm.
A dataset dominated by founders will crown the morning. A dataset dominated by marketing will crown the late afternoon. The "best hour" of a study is partly a description of who was in its list.
Our benchmark covers B2B calls placed in France. Lunch habits, office hours and the share of mobile numbers differ between countries, and so does the way people treat an unknown number. A window measured in one market is a hypothesis in another. We do not have a country-by-country cut, and we will not pretend the French windows apply everywhere.
An hourly table has few cells. An hour by day by persona table has hundreds, and most of them are thin. We exclude every segment under 200 calls from connect and time-slot rates for that reason. A study that reports a winner at "Wednesday, 6pm" on a few dozen dials is reporting noise.
Read enough of them and the disagreement shrinks to details. The broad pattern is consistent with what we see:
The fights are about whether the peak is 10am or 11am, Tuesday or Wednesday. Those are exactly the differences a change of persona mix or metric would produce.
We apply the same list to our own numbers. The benchmark states its limits: the gaps may be correlations without demonstrated causality, and they should be read as practices to test in your own context.
A benchmark gives you a starting hypothesis. Your own dial log gives you the answer for your list, your personas and your country.
Two rates, two views: connect rate per dial by hour, and meeting rate per conversation by day. If you have at least 200 dials per cell, you have a result. If not, start from the benchmark windows and collect. The method, with sample sizes and the traps to avoid, is in how to test your best cold calling time in two weeks.
And keep the size of the prize in mind. Timing is worth 1.6x to 1.7x. Persona and opener are worth about 4x each, as we show in does cold call timing really matter.
Sonik is an AI dialer for B2B outbound teams. The benchmark exists because every call placed through the dialer carries a timestamp, a persona, a duration and an outcome.

We publish connect rate by hour and by persona, and meeting rate by day and by function, with the 200-call threshold, because a blended "best time" hides more than it shows. Sessions in the dialer are built per list, so a team can act on its own windows once it has them.
Timing studies disagree because they measure different outcomes on different people in different countries. Inside our own data, the day does not change who picks up and changes who books by 1.7x. Before you move your calendar for a headline, ask what was counted. This week, pull two reports from your dialer, connect rate by hour and meeting rate by day, and compare them with what 1,000,000 B2B calls reveal about the best time to cold call and our day-of-week data.