Best time to make cold calls: Hours, days and cadence

People ask which hour is best as though there is a magic window. There is a pattern, but it is much weaker than the thing nobody wants to hear.

Times that tend to work

Two windows come up consistently: early morning before the day fills with meetings, and late afternoon once urgent work is done. Midday is usually worse. These are tendencies, not rules, and your market may contradict them.

Days to avoid

Monday morning is chaos and Friday afternoon has already left. The middle of the week is generally steadier. Again, treat this as a starting assumption to test rather than a law.

What cold calling connect rates look like

Set expectations before you judge any time slot. On cold B2B dials, only a small share of attempts reaches a live person: direct mobile numbers connect far more often than switchboard numbers, so measure the two separately before you judge either. Most dials end in silence whatever the hour, and a good window lifts the rate by a few points rather than transforming it.

That has a practical consequence: twenty dials tell you nothing about a time slot. The sample is too small and one busy afternoon skews it completely. Give each window a few hundred dials across two or three weeks before ranking it against the others, and compare connect rate and conversation rate separately, because a window that connects well can still produce rushed, distracted conversations.

Repetition matters far more

This is the part that actually moves the number. One attempt reaches almost nobody: people are in meetings, travelling, or away.

The same contact tried several times, at different hours and on different days, is reached far more often than one tried once. "No answer" is not an outcome, it is the first of several attempts.

Track the attempts

Repetition only works if you record what happened and when, otherwise you call at the same hour repeatedly and get the same silence. That belongs in the CRM, not in someone's memory. Set a clear limit for when a contact is considered unreachable.

A weekly calling cadence to test

If you are starting from zero, borrow this structure and adjust once your own numbers come in:

1. Block two windows per day. One before 9:30 and one after 16:00, in the prospect's local time. Protect them in the calendar the way you protect meetings, because prospecting time that is not blocked gets eaten.
2. Rotate windows per contact. First attempt in the morning, second in the late afternoon, third at midday a few days later. Availability patterns differ per person, and rotation is how you find them.
3. Space attempts across the week. Two attempts on the same day irritates. Two attempts four days apart does not.
4. Cap the sequence. Five to eight attempts over three to four weeks is a common ceiling. After that, mark the contact unreachable and move them to an email-only track rather than deleting them.
5. Leave one voicemail, not five. A single short voicemail early in the sequence gives your later calls context. Repeated voicemails burn goodwill fast.

Common timing mistakes

• Calling when it suits you, not them. Reps dial when their own calendar happens to be free, which is usually midday, the weakest window. Fix: block prospect-friendly windows first and fit admin around them.
• Treating one no-answer as a dead contact. Availability is noise on a single attempt. Fix: set a minimum attempt count before anyone is allowed to disqualify a number.
• Abandoning a window after one bad day. A public holiday abroad or a local event can empty phones for a day. Fix: judge windows on weeks of data, not sessions.
• Ignoring the gatekeeper's schedule. Reception and assistants work fixed hours. Calls before 8:00 or after 17:00 sometimes ring straight through to the decision maker, which changes the conversation entirely. More on that in getting past gatekeepers.
• Batching every dial into one marathon. Energy drops, tone flattens, results follow. Two shorter blocks beat one long one.

Export markets

Time zone and local rhythm decide everything here. A call landing at 7am or during a local lunch break reads as disrespect no matter how relevant the offer. Check public holidays and normal working hours, which vary more across Europe than people assume, and check what the local rules allow before the first dial, because some countries restrict unsolicited business calls far more tightly than others.

How this plays out across Europe

Working hours vary more across Europe than most calling playbooks admit. German-speaking markets start early: reaching a Geschäftsführer at 7:45 is normal, while a Friday afternoon call into Germany is pointless because many offices empty out by 14:00. France protects the lunch break, so 12:00 to 14:00 is dead time. The Nordics wind down early in the afternoon. The Baltics run closer to a 9:00 to 18:00 rhythm, with Lithuania one hour ahead of Berlin.

Selling into several markets at once turns the calling calendar into a planning exercise. A Vilnius-based team calling into DACH loses the local early morning but keeps a clean run from mid-morning to late afternoon, which is one reason we sequence markets by time zone when we run outbound into DACH alongside Baltic and Nordic campaigns. If your prospecting spans more than two zones, it pays to structure the whole week around it, as covered in outbound across time zones.

Warm the call first

A call to someone who already saw your email or LinkedIn message goes differently: they recognise the name and the conversation starts warmer. Combining channels affects outcomes more than any hour of the day, which is the point of a multichannel sequence.

If you are waiting for the perfect slot

The perfect-hour question is often a polite form of procrastination: research feels safer than dialling. The reality is that a mediocre window worked consistently beats a perfect window worked occasionally, because volume and repetition compound while timing does not. Once the phone is answered, the clock stops mattering and the first ten seconds take over. Pick two windows, commit to them for a month, log every attempt, and let your own data overrule everything written here.