AI Calling Campaigns

Best time to make calls: what the data says

The best time to make calls depends on who you're calling and why. What the studies actually found, the legal windows, and how to test your own best hours.

Vadim Kouznetsov·Nov 15, 2026·10 min read
A wall clock and a desk telephone on a windowsill in late afternoon light
Key takeaway

There is no single best time to make calls. For consumers on mobiles, early evening on a weekday is a reasonable default, and early morning is usually the weakest slot, but published studies disagree on the details and the differences are a few percentage points. Three things matter more: calling within the legal window in the person's own local time, calling soon after whatever made the call relevant, and spacing retries across different parts of the day. Then test your own hours, with enough calls to trust the result.

Search for the best time to call and you'll find confident answers that contradict each other: Monday at 8am, Wednesday at 4pm, Friday evening. They're not all wrong. They measured different people, for different reasons, in different years. The useful question isn't "what time is best?" but "best for whom, and for what kind of call?"

What the studies actually found

Here is what four of the more transparent sources report, with what each one measured.

SourceWhat was measuredBestWorst
CallHub, 2.2M+ calls, US and Canada, Aug–Dec 2020Answer rate on volunteer campaign calls6–7pm (23.38%); Friday (22.21%)8–9am (17.64%); Tuesday (19.29%)
Weeks, Kulka and Pierson, Public Opinion Quarterly, 1987First-call contact and interview completion, US householdsWeekday evenings and weekendsWeekday daytime
Innovations for Poverty Action, 2021, nine countriesSurvey completion on first attemptsMidday, slightly ahead of morningEvening
Same IPA studyDay of weekNo statistically significant difference—

A few things stand out.

The gaps are small. In the CallHub data the best and worst hours differ by under six points, and the best and worst days by about three. That's worth having, but it won't rescue a campaign that's failing for other reasons.

Evenings win in rich countries and lose in some others. The two US studies, 33 years apart, both found evenings beat daytime. The IPA study across lower- and middle-income countries found evenings the weakest slot. Habits, work patterns and phone costs differ. If you call several countries, don't assume one schedule fits all of them.

The legal window cuts off the best hour. CallHub excluded 9–10pm calls from its analysis because they fall outside the compliant window, yet that hour showed a 34.87% answer rate, along with a 1.98% do-not-call rate. People answer late because they're home. They also get annoyed, and in the US it isn't allowed for most marketing calls.

Sales studies are a different population. Much of what you'll read about "best time to cold call" comes from B2B sales vendors measuring calls to office phones. Business buyers at their desks behave differently from consumers on mobiles. If you call customers at home, B2B data is the wrong benchmark.

Before optimising inside a window, know where the walls are. As of November 2026:

  • US marketing calls. The FCC's rule at 47 CFR 64.1200(c)(1) bars telephone solicitations to residential subscribers before 8am or after 9pm, local time at the called party's location.
  • US debt collection. Regulation F, 12 CFR 1006.6(b)(1)(i), treats a time before 8am or after 9pm local time as inconvenient, unless the collector knows otherwise. If someone tells you they work nights, the default no longer protects you.
  • UK consumer credit collections. The FCA's CONC 7.9.4 says a firm "must not contact customers at unreasonable times" and must pay due regard to customers' reasonable requests about when, where and how they're contacted.
  • States and other countries may set narrower windows or limit how often you call. Check each place you dial.

Notice the common thread: local time at the person's location. A US list spans four mainland time zones. A European list may span three. Calling "from 9am" in your own time zone means calling some people at 6am theirs.

This isn't legal advice. Calling-hour rules depend on the purpose of the call, the country and sometimes the state, and they change. Check your specific campaign with a lawyer.

Why "when after the trigger" beats "what time of day"

For many calls the clock matters less than the gap since something happened. Someone who filled in a quote form at 2:14pm is most reachable, and most interested, in the next few minutes. A player who registered but didn't deposit is worth calling in a day or two, not a month later. A customer whose card payment failed this morning has the problem on their mind today.

So split your calls into two kinds:

  • Triggered calls. The best time is "soon after the trigger, within the allowed window". If the trigger fires at 10:30pm, the call goes at the start of the next morning's window, not at the next Tuesday 4pm slot a study recommended. Our speed-to-lead guide covers the inbound-lead case.
  • List calls. A renewal list, a win-back list, a reminder list. Here time of day and day of week are the levers you actually have, and testing them pays.

A sensible default schedule

If you have no data of your own yet, start here and adjust.

Who you're callingFirst attemptSecond attemptAvoid
Consumers, personal mobilesLate afternoon to early evening on weekdaysLate morning on a different day, or Saturday late morningBefore 9am; anything near the legal cut-off
Existing customers about their accountLate morning or early eveningThe opposite half of the dayLunchtime is mixed; test it
Business contacts on office linesMid-morning or mid-afternoon on weekdaysA different half of the dayMonday first thing, Friday afternoon
Several countriesEach person's local evening (test per country)Local late morningAssuming your home country's habits travel

The second-attempt column is where most schedules go wrong. If someone missed your 11am call because they're in meetings every weekday morning, calling at 11am again tomorrow will fail for the same reason. Move the retry to a different part of the day.

How to find your own best time to make calls

The only number that really matters is the one from your list. Getting it is straightforward, but most people under-sample.

  1. 1

    Pick two windows, not six

    Choose the comparison that would change what you do, such as 11:00–14:00 against 17:00–20:00 local time. Every extra window splits your calls further.

  2. 2

    Split the list randomly

    Put people into each window at random, not by region or signup date. Otherwise you're comparing audiences, not times.

  3. 3

    Keep everything else identical

    Same script, same caller ID, same retry rules, same weeks. Change only the time window.

  4. 4

    Measure the right thing

    Answer rate first (a human picked up), then the outcome you care about: bookings, payments arranged, offers accepted. A window that's answered more but converts less isn't the winner.

  5. 5

    Wait for enough calls

    Check whether the difference is bigger than the noise before acting on it. The arithmetic is below.

How many calls is enough?

Worked example. You call 500 people in each of two windows. Window A gets 18% answered, window B 22%. Is B better?

At around a 20% answer rate, the standard error for 500 calls is the square root of (0.2 × 0.8 ÷ 500), about 1.8 percentage points. The standard error of the difference between two such windows is about 1.8 × √2 ≈ 2.5 points. Your 4-point gap is about 1.6 standard errors. That's suggestive, not conclusive: it could easily be chance.

To detect a 4-point gap at roughly 95% confidence you need about 2 × 0.16 × (1.96 ÷ 0.04)², around 770 calls per window. To detect a 2-point gap you'd need about four times that. If your list is a few hundred people, don't slice it six ways and crown a winner.

Setting it up in Dial Milo

Dial Milo is an AI phone agent for inbound answering and outbound calling campaigns. A few of its settings map directly onto everything above.

  • Calling hours in each person's local time. You set days and hours (the default window is 10:00 to 20:00) and Milo applies them where each person lives. It reads a timezone column if your list has one, or works the zone out from the number, including US and Canadian area codes. In countries with several zones, it only calls when the time is inside your hours in every zone the number could be in. The calling hours help article has the detail.
  • Retries spaced apart. Set attempts per person, for example 2 tries 4 hours apart, so a late-morning miss gets a second go in the late afternoon. No-answers and voicemails are retried; Milo hangs up on voicemail rather than leaving your brand on it.
  • Scheduling per person. If you trigger calls from a CRM or journey tool, the at field queues a call for a specific moment, and calling hours still apply on top. The CRM-triggered calls guide shows the payloads.
  • Callbacks requested on the call. When someone says "call me after six", Milo records a callback with a time and calls back then. That's the person telling you their best time, which beats any study.
  • Callbacks into the campaign number. People often ring back hours later. Milo recognises them and continues the conversation, so a missed call in the morning can become a completed call at lunchtime. See what happens when people call back.

For a timing test, the cleanest approach is two copies of the same campaign with different calling windows. Use Copy on the campaign card (it keeps the settings and scripts but no people), change only the hours on the copy, and upload a random half of the list to each. Compare answer rates and outcomes when both have finished.

What a test costs

Illustrative figures. Testing two windows with 800 people each is 1,600 people. Say 20% answer and calls average two billed minutes. Outbound runs roughly 14–20¢ a minute all in for US, UK and Dutch numbers, so the 320 answered calls cost about $90 to $130. Calls that don't connect have no minute charge, and voicemails cost little because Milo hangs up. For a real reference point, an October 2026 pilot by an online casino operator made 121 dial attempts to 61 players for $11.53, about 9.5¢ per attempt. The spend calculator prices your actual list, and you can check any number on the rates page.

The short version

  1. Call inside the legal window, in the person's local time.
  2. For triggered calls, call soon after the trigger; for lists, start with weekday early evening for consumers.
  3. Put retries in a different part of the day from the first attempt.
  4. Let people tell you their best time, and call back when they say.
  5. Test two windows with enough calls before changing your schedule.

If you want to see how a campaign handles time zones and retries before committing, Dial Milo is pay as you go from a $10 top-up, and the browser demo lets Milo call you before you pay. Create an account and run a test call on yourself first.

Frequently asked questions

What is the best time of day to call customers?
For consumers on personal mobiles, early evening tends to do best: one study of 2.2 million US calls found the highest answer rate between 6pm and 7pm and the lowest between 8am and 9am. But studies disagree, the gaps are a few percentage points, and your own list can behave differently. Treat early evening as a starting hypothesis and test it.
What is the best day of the week to make calls?
There is no reliable universal answer. One large US study found Friday best and Tuesday worst, a survey study across nine lower- and middle-income countries found no significant day-of-week difference, and sales vendors often report Monday. Day matters less than time of day and much less than how quickly you call after a trigger.
What hours am I allowed to call in the US?
Federal TCPA rules bar telephone solicitations to residential subscribers before 8am or after 9pm in the called person's local time, and Regulation F treats the same hours as inconvenient for debt collection calls. State laws may be stricter, so check the states you call.
Should I call at the same time on every retry?
No. If someone didn't answer at 11am on a Tuesday, the same slot next Tuesday is likely to fail for the same reason, such as a work meeting. Space retries so the second attempt lands in a different part of the day, for example a late-morning first try and an early-evening second.
How many calls do I need to find my own best time?
More than most people expect. To tell an 18% answer rate from a 22% one with reasonable confidence, you need roughly 750 to 800 calls in each window you're comparing. With smaller lists, test two windows rather than six, or you'll be reading noise.

Never miss a call. Never forget to call back.

From about 12¢ a minute. Outbound calls nobody answers are free. $1 credit to try it.

Already have an account? Sign in