Best Time to Call Customers: Answer Rates by Hour and Weekday

The best time to call online shoppers in Loop Voice's September 2026 benchmark was Thursday at 1 pm, when 13.6% of calls were answered, about twice the 6.6% rate across all calls. On weekdays, calls placed between 5 pm and 9 pm were answered more often than calls between 8 am and noon (7.0% against 5.4%), with every hour counted in the shopper's local time.

The figures come from the Loop Voice Ecommerce Benchmark, September 2026: 61,389 calls from 7 online stores to 18,361 shoppers who had started a checkout or placed an order, so none of them were cold calls. Loop Voice placed the calls and publishes the report, and the data is free to quote under CC BY 4.0.

Each cell is the share of calls a person answered in that hour of the shopper's local day, so 1 pm means calls placed from 1:00 to 1:59 pm. Voicemail and call screening count as unanswered, and blank cells had fewer than 200 calls, which the benchmark greys out.

Percent of calls answered. Source: Loop Voice Ecommerce Benchmark, September 2026, data file, which adds a 95% interval for each published cell.

• Thursday held the two best hours: 1 pm (13.6%) and 8 am (10.6%). The 8 am cell rests on 1,536 calls (95% CI 9.2 to 12.2) and the 1 pm cell on 290 calls (95% CI 10.1 to 18.0), so 8 am is the steadier number.
• Weekday mornings were the weak spot. Of the 15 weekday cells from 9 am to noon, 13 fell below the 6.6% average.
• Mid-afternoon was steady. Nine of the ten weekday cells at 2 pm and 3 pm were at or above the average, led by Friday at 3 pm (10.1% on 1,084 calls).
• Sunday at 8 am was the low point at 1.9% of 369 calls, and weekend and Thursday hours from 4 pm on had too few calls to publish.

The evening edge has a catch. The 8 pm hour has one published cell (Monday, 5.7%), so the weekday figure for 5 pm to 9 pm comes mostly from calls placed before 8 pm.

Read the grid as a direction, not a schedule. Like the benchmark's other timing figures, these are comparisons across shoppers and stores, not a controlled test:

• Calls weren't assigned to hours at random. Nine in ten of September's calls (55,725 of 61,389, per the data file) were abandoned-checkout calls, timed by when each shopper left checkout rather than by the hour, so a cell mostly reflects when those shoppers were shopping, plus the retries that followed.
• Every attempt counts. The rate divides answered calls by all calls placed, so an hour heavy with second calls can look worse than an hour of first calls.
• Answered is not ordered. The benchmark publishes no order rates by hour. To learn what calls add, run a holdout test.
• It's one month from seven stores. September 2026 had four Thursdays, and Monday's cells include Labor Day on September 7 (OPM). No store makes up more than 15% of a figure, but a quiet week can still move a cell built on 290 calls.

For an abandoned checkout, the shopper picks the hour by leaving. What you control is the wait, and the benchmark's speed-to-call figures favor calling within minutes over holding the call for a better hour.

The wait counts calling-hours minutes only (8 am to 9 pm in the shopper's time zone), so a checkout abandoned at 11 pm and called at 8:04 am counts as 4 minutes. In September, 72% of checkouts were abandoned during calling hours and got a first call a median 31 minutes later.

Source: Loop Voice Ecommerce Benchmark, September 2026. Picked up is the share of shoppers, not of calls, and the under-5-minute band (165 shoppers) was too small to publish. These bands also compare different shoppers and stores, so take the direction, not the exact gap.

1. Don't hold a checkout for a better hour. A checkout abandoned at 10 am on a Tuesday, one of the grid's weakest cells, should still get its first call within 5 to 15 minutes. The delay before that call is a setting in Loop Voice's abandoned checkout template, and Insights shows the time from abandon to first dial, so you can check that first calls land in that band.
2. Let the hour shape the second call. The template calls, texts the checkout link, waits, then calls again with a discount. If first calls land in the morning, a wait of a few hours moves the retry into the stronger afternoon cells, and the text after the missed call covers the gap.
3. Don't write off late-night checkouts. Called the next day, they ordered at 6.5%, more than daytime checkouts that waited 1 to 4 hours (4.7%). A shopper who orders overnight drops out of the flow before the morning call.

A campaign calls a list you choose, once, on the schedule you set, so this is where the grid earns its keep. Loop Voice outbound campaigns (in beta) take a campaign time zone, a daily calling window and a pace in calls per minute, and the window applies in each customer's local time.

Example: 1,800 customers at 5 calls a minute need 360 minutes of dialing, so a weekday window from 2 pm to 8 pm can finish the list in one day.

• Start after lunch. Weekday cells from 9 am to noon were the week's weakest block.
• Stop at 8 pm. The evening edge came mostly from calls before 8 pm, and Florida's limit for sales calls is 8 pm.
• Skip Sunday at 8 am. It was the lowest published cell, at 1.9% of 369 calls.

September's calls were mostly abandoned-checkout calls, and customers who haven't ordered in months may keep other hours. Test before you commit: set up two campaigns for comparable segments with different windows and compare their answer rates, since every call attributes back to the campaign that made it.

Example: at a 6.6% answer rate, 100 calls give a 95% interval of roughly 3.2% to 13.2%, too wide to rank two windows, while 1,000 calls narrow it to about 5.2% to 8.3%. Give each window at least 200 calls, the benchmark's own floor, and more when the two rates sit close together.

The legal window sets the edges of everything above. The federal rule bars telephone solicitations to a residential subscriber before 8 a.m. or after 9 p.m., local time at the called party's location (47 CFR 64.1200(c)(1)). Florida stops commercial telephone solicitation calls at 8 p.m. in the called person's time zone and allows no more than three from any number to a person in 24 hours on the same subject (Florida Statutes 501.616(6)), and other states set their own limits.

Loop Voice applies quiet hours in each customer's own time zone. Before any hour matters, AI-voiced marketing calls need the shopper's prior express written consent, and opt-outs have to be honored. Our guide to the rules for AI calls covers both; confirm the details for your states with counsel.

Most published studies of the best time to call come from calling-software companies and phone-banking tools, and they measure cold or political calls. The benchmark's calls went to shoppers who had just started a checkout or ordered, so the populations differ as much as the methods.

Two of the three agree with the benchmark that mornings are a weak time to call: CallHub's lowest hour was 8 to 9 am, and WAVV calls 6 to 9 am the worst block of the business day. MightyCall found the opposite, with 8 to 11 am strongest, but it counts a call as successful only if it connected and lasted 30 seconds, and it doesn't say whom its customers were calling.

Compare the shape of the day, not the percentages. Each study defines an answer its own way, and WAVV warns that its hours aren't each recipient's local time.

Seven stores' shoppers are not yours. Loop Voice's Insights shows your answer rate against the prior window, who answered (a person, voicemail, an unreachable number or a call screener), pickup by attempt and the time from abandon to first dial, and call records export to CSV.

1. Export a month of calls and count answered and total calls by hour and weekday, in each shopper's local time.
2. Group hours into blocks, such as weekday mornings, afternoons and evenings, so each block passes 200 calls sooner.
3. Move your campaign window or retry wait only when one block beats another by more than the noise in the example above, then check again the next month.

New benchmark issues appear on the benchmarks page, and the answer rates by state show how much a shopper's location moves the same numbers.