B2B cold calling reach benchmarks, July to October 2026: what 7,615 dials show

Summary of findings

• About one dial in eight reached a person: 12.9% of 7,615 rung-out dials (982). About one in nine reached a decision-side contact: 11.1% (843).
• Reach by attempt to the same number: 13.6% on attempt 1, 15.1% on attempt 2, 8.2% on attempt 3 and 5.5% on attempt 4 or later.
• By calendar month, a decision-side contact was reached on 12.2% of dials in July, 10.0% in August, 11.7% in September and 10.3% in October to the 14th.
• Numbers with Canadian area codes reached a human on 16.2% of dials (220 of 1,356), against 12.1% (760 of 6,257) for numbers with other North American area codes (the United States or another territory in the North American numbering plan).
• Voicemail was the outcome of 29.0% of dials (2,209). The share was higher on later attempts: 22.4% on attempt 1 and 44.7% on attempt 4 or later.
• Across hour bands of the prospect's local day, reach of a human ranged from 11.4% to 16.1% (verified-clock period only, 3,628 dials). We do not read that as a best time to call.

What happened to the dials

Of 7,615 rung-out dials, 982 reached a person, 2,209 ended in voicemail, 4,267 were not answered and 157 reached a number that was wrong or not in service.

Reach by attempt number

The first two attempts to a number reached a person at similar rates, 13.6% and 15.1%. The rate fell to 8.2% on the third attempt and 5.5% from the fourth onward. Decision-side reach shows the same shape: 11.6%, 13.1%, 7.0% and 5.3%.

This is observational. Later attempts go to numbers that were already dialed, and whether a number is dialed again can depend on how the earlier attempts ended, which can select for numbers that are wrong, unattended or screened. The data do not show that a fourth call is a worse call, only that numbers still being dialed a fourth time were less likely to produce a person than numbers dialed for the first time.

Reach by month

Reaching a decision-side contact stayed within a narrow range each month: 12.2% in July, 10.0% in August, 11.7% in September and 10.3% in October. October covers October 1 to 14 only. Months are calendar months in UTC.

Be careful with the first bar of each pair. The "gatekeeper" outcome was logged 1 time in July and 15, 57 and 66 times in the following months, so the "reached a human" figure contains a very different share of gatekeepers in July than in later months. The decision-side measure excludes gatekeepers and is the one to compare over time.

Canadian and other North American numbers

Country is read from the area code. Canadian numbers reached a human on 16.2% of 1,356 dials and a decision-side contact on 12.6%. Numbers with other North American area codes (the United States or another territory in the North American numbering plan) reached a human on 12.1% of 6,257 dials and a decision-side contact on 10.7%. Voicemail was 31.3% and 28.5%. 2 dials had no usable area code and are in neither group. There were fewer Canadian dials, and we have not controlled for caller, list or month, so the gap should not be read as a country effect.

Voicemail share

Voicemail ended 29.0% of dials. By attempt it rose from 22.4% on the first attempt to 41.0%, 39.6% and 44.7%. By month it was 28.9% in July, 34.2% in August, 12.1% in September and 37.6% in October. The September figure is low enough that we doubt it reflects who picked up. How callers chose between logging voicemail and logging no answer is one candidate we have not checked, so treat the monthly voicemail shares as weaker than the reach rates.

Reach by hour of the prospect's local day

Reach of a human across the five bands runs from 11.4% to 16.1%. The bands are not balanced for caller, list or month, and we have not tested whether the differences are larger than chance. The chart supports one narrow statement: no band reached a person on fewer than one dial in nine, and none on more than one in six. The voicemail share is not charted by hour because the last band is under the 20-event minimum.

This chart uses only the 4,737 rung-out dials from the period whose clock we could verify (August 12 to October 14), of which 3,628 have a resolved prospect time zone and a local time from 08:00 to 18:59. Our first pass found 508 rung-out dials that map to a local time from 04:00 to 07:59. 39 of them are in the verified period and fall outside this chart's 08:00 to 18:59 window. We could not explain the other 469, so they, and every dial from the earlier period, are left out of this chart. The method is explained under time zones below.

How we calculated this

Population and period

The source is the call log of Vivin's own dialer. It holds 9,744 call attempts from 2026-07-03 to 2026-10-14 to 5,694 distinct business phone numbers, placed by 13 callers. The busiest caller made 35.5% of the rung-out dials (2,705 of 7,615). Reach varied several-fold between callers with 100 or more dials, so these figures describe one team, not callers in general.

A rung-out dial is a logged attempt that is not a technical failure and was not ended before an answer. We exclude 2,129 attempts: 322 dial errors, 36 blocked by the telephony provider's policy, 1,705 cancelled while ringing and 66 abandoned. That leaves 7,615 rung-out dials, the population for this page. Each rate states its own denominator, the dials in that group, and the hour chart uses only the subset described under time zones. Of these, 7,613 had a Canadian or other North American area code and 2 had no usable area code.

Definitions

• Reached a human: the call was logged as one of gatekeeper, not interested, callback scheduled, booked, interested for follow-up, information requested or do not call.
• Reached a decision-side contact: the same list without gatekeeper.
• Voicemail: logged as voicemail or as a voicemail message left. Wrong or out-of-service numbers are counted in neither.
• Attempt number: the order of non-technical dials to the same phone number, from the start of the data.
• Country: inferred from the area code of the dialed number.

Time zones

Timestamps are stored in two forms. Dials from July 3 to August 10 carry a UTC label; all of them were loaded into the log in one import on August 12, so there is no independent creation time to check them against. Dials from August 12 carry no zone label. For all 6,059 of those we compared the stamp with the system's own UTC creation time and confirmed it is Toronto local time (6,059 of 6,059 matched within five seconds).

The prospect's local time comes from the area code, using Eastern, Central, Mountain, Pacific and Atlantic zones. It resolves for 75.2% of all logged attempts. It is an inference: a number can belong to someone who has moved.

The early-morning dials. 508 rung-out dials map to a prospect-local time between 04:00 and 07:59. 469 are in the imported period and 39 are in the verified period. Of the imported period, 240 dials fall before 08:00 on the Toronto clock itself, while no dial in the verified period falls before 08:00 or after 21:00 Toronto time (0). Two explanations fit the imported period: the dials really happened that early, or the UTC label on the import was wrong. We could not tell which, so we exclude the whole imported period from every hour-of-day figure. The other figures on this page do not depend on the hour, so they use all 7,615 dials.

Cells under 20 are withheld

We publish a rate only when at least 20 events sit behind both its numerator and its denominator. Smaller cells are shown as "withheld". This page contains no names, phone numbers, area codes, company names, list names or caller names, and no individual call dates.

Caveats

• Observational data. Repeat dials go to numbers that were dialed before, and whether a number is dialed again can depend on how earlier attempts ended, which biases later attempts.
• One team's calling, with its own scripts, lists and callers. Mix changed between months.
• There is no industry, company size or title field attached to the calls.
• There is no outcome data after a meeting is booked.
• The share of dials logged as gatekeeper and as voicemail differs sharply between months, and we have not established why.

What we cannot claim

• Show or no-show rates, because attendance was not recorded.
• Booking rates. Most cuts have fewer than 20 bookings, and the bookings are concentrated in one month, so a rate by cut would mostly describe that month's team, script and lists.
• Any cut by industry, company size or seniority.
• Anything about regions outside North America.
• A best time of day or week to call.
• That these rates will repeat for another team, list or offer.

How to use the numbers

Treat them as a reference point for your own call log, not a target. Compare like with like: use rung-out dials as the denominator, report decision-side reach separately from gatekeepers, and count attempts per number. If your reach on the third attempt is far below your reach on the first, check the numbers before you check the caller. Read the figures alongside your own lists and scripts.

Next edition

The next edition is planned for the week of January 18, 2027 and will cover the quarter ending December 31, 2026, with the same definitions. Any change to a definition will be listed here. We plan to add appointment outcomes and industry fields to the records so that show rates and industry cuts can be reported when each cell has at least 20 events.