Bad Calls.
A number worth staring at.
A campaign manager asked me last year why her connection rate had dropped. She assumed the phone was dying. Spam labels, call screening, younger voters who never pick up. All the usual explanations.
So I asked what her connection rate actually was. Around 9 percent. Then I asked what she thought a strong operator hits. She guessed 15.
That gap is what this report is about. We pulled 82 million dials across 2.5 years of real calling, and the same pattern held everywhere. The operators who struggle and the operators who win are usually working the same lists, the same carriers, the same voters. What separates them is not the channel. It is when they call, how many times, on which dialer, and how clean their numbers are.
None of that is luck. All of it is a decision.
We built this benchmark so you can see where you land, and so the decisions that move you up a tier stop being guesswork. Where do you think you'll land?
What you're holding, and what to do with it.
This is a benchmark, not a manifesto. We pulled 83 million real outbound dials, spanning political, advocacy, nonprofit, and labor union campaigns running on CallHub from 2024 through today, and lined them up side by side. The question wasn't whether calling still works. It was: on the exact same lists, the exact same carriers, the exact same voters, why does one operator connect on 3 dials in 4 while another barely clears 1 in 10?
That gap between an average campaign and an elite one is what this report measures, broken into the specific decisions that separate them.
Specifically, here's what you'll find inside:
Connection rate by operator tier: where the platform average sits, where the typical (median) operator sits, and where the top decile sits, so you know which of the three you're actually being compared against.
Timing: an hour-by-hour, day-by-day heatmap showing when a dial is most likely to connect, broken out by vertical.
Retry discipline: exactly how much a second and third dial is worth, and where retrying stops paying off.
Dialer mode: why a lower connection rate on a predictive dialer can still mean more live conversations per agent-hour than a higher rate on power dialing.
Four verticals, three years: connection-rate trend lines for political, advocacy, nonprofit, and union campaigns, including which are shrinking their lists on purpose and which are scaling into trouble.
What elite operators do differently: the specific, testable levers, like number rotation and spam shield, that show up disproportionately at the top of the distribution.
- Three numbers describe three different stories, and most of this report is about not confusing them.
- Platform average (13.7%): dragged down by a small number of very large operators dialing enormous cold lists. It describes them, not you.
- Median operator (34.2%): the fairest peer comparison for most teams reading this.
- Top decile (73.6%): the aspirational ceiling, proof of what's achievable on the same lists and the same carriers everyone else is using.
- The move: benchmark against the median, aim for the top decile, and ignore the average.
The best operators reach 3 people in 4. The typical operator reaches 1 in 3.
Across 83 million dials, the median operator connects on about 1 in 3. The platform-wide average, by contrast, is closer to 1 in 7, and it is tempting to read that lower number as proof the phone is finished. It is not. The average is dragged down by a few very large operations dialing enormous cold lists. Look at where an individual operator actually lands and the picture changes completely.
Stop benchmarking against the platform average. A few giant operations dialing enormous lists drag it down, so it reads far lower than a normal operator. Measure against the median operator at 34.2% instead, and chase the 73.6% ceiling.
This report shows what is associated with stronger connection rates. It does not prove why. Connection rate is shaped by many things at once: list quality, timing, dialer mode, number reputation, retry discipline, and audience type. Our data shows the variation without ranking the causes. Where a lever clearly works on a fixed list, like time of day, we say so. Everywhere else, treat these findings as the highest-value things to test, not guarantees.
Find yourself on the board.
This is the benchmark itself: three tiers, three numbers. Compare your own connection rate against the median, top 25%, and elite figures below to see exactly where you stand, and how far the climb to the top really is.
| Operator tier | Connection | Talk time | Dials / contact | Cost / conversation |
|---|---|---|---|---|
| Median operator | 34.2% | 43s | 1.46 | $0.26 |
| Top 25% | 64.5% | 46s | 1.45 | $0.18 |
| Top 10% “elite” | 73.6% | 50s | 1.37 | $0.19 |
Drop your own connection rate onto this table. Under 34.2% puts you below the median operator, and the climb to the top is practice, not budget.
Big operations carry a hidden connect-rate tax.
Group every operator by how much they dial and a clear pattern appears: connection rate falls as volume climbs. Part of that is a simple statistical effect, since bigger lists reach deeper into colder, harder-to-verify contacts. But it is not a law. The best large operators beat the curve, and the gap between them and the rest is the opportunity.
| Operator size (annual dials) | Median connection | Share of all dials | |
|---|---|---|---|
| 1,000 – 10,000 | 41.1% | 1.7% | |
| 10,000 – 100,000 | 28.9% | 8.6% | |
| 100,000 – 1 million | 22.5% | 22.1% | |
| 1 million and up | 9.4% | 67.5% |
The lesson is not "dial less." Volume is the point for most outreach programs, and scale is a strength. The lesson is that scale and connection rate pull against each other unless you actively manage the list, so the biggest operators have the most to gain from tighter targeting and cleaner data before the dial.
If you dial at scale, treat connection rate as a list-quality dial you can turn. Segment, verify, and suppress before you dial. That is how the best large operators beat the curve without dialing less.
There is no best time to call. There is a best time to call each audience.
Connection rate swings hard by day and hour. Most campaigns dial the 2pm to 6pm block, the coldest stretch on the board. The hours that connect sit at the edges, in the recipient's local time. Hover any cell to read it.
The window flips by vertical
| Vertical | Best day | Best window |
|---|---|---|
| Union | Mon–Thu | Mornings, 7–9am |
| Nonprofit | Tue–Thu | Mornings & midday |
| Political | Mon–Thu | Evenings, 8–9pm |
| Advocacy | Tue–Thu | Evenings, 7–9pm |
Two numbers to keep
On a power dialer, weekday 8–9pm connects at 53–56%, versus 39–43% in the 5–6pm slot most campaigns favor.
Canadians answer far more than Americans: 53.7% vs 36.3% on a weekday power dialer, a 17-point gap that holds across every hour.
Move dials out of the 2 to 6pm dead zone. Political and advocacy skew to 8 to 9pm, union and nonprofit to 7 to 9am. Same list, same budget, and connection rate can climb sharply. Stay inside calling-hour rules while you do it: reach each contact in their own local time. Unified, time-zone-aware scheduling makes that automatic.
The first 3 dials do the real work.
Connection rate is highest on the first dial. The real cliff is between the first and second attempt. After that, retries keep paying off at a much gentler decline.
| Share of all dials | Share of all connections | Share of all outcomes |
|---|---|---|
| 84% | 91% | 89% |
A retry is not wasted effort. A second dial still connects roughly 1 time in 10, and a third about 1 in 12. That's a far gentler decline than the drop from dial one to dial two. Those are real conversations single-dial campaigns never claim.
Retries pay off most where the contact is warm and durable: current donors, dues-paying members, past voters you've reached before. A missed call costs nothing there, and a live one is worth a lot.
They pay off least on a large, one-time cold-prospect push against a hard deadline. A fourth attempt at a contact who hasn't picked up is often better spent as a first attempt at someone new.
The discipline is retrying well: about 3 attempts, well spaced, on fresh numbers that have not been burned. CallHub's number rotation and spam-label shield keep the second dial from landing as "Scam Likely."
Cap cadence at 3 attempts, then stop. Move the budget from dials 4 and beyond onto fresh, un-burned numbers, and space the 3 attempts across different days and day-parts.
A low connection rate can still be the most productive dialer.
Predictive dialing connects at a lower rate than power dialing, yet it puts far more live conversations in front of an agent every hour. The pacing engine does the work in the background.
Connection rate
Successful outcomes / agent-hour
On predictive, stop grading by connection rate and track conversations per agent-hour instead. On cold lists, test predictive: it can quadruple outcomes per hour at half the cost per contact.
Same phone, different playbook.
Political, Advocacy, Nonprofit, and Union make up the vast majority of outreach calling, and they run the identical infrastructure to very different ends. "Successful" below means an agent-logged positive outcome, such as a pledge, an RSVP, or a completed survey, as a share of connected calls.
| Vertical | Dials | Connection | Voicemail | Drop | Talk | Successful |
|---|---|---|---|---|---|---|
| Union | 4.5M | 28.6% | 39.1% | 1.2% | 69s | 15.4% |
| Nonprofit | 1.4M | 22.9% | 38.9% | 4.6% | 57s | 8.4% |
| Advocacy | 2.5M | 19.1% | 45.6% | 5.6% | 46s | 3.6% |
| Political | 70.9M | 11.8% | 44.4% | 2.8% | 39s | 14.1% |
The verticals split cleanly by list type. Union and Nonprofit call people who already know them, and they connect deep. Political and Advocacy run broad persuasion and turnout at massive volume, which trades connection rate for reach. Neither is wrong. They are different jobs, and they need different playbooks.
Know which game you are in, and benchmark against your own vertical, not the blended average. A warm member list should expect union-style depth. A cold turnout universe should plan for reach over connection rate, and resource it accordingly.
Connection rate by vertical, by year
Each number is that vertical's annual connection rate: the share of its dials that reached a live person. The blended average looks like a tidy election cycle. Underneath, every vertical is on its own track.
| Vertical | 2024 | 2025 | 2026* | Direction |
|---|---|---|---|---|
| Advocacy | 14.3% | 20.3% | 26.9% | Rising every year |
| Political | 9.8% | 17.3% | 10.8% | Tracks the election calendar |
| Nonprofit | 29.6% | 23.4% | 14.4% | Falling |
| Union | 41.9% | 31.8% | 20.6% | Falling; volume grew ~5x over the same period |
| All verticals | 10.9% | 19.6% | 13.6% | Political-driven cycle |
Scale and connection rate move together
Advocacy got smaller and tighter, and its connection rate climbed three years running. Union moved the opposite way on both dimensions: volume grew roughly fivefold while its rate fell by half. We can't say scale caused the drop, but the two track closely enough to treat list size as one of the first things to check when a vertical's rate is under pressure.
Conversations are getting longer
Talk time rises every year regardless of the cycle: 41s to 43s to 49s overall, and 48s to 83s for Union. Whatever else moves, the conversations that connect are deepening.
Watch your trend, not just your number. A falling connect rate while volume climbs means you are scaling past your list quality. Advocacy reversed exactly that by getting tighter, not bigger.
The elite don't dial harder. They dial smarter.
The top 10% do not connect twice as often because they push more volume. They run a different kind of operation.
The highest-connecting operators are not the biggest dialers. They run smaller, cleaner, better-targeted lists, and they reach a person in fewer attempts and hold longer conversations once they do. We cannot prove which of those causes the others. What the data shows is that these traits travel together at the top, so the safest instruction for everyone else is not "dial more." It is "target better, and protect the numbers you dial with."
Reaching a person in 1.37 dials instead of 1.46 tracks with cleaner lists and protected numbers. De-dupe, verify, and suppress before the dial, and guard your caller reputation. These are the habits that cluster at the top, and the highest-value things to test on your own operation.
See where you land.
The benchmark tells you what good looks like. CallHub helps you get there: better timing, smarter dialers, cleaner numbers.
Book a DemoHow we counted.
- Dataset
- 82,760,961 outbound dials to US and Canadian recipients. Static snapshot, data through mid-2026.
- What counts as a connection
- A real human answer: agent-set disposition, distinct from the sub-account, duration greater than zero, system ANSWER, human verdict from answering-machine detection where enabled, and agent disposition not equal to MACHINE. This mapping is fixed across every figure in the report.
- Verticals
- AI- and manually-verified account classification covering 99.4% of dial volume, not self-reported profiles. This report focuses on the four core outreach verticals: Political, Advocacy, Nonprofit, and Union.
- Tiers and the distribution
- Operator tiers rank the 582 accounts with at least 1,000 dials by connection rate, reported as the median account within each tier. Platform-wide aggregates are volume-weighted across all dials and run lower than the median operator, because the largest operations account for the majority of dial volume and connect at lower rates. Both views are true, and the report states which it is using.
- Exclusions
- Test campaigns, internal and staff accounts, and demo traffic removed. Attempt codes at or above 100 excluded (about 1% of dials).
- Privacy
- All published figures are aggregated to the tier, industry, or platform level. No individual account or contact is identifiable in this report.
The terms.
- Connection rate
- The share of total dials that reach a live human. The headline metric.
- Connected ACD (talk time)
- Average connected call duration: the mean length of a live conversation, in seconds.
- Dials per contact
- The average number of dial attempts made to reach a contact. Lower is more efficient.
- Voicemail rate
- The share of dials answered by an answering machine rather than a person.
- Drop / abandon rate
- The share of answered calls dropped before an agent connects. On predictive dialers this is also a compliance signal.
- Successful outcome
- An agent-logged positive result, such as a pledge, an RSVP, or a completed survey, measured as a share of connected calls.
- Dialer modes
- Preview: the agent sees each contact before the dial. Power: the system dials at a set pace per agent. Predictive: the system paces dialing to agent availability and screens non-human answers.
- Operator tier
- Accounts ranked by connection rate into the median, top 25%, and top 10% ("elite").
- Patch-through
- A call connected live and transferred to an agent or a target line.