Outbound metrics and benchmarks: what to track and what good looks like

The seven-stage outbound funnel

Every stage needs one definition, written down, before anyone compares numbers. Most disagreements about benchmarks come down to different denominators.

Pick one denominator for reply rate and keep it. We use sends, which is the simplest to audit, and write down when a report uses delivered instead. Open rate is not on the list. Smartlead states that it does not use open rate because Apple Mail auto-loads images, so reported opens are largely automatic (Smartlead; checked 18 October 2026), and Woodpecker advises readers to "treat reply rate as the number that matters" (Woodpecker; checked 18 October 2026).

The planning benchmarks we run against

When we plan a campaign, we use three ranges: 2 to 3 percent of sends reply, 15 to 50 percent of replies are positive, and 30 to 50 percent of positive replies book. These are not a published study, and a first test on a new list may land below them.

Multiplying them gives a range for every 1,000 sends.

The spread is wide because the three ranges multiply. A campaign at the low end of every range books about one meeting per 1,000 sends, and one at the high end books about seven. The distance between them is mostly the quality of the list and the strength of the offer, which is why we spend more time on those than on subject lines. For the copy side, see how to write cold emails that book meetings.

One of our cases shows what a good start looks like. In the Fema Tape engagement, 850 firms were contacted in month one and produced 32 positive replies against a target of 15, and 37 meetings in two months (Fema Tape case study). That was a strong result from a small, scored list, and you should not read it as an average.

What published benchmark studies say

All figures below are third-party data from vendors that sell email tools, checked on 18 October 2026. Each is measured on that vendor's own customers, so none of them is a market-wide average.

These numbers do not agree because they count different things. A median sits below a mean when a few strong campaigns lift the average. A rate per delivered email is higher than a rate per sent email. A rate per sender, weighted by who is in the sample, differs from a rate per campaign. Smartlead filters to senders with at least 500 emails, Woodpecker excludes agencies from some tables, and Instantly counts only human replies after removing bounces and auto-responses.

Our 2 to 3 percent planning range sits above Woodpecker's median and Smartlead's typical sender, and just below Instantly's average. We plan for targeted campaigns on scored lists, so the range assumes that work has been done. A first campaign at 1 percent sits inside the typical range of these datasets. It has not yet shown that your list and offer beat the typical.

Two more cuts from the Woodpecker data apply here. Woodpecker found reply rate falls as campaigns grow: independent companies had a 2.7 percent median at 0 to 50 prospects per campaign and 0.5 percent at 10,001 or more. And bounce rate matters: campaigns with 0 to 2 percent bounce had a 1.6 percent reply rate against 1.0 percent for campaigns with 20 to 40 percent bounce. Instantly advises keeping bounces below 2 percent, which matches what Smartlead calls its danger line.

A typical day-90 funnel (planning model)

The table below is a planning model, not a client result and not a promise. It shows the volumes a well-run outbound system might reach about 90 days after launch, once infrastructure is warm and the list and offer have been tuned.

The implied rates sit inside the three planning ranges. The sends range depends on whether you send on weekdays only (about 22 days) or every day of a 30-day month. The model does not say how many emails each contact receives, so read the signal and enrichment rows as the intake that feeds the sending row, and check your own ratio of contacts to sends. A contact that gets a four-email sequence uses four sends.

The model also stops at booked meetings. For held meetings, the only figure we fetched is RevenueHero's December 2024 analysis of 6,428 scheduled meetings across 15 industries, which found a 6.5 percent no-show rate (RevenueHero; checked 18 October 2026). The page does not say whether the meetings were inbound or outbound, and the data comes from one scheduling vendor's customers, so cold-booked meetings may see more no-shows. As an illustration, 65 booked meetings at 6.5 percent no-shows is about 61 held, and at 20 percent it is 52.

We do not publish a planning rate for opportunities. The share of held meetings that sales accepts depends too heavily on who takes the call and how the offer is qualified, and we found no third-party figure with a stated sample. Measure it yourself from the first 20 or 30 held meetings. For a full view of the system behind these numbers, the cold email playbook for B2B shows how the stages connect.

Diagnosing each stage

When the booked-meeting number is low, work backwards from the first stage that falls below range. Fixing a later stage first wastes effort, because everything downstream inherits the earlier problem.

Check for low sample sizes before acting. At 1,000 sends and a 2 percent reply rate you get 20 replies, and a swing of five replies either way is within normal variation. Evan Miller's guidance on A/B testing makes the same point from the other side: decide the sample size in advance and avoid stopping an experiment the moment it looks significant, because repeated peeking raises the false-positive rate (Evan Miller; checked 18 October 2026). For outbound, that means judging a change after a few thousand sends, not a few hundred.

A weekly scorecard

Review the same table every week. The thresholds below mix published limits with our own working targets, and the source column says which is which.

Google tells senders to keep the spam rate reported in Postmaster Tools below 0.10 percent and to avoid ever reaching 0.30 percent (Google; checked 18 October 2026). Our cold email systems service measures success on booked meetings, and the stages above explain why a number is low.

Questions

What is a good cold email reply rate?

Our planning benchmark is 2 to 3 percent of sends. Third-party vendor data is lower or higher depending on definitions: Woodpecker reports a 1.5 percent median, Smartlead 0.74 percent for the typical sender, and Instantly a 3.43 percent average on delivered emails.

What counts as a positive reply?

A reply that asks for or accepts a next step, such as a call, a sample or more information. Auto-replies, out-of-office messages, unsubscribe requests and clear refusals do not count.

How many meetings should 1,000 cold emails book?

Using our planning benchmarks, roughly 1 to 7 booked meetings per 1,000 sends. The range is wide because list quality and offer strength change every stage of the funnel.

Should I track open rate?

It is not a reliable signal. Smartlead does not use it because Apple Mail auto-loads images, and Woodpecker advises treating reply rate as the number that matters.