How long cold email takes to work
Expect two to three weeks before anything can go out cold, four to eight before a first meeting is plausible, and a full quarter before the number of meetings means anything. That is the calendar question, not the how-many-words-should-my-email-be one. Here is what you should be seeing at each point, and what would justify stopping.
By Kshitij Maheshwari, co-founder · Updated August 2026 · 14 min read
What you should actually be seeing by now
Most of the wait is machinery, and machinery works the same in Paris as in Denver. Find the row you are standing on, then read the last column.
| When | What is actually visible | What it means | Kind of claim |
|---|---|---|---|
| Weeks 0 to 2, before any cold send | Nothing sent. Records set, new addresses building a track record. | Nothing here is about your market. Judge the build. | Mechanical |
| Week 1 of sending | Bounces and delivery failures, back within hours. | Your first real feedback, and it is about your list. | Mechanical |
| Weeks 1 to 2 | A handful of replies, nearly all of them a no. | Words, not a rate. Read them one at a time. | Practice |
| Any day, all quarter | Most contacts are still partway through their follow-ups. | Count who has finished before you count anything else. | Mechanical |
| Weeks 3 to 5 | The first contacts reach the last touch you planned. | Your first readable date, not the day you started. | Mechanical |
| Weeks 4 to 8 | A first meeting, if your weekly pace supports one. | Proof a person clicked. Not proof a market responds. | Mechanical |
| Weeks 6 to 10 | The same objection coming back from every slice. | The cheapest verdict of the quarter, and an early one. | Practice |
| Around day 60 | A read on which slice answers, and why. | The real output of a first quarter, not a meeting count. | Practice |
| Beyond your quarter | Deals close, on a calendar that is not yours. | Under ten months on average in one European study. | Measured |
Sources: 6sense's November 2025 survey of European B2B buyers, on purchases of $25,000 and up; Eurostat's short-term business statistics, monthly averages for 2000 to 2019. Rows marked mechanical follow from how sending works and carry no country. Rows marked practice are how we run the motion.
Three things are running, and only one of them is yours
Founders judge outbound against one calendar, their own. Three are running at once, they start on different days, and two of them ignore you completely.
The build is everything before a cold email can leave: domains, records, and the stretch where a new sending address earns a track record. The sequence is the set of follow-ups one contact receives, spread over weeks. The buyer decides on their own schedule, and started long before you.
- •Which slice you write to, and what you offer them
- •How many contacts you can research in a week
- •How fast you answer a reply, which is the same day
- •Whether the build was done properly the first time
- •A sequence occupies about a month per contact by design
- •Providers throttle a new sending address on their schedule
- •The buyer starts when their problem gets loud enough
- •None of it moves faster because you are anxious
The first weeks feel like failure because two of the three have not started yet. You cannot send cold at any price until the build is done, and the reasons are published by the providers themselves: the limits and the order of operations are in our cold email infrastructure guide.
You can speed up the build, and that is the whole list. The other two run on their own schedule.
Your campaign is younger than your calendar
One founder's first quarter, told as what she could actually see. What she built in each phase, and what it cost, is in our seed outbound playbook.
An illustrative walkthrough of the method, not a specific client result. We report real numbers only when they are real.
Nothing, then a little
She starts on 1 September and researches 50 contacts a week. By week four: 100 contacts in, four bounced, and three replies that all say some version of we built this in-house.
The number does not exist yet
200 contacts in, none finished, no meetings. She compares her rate to one she read somewhere and decides it is not working. Every one of the 200 is mid-sequence.
What the quarter produced
500 in, about 300 finished, three meetings held, which is not yet a pattern. Four objections she heard over and over, and one slice answering differently. That is the output.
Her sequence has touches on days 1, 3, 6, 13 and 20 and closes on day 34, so one contact takes 34 days to finish. At 50 a week, week six holds 200 started and none finished. The waiting is not idle either: the follow-ups are the work, and what to put in them is in our post on following up after no response.
Count finished contacts, not started ones. Replies keep arriving on the later touches, past the tenth in Instantly's 2026 benchmark report, so a contact who is halfway through has not really been asked yet.
The thing that moves before the meetings do
Evidence arrives in almost the exact reverse of how much you care about it. Everybody watches the last one first.
Hours, and it is about your list
A bounce is an email that came straight back because the address was wrong, and it says nothing about your offer. If this one is ugly, fix it before you read anything else. The mechanics are in our deliverability guide.
Week one, and worth more than they look
A handful of replies will not divide into anything, and they are still the best evidence you get all quarter. What each volume of contacts actually buys you is laid out in our guide to turning outbound into GTM learning.
Month two, if your volume is there
A rate needs enough finished contacts to stop swinging around, and most seed teams never send enough to buy one they can defend. Before you call a variant a winner, read why your A/B test is lying to you.
Week four to eight, and slowest of all
The last thing to move and the first thing everybody checks. The arithmetic behind a meeting target is in the seed outbound playbook, and if replies arrive without meetings, run the reply to meeting audit.
Your first meeting is a coin flip. Your fifth is information
The first one feels like the verdict on the whole quarter. It is the least informative thing that will happen in it.
Luck, not evidence
Meetings are rare next to contacts however good you are, so whether the first lands in week four or week nine turns on which few people happened to answer. Same campaign, same underlying rate, different Tuesday.
Write down why it happened
First meetings often arrive for a reason you cannot repeat: somebody who already knew you, or a lurker who was going to reach out anyway. If you cannot name the reason, you cannot run it again.
The fifth one carries the information
By the fifth you can see whether the same kind of person keeps saying yes, and for the same reason. The objections you collected on the way there tell you more than any single meeting.
One meeting proves a person clicked. Five meetings and four repeated objections prove a market is talking back.
What working even means, and when you get to ask
Most arguments about whether outbound is working are two people using different finish lines. A quarter can settle some of them and not others.
"Did outbound pay for itself in 90 days?"
- ✕Uses closed revenue as the test
- ✕Puts your quarter on the buyer's calendar
- ✕Fails a working motion for finishing late
"Who answered, what did they say, and can we do it again?"
- ✓Replies and meetings sit inside your control
- ✓Repeated objections arrive by week four
- ✓Revenue gets judged over a longer stretch, honestly
was the average European buying cycle in 2025, meaning how long a buyer takes from noticing the problem to signing.
6sense, from a November 2025 survey of buyers in Belgium, France, Germany, the Netherlands and Sweden. Those are six-figure purchases, not yours. The same survey found European buyers start the conversation themselves about eight times in ten.
So judge a quarter of outbound on conversations, and never on closed revenue. Whatever you sell and wherever you sell it, the buyer decides on their own calendar, and some of what you send this quarter gets paid back in the next one.
Not sure if yours is too early to judge or genuinely stuck?
Book a Fit CheckCount working weeks, not calendar weeks
A 90-day test that runs through the month your market is on holiday is a 60-day test, and nobody warns you at the time.
is where EU industrial output sits in August, with Italy closer to 40% down and the Nordics emptying out in July instead.
Eurostat, short-term business statistics, EU-27. It measures factories rather than inboxes, and Germany barely moves at all. It is still the same holiday.
The dead weeks move by country. August across southern Europe, July across the Nordics, the last two weeks of December nearly everywhere, and Thanksgiving week in the United States. Those last two are how we plan rather than a published figure.
So count working weeks for the country you are writing to, and start a test after the dead ones rather than through them.
Book the review off the calendar
Kicking off a 90-day outbound test on 15 July. Review it in October.
- ✕Two of those weeks were never going to work
- ✕The review date arrives before the evidence does
- ✕A quiet August reads as a dead market
Book the review off the work
Ten sending weeks from 1 September, skipping the last two weeks of December. Review when the list has been through once.
- ✓Counts weeks your market is at its desk
- ✓The finish line is a finished list, not a date
- ✓Written down before anybody is anxious
What actually justifies stopping
Four things are real reasons to stop. Every one of them is something you observed, not a date you reached.
You worked the whole list and it is quiet
Every reachable, verified contact through every touch, and nothing came back. That is not impatience, that is the input running out, and it is a complete answer.
Everybody gives you the same specific no
Quiet everywhere points at the offer. The same specific no everywhere points at the list. Our guide to turning outbound into GTM learning makes the case that this is the cheapest read you get, and it arrives by week four.
You cannot get delivered at all
This kills the infrastructure, not the motion. Stop the count, fix the mechanics, restart the count, and do not read the burned weeks as evidence about your market, because they hold none.
You ran out of hours before you ran out of list
The honest kill, and the commonest. It is a decision about who runs the motion rather than whether it works, and both sides are costed in agency versus in-house.
Five reasons to stop that are really just a bad week
From the inside, these four feel exactly like the four above. The fifth, in red, is the one that costs the most.
Six weeks at 50 contacts a week and six weeks at 300 are different experiments with the same date on them. Elapsed time on its own is not a finding.
Usually a divide-by problem. Your dashboard and the report you are comparing yourself against are rarely dividing by the same thing.
Small counts swing hard, and a silent week inside a busy quarter is noise. Change what you send before you change your mind about the market.
A test result is a verdict on one line, not on the motion. Rewrite the line and keep sending; the test was never asking the bigger question.
Stopping while most contacts are unfinished
The most common false kill on this list, and the most expensive. Six weeks into a month-long sequence, most of the people you wrote to have not had every touch, so the rate you are about to quit on is not a rate yet. You are paying for a quarter and reading it at week six.
Kill the slice, not the motion
A founder asking whether to kill outbound usually means something much smaller: whether to keep sending this thing to these people.
- 1 Decide which question you are asking before you answer either one.
- 2 Killing one slice is cheap and reversible. Killing the motion is neither.
- 3 Write the stopping rule in week zero, while nothing is at stake.
- 4 A rule written down is a decision. The same call on a quiet Friday is a mood.
A rule can be one sentence: if the whole list has been through the whole sequence and nobody has said anything specific, we change the offer.
What we tell founders to expect
We say roughly the same thing on every fit check, and none of it is a promise about how many meetings land.
The week-three question is always a version of am I allowed to stop yet. I ask two things back: how many of your contacts have finished, and what did the ones who answered actually say. Almost nobody has the first number, and almost everybody has the second written down somewhere unread.
From kickoff to a clear verdict in 60 days, on a two-month start, then month to month with no lock-in. The verdict is a read on who answers and why. We do not promise a number of meetings, and nobody honest will.
Questions founders ask
How long before I get my first meeting from cold email?
Is three weeks long enough to tell whether cold email is working?
Why am I getting no replies after two weeks of sending?
When should I stop a cold email campaign?
Does the time of year matter for outbound?
My first meeting was great. Does that mean it is working?
Co-founder of Real Good GTM. He has been the first business hire and Chief of Staff at seed-stage B2B startups, building outbound pipeline before any playbook existed. This post is the answer he gives founders in week three, when the campaign looks dead and the only honest reply is a question about how many contacts have finished.
Connect on LinkedInWhat to do with the weeks you have
The build plan, the evidence each volume of contacts buys you, and the motion for changing one slice.
The seed outbound playbook
The first 90 days phase by phase, what each one costs, and the arithmetic behind a meeting target.
Read the playbookTurning outbound into GTM learning
What each volume of contacts actually buys you, and which conclusions you have already paid for.
Read the guideThe ICP slice experiment
How to change one slice, size the batch so it can be read, and act on what comes back.
See the playWant a straight answer on whether to keep going?
Book a fit check. We'll look at how much of your list has actually finished, read what came back in words, and tell you straight whether this is too early to judge or genuinely stuck.
Book a Fit CheckNo hard sell. No fake numbers. Real good work speaks for itself.