Nobody talks about the 90-day ramp. Here is exactly what happens each week when you build a cold email program from scratch.
Rees Bayba
Founder, Astra GTM
TL;DR
The most common question we get from new clients: 'How long until we see meetings?' The honest answer is 90 days for a new program built from scratch. That timeline frustrates people who want pipeline yesterday. But skipping steps to move faster is how you end up six months in with burned domains, a 2% inbox placement rate, and zero meetings to show for it.
This is the exact ramp we see across every new client engagement. Not theory -- the actual week-by-week reality of building a cold email program that produces consistent results. We have launched 8+ programs from zero in the last 18 months. The timeline varies by a week or two, but the phases are always the same.
Nothing goes out in the first two weeks. This is the part most people want to skip. Do not skip it. Everything you build in months 2 and 3 depends on the infrastructure you set up now. Rushing this phase is the single most common reason cold email programs fail.
Buy dedicated sending domains. Plan for 3 domains per 1,000 emails you want to send per day. If your target is 500 emails/day, that is 2-3 domains minimum with room to grow. Use variations of your brand name -- not your primary domain. If your company is acme.com, buy acmeHQ.com, getacme.com, acmeteam.com. Never send cold email from your primary domain.
At 30 sends per mailbox per day with 3 mailboxes per domain, each domain handles ~90 sends/day. Plan infrastructure before you need it -- adding domains mid-campaign means another 2-3 week warmup.
Want this built for your team?
We implement these systems end-to-end. First campaigns live in 14 days.
New mailboxes have no sender reputation. ISPs treat them like strangers -- which is exactly what they are. Warmup services send and receive automated emails between your new mailboxes and a pool of established accounts, gradually building a history of legitimate activity. Minimum 14 days. We run 21 days for any client where deliverability risk is high (regulated industries, large enterprises on the ICP).
During warmup, do not sit idle. This is when the real strategic work happens. Define your ICP with precision -- not just job titles and company sizes but signals that indicate timing and need. Build your target account list. Start contact enrichment so verified emails are ready the moment warmup completes.
Use the warmup window to build a launch list of 500-1,000 verified contacts. Run every email through verification (we require a deliverability score of 97+ before any contact enters a campaign). This is not optional. Sending to unverified lists in your first week of real sends will damage the reputation you just spent two weeks building.
Warmup hits 70-80% completion around day 14. You can start sending real emails, but at low volume. Think 50-100 total sends per day across all mailboxes. This is not the time to blast your entire list. It is a controlled ramp designed to build reputation through real prospect engagement while the warmup process finishes in the background.
Your first sends should go to your highest-confidence contacts -- verified emails at companies that closely match your ICP. You want early engagement (opens, replies) to signal to ISPs that your emails are wanted. Send to your weakest contacts first and a wave of bounces or spam complaints will undo your warmup progress.
At 50-100 sends per day, you will accumulate 350-700 sends by end of week 4. That is not enough to judge copy performance. You need 500+ sends per variant before any conclusion is meaningful. Use this phase to verify deliverability, not to evaluate creative. Send test emails to seed accounts at Gmail, Outlook, and Google Workspace. If 90%+ land in the primary inbox, your infrastructure is working.
You might get your first reply in this window. Or you might not. Both are normal. One reply from 400 sends is a 0.25% rate -- impossible to distinguish from noise at this sample size. Do not panic. Do not rewrite everything. Stay the course.
Below this, you are reading statistical noise. A single reply changes your 'rate' by 0.2 percentage points. Wait for real data before making changes.
Month 2 is the inflection point. Warmup is complete. Volume ramps to 500-1,000+ emails per day. You have enough sends to start seeing statistically meaningful patterns in your reply data. And this is exactly where most people and most teams give up.
The pattern is always the same. A founder or VP of Sales launches cold email expecting immediate results. Week 3 passes with no meetings. Week 4 brings one lukewarm reply. Week 5 they start questioning the channel. Week 6 they pull the plug and declare 'cold email does not work for our market.' Meanwhile, the data was just beginning to get useful.
By the end of month 2, you should have 4,000-8,000 emails sent. At a 0.5-1.5% positive reply rate (normal for mid-market B2B), that is 20-120 positive replies. With 60-70% of positive replies converting to meetings, you are looking at 12-84 meetings. The range is wide because it depends entirely on your ICP, copy quality, and deal size.
Realistically, most new programs land in the 3-8 meetings range in month 2. Not 84 -- that is the ceiling for a scaled program with exceptional copy and a responsive ICP. Three to eight meetings from a program that did not exist 60 days ago is a strong signal that the system is working. The question is whether you have the patience to let it compound.
| Metric | Concerning | Normal | Strong |
|---|---|---|---|
| Total sends (month 2) | Under 2,000 | 4,000-8,000 | 8,000-15,000 |
| Positive reply rate | Under 0.2% | 0.3-1.0% | 1.0%+ |
| Meetings booked | 0 | 3-8 | 8+ |
| Bounce rate | Above 5% | 1-3% | Under 1% |
| Spam complaint rate | Above 0.3% | Under 0.1% | Near 0% |
This is where most teams quit. They expected 20 meetings by now and got 5. But month 3 is where compounding kicks in -- you have data to optimize against and infrastructure that is fully seasoned.
Month 2 is when you start making data-driven copy decisions. You have enough sends per variant to identify winners and losers. Kill the variants pulling below 0.3% positive reply rate. Double down on anything above 0.8%. Start A/B testing new openers against your best performer -- but only one variable at a time. Changing the opener, CTA, and subject line simultaneously tells you nothing.
Also review your targeting. Look at which segments are replying and which are silent. Sometimes an entire industry or company size band is unresponsive -- not because the copy is wrong but because your product is not solving a problem they recognize. Remove dead segments from your active campaigns and redirect volume toward responsive ones.
Month 3 is where a well-built program starts feeling like a machine. You have 60+ days of send history building your domain reputation. Your copy has been tested against real data. Your targeting is refined based on who actually responded. The work you did in months 1 and 2 compounds here.
At this stage, most of our clients are sending 1,000-2,000+ emails per day across multiple campaigns and segments. Monthly meeting output ranges from 10 to 25+ depending on volume and ICP. The cost per meeting has dropped because your infrastructure is amortized and your copy conversion rates are higher.
For a program sending 1,000-2,000 emails/day with optimized copy and targeting. Volume depends on ICP accessibility and infrastructure capacity.
| Week | Activity | Send volume | Expected meetings |
|---|---|---|---|
| 1-2 | Infrastructure setup, DNS, warmup begins, ICP research, list building | 0 prospect emails | 0 |
| 3-4 | Warmup completing, first sends, deliverability testing | 50-100/day | 0-1 |
| 5-6 | Volume ramp, first meaningful reply data | 200-500/day | 1-3 |
| 7-8 | Full volume, first copy optimization cycle | 500-1,000/day | 3-6 |
| 9-10 | A/B testing at scale, segment expansion | 1,000-1,500/day | 5-10 |
| 11-12 | Steady state, predictable pipeline | 1,000-2,000+/day | 8-20+ |
These numbers assume you are doing the work at each phase -- not just sending emails and hoping. The teams that hit the high end of these ranges are the ones optimizing copy weekly, cleaning their lists aggressively, and expanding infrastructure ahead of volume needs.
Not every slow start means something is broken. But some signals are real problems that patience will not solve. Here is how to tell the difference.
Everything above assumes a cold start -- new domains, new mailboxes, no existing sender reputation. If you already have warmed infrastructure (maybe from a previous campaign or an existing outbound program), the timeline compresses significantly. You can skip the 2-3 week warmup phase and start sending real emails in week 1.
With pre-warmed infrastructure, first meetings can arrive as early as week 2-3. By week 6 you should have meaningful data and be well into the optimization phase. This is why agencies with existing infrastructure can promise faster results -- they are not starting from zero on the infrastructure side.
But even with pre-warmed infra, copy still needs testing, ICP still needs validation, and data still needs enrichment and verification. You save 2-3 weeks on warmup. You do not skip the learning curve on what message resonates with your specific market.
Changing everything at once. A campaign runs for 5 days with no replies. The founder rewrites the subject line, changes the opener, swaps the CTA, adjusts the target list, and increases send volume. Now they have no idea what was working and what was not. They have reset their experiment to zero.
The second biggest mistake: optimizing for the wrong metric. Teams chase open rates, which are increasingly unreliable due to Apple's Mail Privacy Protection and corporate email scanners that auto-open everything. A 60% open rate tells you almost nothing in 2026. Focus on positive reply rate and meetings booked. Those are the only two numbers that connect to revenue.
From zero infrastructure to predictable pipeline. You can see first meetings by week 5-6, but consistent, optimized output takes the full quarter. Plan for it.
Can I speed up the warmup process?
Not meaningfully. Some warmup tools claim 7-day warmup is sufficient. Our data shows that mailboxes warmed for fewer than 14 days have 15-20% lower inbox placement rates in the first month of real sending. The extra week of patience saves you months of deliverability problems. Budget 14 days minimum, 21 days for high-stakes ICPs.
How many domains do I need to start?
For a starting volume of 300-500 emails per day, plan for 5-6 domains with 3 mailboxes each (15-18 mailboxes total at 30 sends each). If you are starting smaller at 100-200 per day, 2-3 domains will cover you. Always have at least one backup domain warming up in case you need to rotate one out.
When should I start A/B testing copy?
Not until you have 500+ sends on your baseline variant. Starting A/B tests before that means you are splitting an already-small sample in half. Run your best guess as the only variant for the first 500-1,000 sends. Then introduce a challenger variant and split traffic 50/50. Wait for each variant to hit 500 sends before declaring a winner.
What if I get zero replies after 2,000 emails?
First, confirm your emails are actually reaching inboxes (seed test). If deliverability is fine, your copy is the problem -- not a tweak, a fundamental rewrite. Change your opener from a question to a situation-naming statement. If that still fails after another 1,000 sends, revisit your ICP. You may be targeting people who do not have the problem you solve.
Is it worth hiring an agency vs. doing this myself?
If you have never run cold email before, an agency compresses the learning curve from 6-12 months to 90 days. They bring pre-warmed infrastructure, tested copy frameworks, and operational knowledge that prevents the most expensive mistakes. The trade-off is cost ($2,000-8,000/month) and less direct control. If you have time, technical comfort, and tolerance for a slower ramp, DIY is viable.
How do I know if my ICP is wrong vs. my copy is wrong?
Check the negative replies. If people say 'not interested' or 'bad timing,' your ICP is probably right but your copy or timing is off. If people say 'we don't do that,' 'wrong person,' or 'this isn't relevant to us,' your targeting is the problem. Negative replies are diagnostic gold -- read every single one.
What should my follow-up sequence look like?
Three to four total touches over 10-14 days. First email is your best shot -- situation-naming opener, specific ask. Follow-up 2 (day 3-4): add a new angle or proof point, not 'just checking in.' Follow-up 3 (day 7-8): share something genuinely useful (a relevant stat, a case study). Follow-up 4 (day 12-14): breakup email, short and direct. More than 4 follow-ups show diminishing returns and increase spam risk.
What results should I promise my CEO or board?
Month 1: infrastructure live, first sends flowing, zero or near-zero meetings. Month 2: 3-8 meetings from the ramp phase. Month 3: 10-25+ meetings depending on volume and ICP. Set expectations for the 90-day ramp explicitly. If someone expects 20 meetings in month 1, you need to reset that expectation before you start, not after.
We implement these systems end-to-end. First sends within 14 days.