Cold Email9 min read

I sent 100 cold emails and got 0 replies — here is what was actually wrong

Everyone assumes the copy is broken. In most zero-reply post-mortems we run, the copy is fine. The emails simply never reached an inbox a human opens.

The founder who sent us the campaign was ready to scrap the whole thing. One hundred hand-built messages to ICP-fit prospects, personalised first lines, a genuinely interesting offer, and nothing. Not a single reply. Not even a “not interested.” The dashboard said 98 delivered, 41 opened, 0 replied.

He asked us to rewrite the copy. We asked him to run an inbox placement test first. The copy was fine. Out of those 98 “delivered” emails, 22 landed in the primary inbox. The rest were in Promotions, Junk, Quarantine, or invisible behind Microsoft Defender's spam policy. You cannot reply to an email you have never seen.

TL;DR

Zero replies on a cold campaign is almost never a copy problem. It is a funnel problem: list → delivered → inbox → read → reply. Most teams optimise step 5 and ignore steps 2, 3 and 4, which is where 60–80% of the leakage lives.

The cold email funnel nobody draws

Your CRM shows you two numbers: sent and replied. Between them are four invisible steps where your pipeline quietly dies:

  1. Accepted by the receiving server — no 5xx SMTP error at connection time.
  2. Placed in a folder — Inbox, Promotions, Junk, Quarantine, Focused-Other, or silently dropped.
  3. Surfaced to the human — did it show up in the folder they actually read today?
  4. Read by a human — not a bot, not Defender, not a preview pane prefetch.
  5. Motivating enough to reply — this is where copy matters.

Every cold-email course on the internet teaches step 5. Very few teach steps 2, 3 and 4. That is why the average reply rate for a “good” SDR campaign in 2026 is under 1.5%.

The autopsy: what we found in those 100 emails

We re-ran the same subject line and body through a 20-mailbox seed test. The result:

Gmail primary:        31%
Gmail Promotions:     44%
Gmail Spam:           25%
Outlook Focused:       8%
Outlook Other:        55%
Outlook Junk:         37%
Yahoo Inbox:          60%
Yahoo Spam:           40%
Apple iCloud:         72% / 28% Junk
Corporate (Defender): 12% / 88% quarantined

Of 100 prospects, roughly 35–40 physically saw the email in a folder they read. Of those, the true conversion to reply looked like 2–3% — which means the expected number of replies was 1 or 2, not the 10 he was aiming for. The founder wasn't failing at copy. He was failing at arithmetic he could not see.

Why Promotions counts as dead

Cold outbound in the Promotions tab is functionally worthless. B2B buyers filter Promotions as noise, and most never see a Promotions message until batch-deleting it on a Sunday evening. A reply rate of 0.1% on Promotions placement is normal.

Why the “sent” column keeps lying to you

Every ESP dashboard — Apollo, Instantly, Smartlead, Lemlist, Mailshake, Woodpecker — shows you the same column labelled “delivered” or “sent.” All it measures is: did the receiving server accept the SMTP handshake without a hard bounce. That is a fantastically low bar. A server can accept a message, then silently drop it, silently route it to Junk, silently quarantine it, or silently hand it to a filter that swallows it whole. All of those states count as “delivered” in your dashboard.

Inbox placement is a completely different measurement. It requires actually logging into a seed mailbox and checking which folder the email arrived in. You cannot derive it from SMTP logs. You cannot estimate it from open rates (opens are broken for other reasons). You have to test it.

A 10-minute diagnosis you can run today

  1. Pick one prospect-sized audience. A recent 50-to-500 sequence where you already know the reply rate.
  2. Send the same template to seed mailboxes. Gmail, Outlook, Yahoo, Mail.ru, Yandex, GMX, ProtonMail, Apple iCloud, and one corporate Defender inbox.
  3. Record folder placement per provider. Inbox, Promotions/Focused, Junk/Spam, Quarantine, Missing.
  4. Check SPF, DKIM and DMARC alignment. Any fail on any provider = re-auth.
  5. Look at spam engine scores. Rspamd, SpamAssassin, Barracuda-style verdicts.
  6. Compare expected vs actual reply rate. If inbox % is 30%, your real reply-rate denominator is your prospect list × 0.30, not the full list.
Run the diagnosis free

Inbox Check sends your template to 20+ real provider seed mailboxes and returns per-provider Inbox/Spam/Missing verdicts plus SPF/DKIM/DMARC alignment and spam-engine scores. Free, no signup. If you prefer to automate it, the API does the same thing on a schedule.

What to fix first, second, third

  1. Authentication. SPF, DKIM and DMARC must pass on every provider. One misaligned selector can tank Gmail placement.
  2. Domain & IP reputation. New domain? Needs 3–6 weeks warm-up. Purchased domain? Check history. Shared IP? Know your neighbours.
  3. Volume pacing. 50/day ramping by 10%/week until a few hundred. Spikes trigger filters.
  4. Content fingerprints. Excessive links, tracking pixels from cheap domains, and heavy image-to-text ratios are three of the easiest wins.
  5. List hygiene. Invalid addresses, catch-alls and role addresses hurt sender reputation. Verify before sending.

The mindset shift: deliverability is not a tax, it is the product

Cold email is not a copywriting competition. It is a distribution competition. The best message in the world in a Junk folder is worth exactly zero. Founders who treat deliverability as plumbing — something to “set up once” — keep getting the zero-reply autopsy back. Founders who treat it as a weekly KPI see reply rates triple without touching a single word of copy.

FAQ

How many replies should 100 cold emails actually produce?

With 95% inbox placement, good ICP fit and decent copy, 2–5 replies. With 30% placement (typical for untested campaigns), 0–1. The gap between those two outcomes is almost entirely deliverability, not craft.

Can I just improve copy until replies come?

No. Copy only multiplies what reaches the inbox. If your placement is 20%, doubling copy quality doubles 20%-of-something — still small. Fix placement first, then tune copy against a denominator that is worth tuning against.

Is Promotions tab placement fixable?

Mostly, yes. Reduce link count, remove marketing patterns (footer with 8 social icons, unsubscribe styled like a newsletter), send from a person-style address, and drop obvious marketing subject lines. It is not a binary switch but it moves.

How often should I re-test placement?

Before every new sequence, after any infrastructure change (new domain, new ESP, new IP), and monthly as a baseline. Provider filters change; your setup will drift even when you don’t touch it.
Related reading

Check your deliverability across 20+ providers

Gmail, Outlook, Yahoo, Mail.ru, Yandex, GMX, ProtonMail and more. Real inbox screenshots, SPF/DKIM/DMARC, spam engine verdicts. Free, no signup.

Run Free Test →

Unlimited tests · 20+ seed mailboxes · Live results · No account required