A survey says most cold messages get deleted unread. I send them anyway. Here is what I took from it.

On 2 October 2026, PPC Land wrote up a survey that sales data company Firmable published on 19 August.

Firmable asked 1,009 full-time professionals what they do with a cold business message. According to PPC Land, 52% said they delete it without reading a word.

I run cold outreach for more than one of my businesses. So I read the whole write-up, twice.

The headline is weaker than it looks. One finding underneath it is worth more than the headline.

Step 1: what the survey found

Firmable split the answers into five bands, which PPC Land lists as: delete without reading 52%, skim then delete 30%, read but rarely reply 11%, reply if relevant 6%, almost always reply 1%.

Those five add up to the whole sample, which is the first thing I check on any chart.

Then the triggers. Respondents could pick several, and looking obviously mass-sent came first at 63%, according to Firmable's figures as PPC Land reports them.

A generic "Hi {FirstName}" greeting came next at 37%, and getting the name, role or company wrong at 36%, per PPC Land's report.

None of that is new to anyone who has sent outreach. It is still useful to see it ranked.

Step 2: why the headline is the weak part

Here is the complication.

When Firmable shared the study with PPC Land on 2 October, its summary described 82% as deleting cold outreach on sight. PPC Land pointed out that this merges the 52% who delete unread with the 30% who skim first.

A skim is a read. A short one, but your first line got seen.

Then the method. PPC Land says Firmable's methodology note runs to three sentences. It does not say when the fieldwork happened, which countries respondents live in, how they were recruited, or whether the results were weighted.

And the company that ran it sells sales data and outreach tooling. That does not make it wrong. It does mean the survey was built to say something about outreach, by people with a product in the answer.

So I am not quoting Firmable's 52% as a benchmark for my own numbers. I do not know whose inboxes it describes.

Step 3: the finding I am keeping

This one survives the caveats, because it is about how people rank reasons, not about a rate.

According to Firmable, 13% of respondents said human authorship counts most when deciding whether to reply, while 58% said relevance or genuine value.

So the question I had been asking, whether a message sounds like a person wrote it, is the smaller one.

But it does not get you off the hook. The same survey found 61% distrust AI outreach they receive, per PPC Land. People do not reply because a human wrote it. They do delete things that read as generated.

The way I hold both: the message has to be relevant first, and it must not read like a template second. Those are different jobs.

Step 4: the double standard

The survey's own framing, quoted by PPC Land, is "a double standard".

According to Firmable, 54% of respondents send cold outreach themselves.

I do not read that as hypocrisy. Everyone hates bad outreach. Nobody complains about the message that solved a problem they had that week.

Step 5: four rules worth keeping

Four rules, all cheap. One is mine already; the other three are where the survey points.

1. First line about them, from something public. If I cannot write one true sentence about the recipient's business from what they have published, I do not send the message. That rules out most of a scraped list, which is the point.
2. Names checked by hand. Getting the name, role or company wrong was a trigger for 36% of respondents in Firmable's survey. A merge field that breaks is a send I chose not to check.
3. No follow-up that says "just following up". It shows up in Firmable's list of stock phrases that put readers off. A follow-up has to add something new or it does not go.
4. Fresh leads only. I already throw away any lead older than 48 hours, for the reasons in I throw away every lead older than 48 hours. Relevance decays. A message about a problem someone posted last month is about a problem they may already have solved.

The survey did not invent any of these. It ranked the reasons people give, which is the most I would ask of it.

Step 6: how I read any outreach survey now

This one gave me a checklist I will use on the next one.

1. Who ran it, and what do they sell? A sales data company asking about cold outreach is not neutral. Read it anyway, with that in mind.
2. Do the bands add up? If a chart sums to more than the whole, it is a multiple-choice question, and the bars cannot be compared like shares.
3. Is the headline a merge? Two groups added together make a bigger number and a weaker claim.
4. Where do the respondents live, and how were they found? If the method does not say, the number describes nobody in particular.
5. Is it a rate or a ranking? Rankings of reasons travel better than rates, because a skewed sample can still rank things the same way.

Most outreach advice I see quotes one number from a survey like this and stops there. The number is the least useful part.

Step 7: what this means if you are the one being pitched

If you run an agency, you get pitched too.

The survey's best line for a buyer is the one about relevance. A cold message that names something specific about your business, and offers something useful before asking for anything, has done more work than most.

If you are an editor or clipper doing the outreach yourself, the channels matter as much as the message. VALORAE Cast wrote that up in Your first video editing client: the three channels that work.

The short version

• Firmable's survey of 1,009 professionals, per PPC Land, found 52% delete cold messages unread and 30% skim then delete.
• Firmable's headline 82% merges those two groups, as PPC Land notes. A skim is still a read.
• The method does not say where respondents live, how they were found, or whether results were weighted. Use it as direction.
• The finding worth keeping: relevance beats human authorship as the reason to reply.
• Do not let a message read as generated. Respondents distrust that.
• Check names by hand, cut stock follow-up lines, and only message fresh leads.
• Treat every survey from a company selling the cure as a pitch, then read it anyway.

I write up what I learn running VALORAE at tgsidd.com.