Read before you write

You know the message. You have received it, probably this week. Hi Priya, I see you're the Head of Marketing at a fast-growing SaaS company. We help marketing leaders like you drive pipeline with less effort. Would you be open to a quick chat? It is grammatical, it is polite, and the only fact in it is your job title, which is the one fact about you that needed no reading to find.

I have received that message more times than I can count, and the first tool I looked at would have had me send it: it merged a first name and a title into a template and called that personalisation. People accept invitations, so the accept rate on it would have been fine. The reply rate is the problem, and so is the feeling of pretending to have read a profile you never opened.

Two messages, same person

Here is an invented person, so I can show the difference without quoting anybody real. Call her Priya, Head of Marketing at a forty-person company that sells scheduling software to clinics. Last Tuesday she wrote a post about turning off her paid search for a month to find out how much of the pipeline it actually owned, and what she learned.

Built from the title: Hi Priya, as a marketing leader at a growing healthtech company, you're probably under pressure to prove channel ROI. We help teams like yours attribute pipeline accurately. Open to a quick chat next week?

Built from the post: Hi Priya, the paid-search blackout experiment was the most useful thing I read this week, mostly because almost nobody is brave enough to run it. Curious what the organic numbers did in week three. I work on something adjacent and would rather compare notes than pitch.

The first one could have been sent to every Head of Marketing on the platform, and she knows that, because she has received it. The second one could only have been sent to her, and the reason it reads like a person wrote it is that a person, or something that read what she wrote, had to.

What the product does about it

Every person in your morning queue arrives with the message already written, and under it a line that says what it was written from. Sometimes that is the profile: the role, what the company does, what they did before. The warmer leads come with the thing they actually said: the post they wrote about the problem you solve this week, or the comment they left under a competitor's post. That line is there so you can judge the draft against its evidence in a glance, and skip the ones where the evidence is thin.

A draft too close to a recent one is thrown away and rewritten. The cheapest thing for a recipient to spot is the same note sent a hundred times with the name swapped, and the second cheapest is the same opening line. Then you approve, edit or skip, one at a time, for ever. Nothing goes out under your name that you have not read, which is also the only honest way to run a tool that writes in your voice.

Why this is a safety feature and not a nicety

It is tempting to file this under reply rate, and it does matter there. But look at LinkedIn's own reasons for restricting an account, which are on the limits page with the source: many invitations ignored, left pending, or marked as spam. The title-built message is the one that gets ignored. The post-built one is the one that gets accepted, and acceptance is the number the breaker watches. Reading before writing is how the accept rate stays above 15% in the first place, and it is why the product sends 20 of these a day at most rather than a hundred of the other kind.

The worst cold messages people have received, pasted in and rewritten the way the product would write them, are at the clinic. None of the examples above is a real person; Priya is invented and so is her post. Everything else is real, and what the real accounts did with it is on the numbers page.