What we measure, and publish

We run outbound for B2B teams and for ourselves. When a result surprises us, we write it down with the method and the limits, so you can argue with it. Two studies so far, and the shorter notes we post as we go.

Studies

1

925 cold emails from Microsoft 365, zero human replies

Same copy, same lists, same five weeks. Our Google Workspace inboxes got 15 people to write back, our Microsoft 365 inboxes got none, with every deliverability check green. What we ruled out, and the free placement test that caught it.

2

Buying signals beat better lists

Eleven list attributes tested across our campaigns, none predictive of a reply. Two things were: the market, and how long a lead waited in the queue. Plus the two campaigns that actually moved deals, both built on a dated deadline.

Both pieces use our own outbound data, measured on campaigns with at least 100 prospects contacted, each prospect counted once and replies counted as distinct people. Sample sizes and biases are stated inside each study rather than hidden behind a headline.

Shorter notes, on LinkedIn

Between two studies, the thinking happens in public on our founder's account, in French.

1% vs 15%, same email
The only difference between the two campaigns was the moment it was sent: a VP Sales hiring reps and switching tools, versus a static list.
Read on LinkedIn →
Most people send one email, then quit
Between the second and the fifth touch is where the replies are, as long as every follow-up carries something new.
Read on LinkedIn →
Do you know your cost per meeting?
Eight founders out of ten cannot answer. We hear the same meeting priced at 90 euros and at 600.
Read on LinkedIn →
Your outbound is not too expensive
You just do not know what it costs, all in: loaded salaries, tools, data and the time actually spent.
Read on LinkedIn →
The biggest problem in cold email is not your copy
It is sending everyone the same number of follow-ups, whatever the lead came from.
Read on LinkedIn →

See the whole feed →

Want outbound built on signals, not lists?

We build the system, run it, and measure what it actually produces.

Book a GTM audit