Big lists feel productive: lots of rows, lots of potential hits. In practice they often mean less learning per hour — because feedback gets mixed and you can’t tell what actually failed.
I work in small batches: enough contacts to see a pattern, not so many that I lose the plot.
What a batch means to me
A clear slice — same role, similar context, same style of first message. Not “every marketing lead in DACH” but e.g. “marketing leaders at mid-market companies who just announced X.”
Then I send ten to fifteen messages — not fifty. The replies (or silence) belong to one experiment.
Why it improves faster
- You see whether the first line lands — without noise from three other audiences.
- After a batch you can change one thing: hook, context, CTA — not everything at once.
- You stay more human: less copy-paste stress, more attention per contact.
The big-list mistake
When I send fifty at once, I skim profiles before and debrief after — then I repeat the same pattern with more volume. That’s not a test; it’s escalation.
After each batch: a five-line debrief
I write this before I touch the next list — otherwise I forget what I learned:
- Batch size and date
- One-line hypothesis (“first line was too vague about their stack”)
- Replies / no-reply count (rough is fine)
- One thing to try next batch
- One thing not to repeat
Practical: how I start a batch
- Write one ICP sentence (see also One-sentence ICP).
- Curate the list — cut “sort of fits” rather than keep it.
- Use the first message as structure only, not copy: three lines, then personalize.
- After the batch: note reply rate, one hypothesis for the next batch.
When a batch “fails”
Zero replies isn’t a character judgment — it’s a design problem. I change the hook or the list before I change my work hours. Persistence without revision is just volume.