A note on this page. It replaces a named list of ten people. That list is not restated, for two reasons: a specific ranking of individuals dates within a year, and republishing an old one would present stale endorsements as current. The tests below are what the list was a proxy for, and they keep working.
Four tests
1. Do they show the work, or only the conclusion?
"Personalisation matters" is a conclusion everyone agrees with and nobody can act on. A teardown of an actual email, with what was changed and what happened, is the work. Follow the second.
2. Do they publish numbers with context?
A reply rate with no sample size, no time period, and no definition of a reply is a marketing claim. Someone who says "on 1,200 sends across six weeks, counted as any human response including negatives" is doing something different. The definitional discipline in the glossary applies here.
3. Do they ever say something did not work?
An account that only reports wins is either extraordinarily lucky or curating. Public failure is the strongest available signal that the successes are real.
4. Do they still do the job, or did they leave it in 2019?
Advice from someone who last carried a quota before deliverability collapsed and before AI drafting arrived is history rather than practice. Both can be valuable; know which you are reading.
What to discount
- Volume of posting. Reach measures posting frequency more than insight.
- Advice with no constraint attached. Real practice always has trade-offs.
Advice that costs nothing usually delivers nothing.
- Anything selling the thing it recommends, without disclosure. Not
disqualifying, and you should know.
- Screenshots with the account cropped out. Unverifiable by design.
Where the useful conversation happens
Increasingly not on public feeds, where the incentive rewards confidence over accuracy. It happens in smaller rooms: practitioner communities, private groups, and calls where people bring problems rather than takes.
The Tenbound community is one of those, and the Tuesday call is built around that format. Whatever you use, the pattern to look for is people sharing what did not work.
Build your own list
Watch for a month. Keep the accounts that passed the four tests, drop the rest, and repeat quarterly. That produces a list that stays current, which no published ranking does.
Where this sits
This is Mastery in the Tenbound Pipeline Architecture Standard. Learning from practitioners is a real part of how reps and managers improve, and it is the part with the least quality control, which is why the filter matters more than the list.