Fullbound retweeted
the number from my teardown i trust least is the one that looked best: @boardyai at 50% on the investor-intro prompt.
that came from one run per provider. if chatgpt only reuses about 41% of its cited domains on a same-day rerun, that 50% could be real or it could be one good afternoon.
so before round two, the setup:
same 10 prompts
same 6 providers
same search mode
same location
run again over time
this time i'm counting how often @boardyai gets recommended per prompt across runs, not whether it showed up once.
the other thing i'm watching is the zeros on customer intros and talent discovery. if those hold across runs, that's the actual gap.
Fullbound retweeted
the best reply to my ai visibility teardown was my own advice handed back to me.
i've literally posted "ask the same question again" about chatgpt screenshots.
then @mostlymktgX asked if i'd rerun the prompts. axios found only 41% overlap in chatgpt's cited domains on same-day reruns.
i hadn't. one run per prompt per provider. 60 answers, @boardyai in about 6.
so it's a baseline. next round: same prompts, same search mode, same location, run again over time.
6 out of 60 is not that bad @boardyai
Fullbound retweeted
Sam figures out LinkedIn outbound
Sam writes a message.
“Hey {{FirstName}},
saw you’re scaling {{Company}}.
We help companies like yours generate more qualified meetings.
Open to a quick chat?”
She stares at it.
“This is shit.”
So she personalizes it.
“Hey Rahul — saw you recently hired 3 SDRs.”
Much better.
Rahul gets the message.
Rahul has also received the exact same message from 19 other people.
Sam figures out follow-ups.
“Just bumping this up.”
No reply.
“Wanted to make sure this didn’t get buried.”
No reply.
“Worth exploring?”
No reply.
“Should I close the loop?”
Rahul finally replies:
“Yes.”
Sam celebrates.
“Outbound works.”
She then realizes Rahul meant close the loop.
Sam figures out personalization.
She starts mentioning:
their recent funding
their latest LinkedIn post
their office location
their dog
their mother’s maiden name
Still no reply.
Because apparently knowing someone’s dog’s name doesn’t create demand.
Sam figures out the actual problem:
Nobody gives a shit how personalized your message is.
They care if you understand their problem.
And if your entire pitch can be sent to 4,000 people by changing the first line…
you don’t have outbound.
You have mail merge.
Fullbound retweeted
1/
one linkedin outbound thing we’ve been testing at fullbound:
don’t ask for the meeting in the first message.
instead, give them a small reason to have one.
sounds tiny.
it changes the entire conversation.
🧵








