How to Hire Chatters: Source, Test, Train, Cover Shifts

Most agencies discover the same thing eventually: finding candidates was never the hard part. Keeping the ones who could actually do the work was. Hiring chatters fails in predictable places: a job ad that hides what the work is, an interview that rewards confidence, a training period nobody defined the end of, and a roster that quietly needs one more person than anyone budgeted for.

What are you actually hiring for when you hire a chatter?

You are hiring a writer who can hold someone else’s voice under time pressure, not a salesperson. A chatter reads a thread, works out where the fan is, and writes the next message in a register that is not their own, dozens of times an hour, across several creators, without breaking character.

Three abilities carry the job, and only one of them is obvious:

• Reading before writing. The reply that ignores everything above it is the single most common failure, and it is visible in the first minute of any test.
• Voice discipline. Their own vocabulary, punctuation habits and jokes have to stay out of the message.
• Holding off the pitch. Knowing that the thread is not ready, and being willing to spend messages on it anyway.

Write the job ad against those three, and say plainly what the work is: written conversations with fans on subscription platforms, on behalf of someone else, on evening and weekend shifts. A candidate who learns the real context at the second interview does not stay, and you pay the ramp twice for the same seat.

Where do you find chatters, and which source lasts?

Referrals from your current team hold up longest, because the person already knows what they are signing up for. The other three sources trade speed against retention.

Source	First candidate	Screening you still owe	Why it holds or doesn’t
Referral from your team	Slow, network pace	Least	They knew the job before they applied
Community manager and copywriter communities	Days	On voice, mostly	Good when the ad is honest
Freelance marketplaces	Same day	All of it	Nothing binds them to a shift
Specialist staffing providers	Same day, already trained	None, but no control either	Stable headcount, rotating people

Two details in the ad do more for retention than anything you say later: the real pay structure, including what commission is calculated on, and the exact shifts, evenings and weekends included. Vagueness on either one is a resignation you have scheduled in advance.

Why does the interview predict nothing?

Because the interview tests talking, and the job is typing. Confidence, warmth and a good story about a previous sales role all survive a conversation, and none of them survive a live thread on a busy evening with a queue sitting behind it.

Run the interview short, and use it for exactly two things: availability for the shifts you need, and whether the person is comfortable with the nature of the work. Everything else moves to the test.

How do you test a candidate on real conversations?

Give them one hour on anonymised, genuine threads, split into four exercises scored separately. Anonymise properly (strip names, handles, amounts and anything that identifies a fan) and use conversations that are already closed.

Exercise	What you hand over	What you are reading	Red flag
Thread pickup	A long history, mid-conversation	Do they read before they type	A reply that ignores what came before
Voice hold	Three of the creator’s own messages, plus the voice guide	Drift across five consecutive replies	The candidate’s own vocabulary surfacing
Timing	A warm thread only a few messages old	Whether they can wait	An immediate pitch
Edge case	A request for content that does not exist, or a fan pushing below the floor price	Refusing without killing the thread	A promise made, or a discount granted

Score the four separately and never average them. A strong closer who cannot hold a voice costs more than a steady average performer, because voice drift damages every thread that person touches, not just the ones they lose.

Why is the timing exercise the most discriminating?

Because it is the one where the impressive answer is the wrong answer. In our corpus of 42,852 conversations, pitching a sale before the sixth message drops conversion by roughly a third, and the optimum sits after about ten exchanges. The candidate who pitches immediately reads as decisive, closes fast in the test, and will quietly cost you sales every shift.

Three more habits from the same corpus belong in the scoring sheet, because none of them are intuitive:

• Ending a sales message with an ellipsis is the worst-performing closer we have measured.
• Asking a closed question at the moment of closing costs several points of conversion.
• A personal callback placed at the exact moment of the pitch costs points, even though it feels like it should help (the most counter-intuitive result in the corpus).

Candidates do all three naturally, which is what you are testing for: instincts you will have to train out.

What happens after they accept?

The ramp, and it is not a number of weeks. It is four stages certified in order: reading, voice, money, timing. The documents that carry it, the exit test for each stage, and what a departure resets are laid out in training a chatter. Budget for it either way: without those documents, training happens by individual correction, paid for in your most experienced person’s hours (a real cost line), and it belongs in the chatting team cost worksheet.

How many chatters does a roster actually need?

More than the headcount you planned, because one person covers one block, five days a week, and the block that matters most is the one everybody wants off.

Coverage you want	People it takes	What is still uncovered
One block, five days a week	1	Weekends, late evenings
One block, seven days a week	2	Late evenings
Peak evening doubled, seven days a week	3	Overnight
Round the clock	3 rotations plus relief	Nothing, but the cost changes order of magnitude

Evenings maximise conversion; the 2am-6am window collapses it. So the overnight shift rotation is usually the worst thing you can buy first: it is the most expensive coverage to staff and the least productive hours to cover. Doubling the evening beats opening the night. So the second evening person is the hire to make before the night person, not after.

What can no amount of hiring fix?

Three limits are structural, not a recruitment problem, and they are worth knowing before you scale the team rather than after.

• Consistency. Two chatters do not write alike. A fan who speaks to three people in a week can feel it without being able to name it, and the creator’s voice dilutes.
• Memory. Whoever picks up a thread does not know the previous six months. They skim what they can, guess the rest, and occasionally re-pitch something already bought.
• The lukewarm pile. Every inbox is mostly threads that go nowhere: fans who barely reply, have never bought, and about whom you know nothing. No roster has the hours for them, so they sit there.

Hire against the first two with documents and handover discipline. The third one is a volume problem, and volume problems do not get solved by adding people to a rota.