The problem with a complete post
There is a rule taught in every communication class that nobody applies to launches. A conversation continues because each person leaves the other something to pick up.
Ask a closed question and you get a closed answer. Say something finished and the exchange is over, however good the thing you said was.
Most launch posts are finished. They explain what the product does, why it is better, who it is for, and what to click. Every question the post raises, the post answers. A reader who is genuinely impressed is left with exactly one socially available response, which is a version of "congrats, looks amazing."
That reply is worth close to nothing. It costs the writer no thought, it ends the thread, and it produces no reason for anyone else to join. Politeness and silence rank almost identically, which is why so many teams come away from a well-received launch confused about the numbers.
The mechanism underneath is the one on the algorithm page. Ranking systems weight engagement per impression and treat replies and quotes far more heavily than passive approval, because a reply is evidence somebody had something to add.
A post everybody nods at and nobody answers is telling the system that people saw it and felt nothing worth typing.
So the design question is how many different things a person could reply to.
Six threads, six different people
An open thread is any element deliberately left hanging: something arguable, funny, personal, surprising or incomplete that a particular kind of person will feel compelled to respond to.
The part teams miss is that they should be aimed at different groups. Six jokes for one community is one thread with six versions of the same reply. Six threads pointed at six audiences is six separate conversations happening in the same comment section, each one dragging the post back into circulation.
| Thread | What it is | Who pulls it |
|---|---|---|
| The disagreement | A claim a knowledgeable person could push back on | Practitioners with a view. The largest and most valuable group. |
| The in-joke | Something only people inside one tool, workflow or scene will catch | That scene, commenting partly to show they are in it |
| The number that looks wrong | A figure specific or extreme enough that people ask how you got it | Sceptics and anyone analytical. They arrive to check your working. |
| The human detail | A family member, a bad office, a thing that went wrong on the day | People who know you, and people who react to a person appearing inside a corporate format |
| The visible flaw | Something imperfect you left in on purpose | Correctors. The most reliable reply-generator in existence and the one people are most afraid of. |
| The reference | A cultural nod that rewards recognition | Whoever gets it, commenting to prove they got it |
Only the first of those is an argument. The rest are emotional rather than intellectual, and that is the point most launch planning misses entirely.
People reply because something made them feel recognised, amused, annoyed or curious. Very few people reply because a feature list was well organised.
How many, and aimed where
Four to six threads across a launch asset is the working range. Below three, one audience carries the whole comment section and the post lives or dies on whether that group happened to be paying attention. Above seven, the product stops being the subject.
Two rules on placement.
- Spread them. In a sixty-second video that means roughly one every ten seconds, so somebody who stopped watching at the halfway point still saw two. In a written launch it means one per section rather than a cluster in the opening.
- Map them to real groups before you write. List the audiences you actually want in the replies: your practitioner buyers, the users of the adjacent tool, your existing customers, your own team's networks. Then plant one thread each. Doing it in that order produces different creative from writing the asset and looking for jokes afterwards.
The threads also need somebody there when they get pulled. A conversation opened at 9am and answered at 6pm is a conversation that finished without you, and the early window is where ranking decides how far the post travels.
Budget the first few hours for replying, and get the handful of people who will engage genuinely, your team, your customers, your investors, to be around at the start rather than at lunchtime.
The threads that cost more than they return
This technique has a failure mode and it is worse than a quiet launch, so it is worth naming the four kinds to avoid.
- The manufactured one. A personal detail invented for effect reads as invented, because the audience for technical products has seen a great deal of manufactured authenticity and calibrated against it. Threads have to come from things that are true, which is a constraint on the creative and the reason the good ones feel specific.
- The one you will not stand behind. Planting a disagreement and then going quiet when somebody takes it up is worse than never planting it, because the retreat becomes the story. If you would not defend it at 9pm, cut it.
- The one louder than the product. The measure of success is whether people can name the company afterwards. A thread that gets pulled so hard the product disappears has produced a viral post for somebody else's amusement and nothing for you.
- Antagonism aimed at a named target. Friction generates replies, which is exactly why it is tempting, and pointing it at a competitor or a person converts a launch into a fight you now own. Aim friction at a widely held belief, a category convention or your own past decisions instead. It produces the same energy with a bill you can afford.
The general test before planting anything: if this is the only thing anybody remembers from the launch, is that acceptable? Sometimes the answer is yes and it becomes the best part of the campaign. When the answer is no, you have found a thread that is too strong for the asset it is in.
This is a writing decision
Threads are decided in the script and cost nothing in the edit, which makes this the highest-return unfunded work in a launch.
A polished film with no threads and a rough one with five cost roughly the same to make, and the second one produces a comment section.
This is the same trade as the film question: production value is not what makes a launch travel, and a team that spends the budget on gloss usually arrives at something impressive, complete and closed.
The practical version is a pass over the script before anything is shot. Read it and mark every point where a specific group would have something to say. If you find one, the asset is complete in the bad sense, and the fix is an hour of writing rather than a bigger crew.
Why it sits upstream of the budget
Here is the part that matters commercially, and it is the reason we raise this before anybody signs anything.
Distribution amplifies whatever it is given, with no opinion about quality. Put a closed post in front of two million people and you have bought two million people with nothing to say.
The impressions are real, the report looks fine, and the launch produces a spike and no conversation, which is the outcome teams then try to fix by buying more reach the next time.
That is the loop most failed launches are stuck in. The asset gives people no reason to respond, the numbers disappoint, and the conclusion drawn is that the campaign needed more budget rather than that the creative needed more openings. Paid support gets scaled up to compensate for a problem paid support cannot touch.
The floor on any launch is the creators engaged multiplied by their median reach. Everything above the floor is decided by whether the thing being carried gave anybody a reason to join in.
Narrative design settles the one claim the launch rests on; this is the layer on top of it, and it is where the difference between a spike and a conversation gets made.
Questions we get asked
Why does our launch post only get congratulations?
Because it was complete. A post that explains what you built, why it is good and who it is for has answered every question it raised, so the only socially available reply is congratulations.
That reply costs nothing to write, means nothing, and ends the thread. Silence and politeness look different and rank almost identically.
How many things should a launch asset give people to respond to?
Four to six, aimed at different groups rather than at the same one repeatedly. Six jokes for the same community is one thread. A claim for practitioners, an in-joke for one tool's users, a number sceptics will query and one human detail is four doors into the same conversation, and different people walk through each.
Does this require a bigger production budget?
No, and that is the useful part. Threads are decided in the script rather than bought in the edit, so a launch full of them costs the same as one with none. The expensive version is the opposite mistake: spending on production, ending up with something polished and closed, and then buying reach to push a post nobody has a reason to answer.