Skip to content
ICON

Example simulation

A full report, start to finish.

This is a worked example ICON ships with, so you can read a complete output without spending a run. It is a simulation of a real post — not measured platform data.

Try your own

The content analysed

I spent six months building a feature users understood in six seconds. Not because they're impatient. Because six seconds is all any of us give anything now. We rebuilt it three times. The version that worked was the one with the least in it.

  • X
  • Engagement
  • Founders

ICON Score

0Strong

The opening line is doing almost all the work, and it is doing it well. The drop-off risk is the middle: line two states something your audience already believes, which spends attention without adding anything.

Confidence reflects how much context ICON had, not how accurate the outcome will be. Clear platform and goal, but no detail on what the feature was or who follows you, so segment sizing is broad.Clear platform and goal, but no detail on what the feature was or who follows you, so segment sizing is broad.
AI-simulated judgement of how strongly the opening earns the next line. Not observed platform data.

88

AI-simulated read on how quickly the point becomes understandable. Not observed platform data.

82

AI-simulated estimate of felt reaction — recognition, tension, delight. Not observed platform data.

71

AI-simulated estimate of whether someone would pass this on. Not observed platform data.

74

AI-simulated estimate of intent to click, reply or buy. Not observed platform data.

41

ICON models possible audience reactions using AI. Scores and responses are simulations, not guaranteed outcomes or observed social-platform data.

Reaction map

How people may react

Simulated distribution across 100 modelled readers
  • Strong positive28%
  • Curious31%
  • Neutral19%
  • Skeptical17%
  • Negative5%

Most likely first impression

Good line. I want to know what the feature was before I decide whether this is insight or just a nice sentence.

Simulated audience feedback, not observed platform data.

Synthetic responses

The comment section before it exists

13 simulated replies · audience archetypes, not real people
  • BuilderPositive

    The last line is the whole post. Least in it wins, every time.

  • SkepticSkeptical

    six months though. was that learning or was that scope creep with a nice ending

  • Design leadPositive

    Rebuilding three times is the part nobody puts in the case study. Respect for saying it.

  • Curious readerCurious

    What was the feature? Hard to judge the lesson without knowing what got cut.

  • EngineerCritical

    Users understanding something in six seconds isn't the same as it being good. Plenty of shallow things are instantly legible.

  • FounderHigh intent

    Do you have a writeup of the three versions somewhere? Would send this to my team.

  • Passive scrollerNeutral

    true

  • MarketerCurious

    Is the six seconds from actual session data or is it a figure of speech

  • Confused readerConfused

    wait did the feature take six seconds to use or six seconds to understand, those are different things

  • Design leadPositive

    'The version that worked was the one with the least in it' — putting this on a wall.

  • SkepticSkeptical

    Every post like this ends with 'and then we removed things'. Would love to see one that ends differently.

  • BuilderPositive

    Currently on rebuild two. Needed to read this today.

  • EngineerCurious

    What made you stop at three? Genuine question — how did you know that one was it?

Audience clusters

Who responds to what

Share values are simulated proportions of the modelled audience

Builders mid-project

34%
Likely interpretation
Reads it as permission to cut scope on something they are already privately worried about.
What they care about
Whether shipping less is defensible to their team, not whether it is true.
Potential friction
Nothing here tells them how you decided what to remove.
Strongest angle
Name the thing you cut that hurt most. That is the line they will screenshot.

Design-minded operators

26%
Likely interpretation
Hears a craft argument and agrees immediately — 'six seconds' matches how they already talk about attention.
What they care about
Evidence that restraint was a decision rather than a compromise.
Potential friction
'Rebuilt it three times' reads as process, not as judgement.
Strongest angle
Show what version two got wrong. Failure specifics travel further than the conclusion.

Skeptical engineers

22%
Likely interpretation
Suspects six months of work is being retold as a tidy lesson after the fact.
What they care about
Whether the six months were necessary or whether this is a story about waste.
Potential friction
No cost is acknowledged. Six months is a long time to arrive at 'less'.
Strongest angle
Admit the first four months were the price of learning it. That earns the rest.

Passive scrollers

18%
Likely interpretation
Registers the first sentence, agrees vaguely, keeps moving.
What they care about
Being entertained or reassured in under two seconds.
Potential friction
Line two asks them to nod at something they already accept.
Strongest angle
Replace the middle beat with the concrete detail — specificity holds a scroll better than a truism.

Surface vs subtext

What they say / what they may be thinking

  • Says

    Great insight.

    Thinks

    I agree with the conclusion but I have not learned anything I could act on tomorrow.

  • Says

    Six months is real dedication.

    Thinks

    Six months for one feature would get me in trouble. Was this actually well spent?

  • Says

    Love the simplicity angle.

    Thinks

    Everyone says simplify. Almost nobody shows the thing they were scared to delete.

  • Says

    What was the feature?

    Thinks

    I suspect the specifics are less impressive than the framing, and I'd like to check.

Friction scanner

Where you lose people

3 issues, most damaging first
  • Buried payoff

    high

    The strongest sentence — 'the version that worked was the one with the least in it' — sits at the very bottom.

    On X, a large share of readers never reach the third beat. The best line is being read by the fewest people.

    Fix. Cut the middle paragraph and let the closer land second. The post gets shorter and hits sooner.

  • Missing proof

    medium

    'Six seconds' is stated as fact with no source and no story behind it.

    Skeptical readers stop at the first unsupported number, and that segment is a fifth of the simulated audience.

    Fix. Either attribute it ('watching seven user sessions back') or drop the precision and say 'seconds'.

  • Generic middle

    medium

    'Because six seconds is all any of us give anything now' restates something the audience already believes.

    It spends a line without moving the reader, which is where scroll-away happens.

    Fix. Replace it with the one thing you deleted that you did not want to delete.

Misinterpretation scanner

How this could be misunderstood

  • Sounds like a humblebrag

    Framed as six months of your effort resolving into a neat lesson, some readers hear self-congratulation rather than a lesson learned.

    Who reads it this way: Skeptical engineers and anyone who has shipped under deadline pressure.

    Defuse. Put the cost in the post. One clause admitting what the first four months bought you removes the read entirely.

  • Reads as anti-depth

    'Least in it' can be heard as an argument that simple always beats thorough, which is not what you mean.

    Who reads it this way: Engineers working on genuinely complex surfaces.

    Defuse. Specify that you removed choices, not capability.

Rewrite engine

Make it land better

Sharper

Clearer, less drag — same voice

I spent six months building a feature users understood in six seconds. The version that worked was the one with the least in it — and the thing I cut last was the thing I'd been most proud of.

Removed the middle truism and pulled the payoff up, so the strongest line arrives while people are still reading.

193 characters

Bolder

More opinionated, higher risk

Four of the six months I spent on this feature were wasted. I was building for a user who reads carefully. That user does not exist. The version that shipped is the one with the least in it, and it took me three rebuilds to stop arguing with that.

Opens with the cost instead of the accomplishment, which disarms the skeptics and takes a position people can disagree with.

248 characters

Cleaner

More premium, more controlled

Six months on one feature. Users understood it in six seconds. Three rebuilds. The one that worked had the least in it.

Stripped to declarative fragments. Same argument, no connective tissue, reads as confidence rather than explanation.

120 characters

Hook lab

Alternative hooks

Each uses a different mechanism
  • Contrast

    Six months of work. Six seconds of understanding. That ratio is the whole job.

  • Tension

    The feature I was proudest of is the one I had to delete.

  • Curiosity

    We rebuilt the same feature three times. Only the third one was honest.

  • Authority

    After three rebuilds I can tell you exactly what users do with a new feature in the first six seconds.

  • Direct benefit

    If you are on rebuild two right now: the version with less in it is probably the one.

The one change worth making

Delete the middle paragraph and promote the closing line to position two. That single move raises the odds that the sentence people would quote is the sentence they actually see.

Now run yours.

Free during beta. No account required.

Run a free simulation