Across 6,569 deduplicated competitor moves, 1,823 public reactions were measured. The top 100 moves — 1.52% of everything tracked — took 62.5% of them.
Only 520 moves, 7.9% of the total, drew any reaction at all. So the distribution is not merely skewed; most of it is empty, and what remains is dominated by a handful.
How concentrated is attention on competitor launches?
Extremely. In a dataset of 6,569 moves from 648 companies, 1.52% of moves captured 62.5% of all measured public reaction, and 92.1% of moves captured none. Attention follows a power law rather than anything resembling a normal distribution, which means the average is close to meaningless as a summary — mean reactions per move is 0.40, a number that describes almost no actual move.
The practical consequence: median and average tell you nothing here. The only question worth asking about a specific move is which side of the cliff it fell on.
What does that mean for a weekly competitor brief?
It means a brief should be short, and that shortness is a finding rather than an editorial preference. If most of the week's attention sits in a small number of moves, a brief that lists everything detected is mostly listing things that provably did not matter. The concentration is what makes it legitimate to put two or three items at the top and treat the rest as background.
It also means alert volume should be low by default. A tool that notifies on every detected event is, given this distribution, interrupting you overwhelmingly for moves in the 92.1% that drew nothing.
Does concentration mean the quiet moves are unimportant?
No, and this is the part worth being careful about. Reaction measures whether a market responded, not whether something mattered. A significant move that drew nothing is not automatically trivial — it may be trivial, or it may be a real product change that failed to reach its audience. Those are opposite situations with the same reaction count.
Which is exactly why the pairing matters. A high-significance move sitting in the silent 92.1% is the most actionable thing in the dataset: someone built something real and the market did not respond. Concentration tells you where attention went; it does not tell you where value went.
How was this measured?
All 6,569 deduplicated moves across 648 tracked companies, with no top-N truncation — moves were ranked by coupled reaction count and cumulative share taken at 10, 50 and 100. Reactions come from Hacker News, Bluesky, GitHub and news, bound to a specific dated move by exact-phrase and domain-scoped matching. 1,823 reactions total across 520 moves that drew at least one.
The limitation to keep in mind: concentration measured on public surfaces will overstate concentration in markets that talk elsewhere. A sector whose buyers never post publicly contributes almost entirely to the silent 92.1% regardless of what it ships, so the head of this distribution is disproportionately drawn from loud, developer-facing markets. Within a comparable set the shape holds; across very different sectors it should be read with that in mind.
Figures are deduplicated. Our own tracking table held several records for a few companies (Linear had seven), and each record independently re-detected the same changelog entry and re-coupled the same public reactions to its own copy — inflating one company's reaction count by 268% before we caught it. A move is now identified by title, category and date, and reactions are counted by distinct source URL, so the same event and the same reaction each count once.
Revised after a coupling audit. Content matching bound reactions to launches without a causality check, so an old forum thread naming a company could attach to a recent, unrelated release — in the worst case a 2007 Hacker News post counted as reaction to a 2026 changelog entry. 1,099 of 2,885 couplings violated causality (Hacker News worst at 80%, because searching a company name surfaces its whole history) and have been removed; the coupler now rejects them. Every finding above survived the correction and two got stronger — but the landed rates here are lower than first published, and that is the honest direction of the error.