Press moves — funding announcements, launches with a comms push, partnership news — drew a public reaction 14.8% of the time. Actual product ships drew one 7.9% of the time. That is an 87% higher hit rate for announcing something than for building something.
The obvious conclusion is the wrong one, so it is worth walking through carefully.
Do product launches get less attention than announcements?
Yes, measurably. In a dataset of 6,569 scored competitor moves from 648 companies, 2,187 press moves produced 323 with a coupled public reaction (14.8%), while 2,490 shipping moves produced 197 (7.9%). Average reactions per move follow the same ordering: 0.47 for press, 0.34 for ship.
This is not surprising once stated plainly. Press moves are engineered for reaction — timed, pitched, seeded with journalists and amplified by the people involved. Shipping is a side effect of building; a changelog entry has no distribution behind it. You are comparing something designed to be noticed against something that merely happened.
Does that mean a company should do more PR?
Not from this data, no. The measurement here is attention, not outcome. A press move that earns 40 reactions and a shipped feature that earns 4 are not ranked against each other on business value — nothing in this dataset knows about revenue, retention or whether anyone adopted the thing.
What the gap actually establishes is a correction factor for reading a rival. If you compare a competitor's press hit to their shipping activity using raw attention, you will systematically overrate the announcement. The right comparison is within category: is this ship loud for a ship, is this raise loud for a raise.
What is the useful signal in the gap?
A competitor whose press consistently lands while their shipping consistently does not is telling you something specific: they are better at narrative than at product, and their attention is bought rather than earned. That is a genuinely different competitor to face than one whose quiet releases get picked up organically.
The inverse is the more dangerous rival. Low press hit rate, high ship hit rate means people react to the software itself with no distribution behind it. That is the profile that compounds, and a recency-ordered feed will make them look quiet right up until they are not.
How was this measured?
Every move is a diff between consecutive observations of a public surface, categorised at detection time as ship, press, hire or open source. Reactions come from Hacker News, Bluesky, GitHub and news, bound to a specific dated move by exact-phrase and domain-scoped matching. A move 'landed' at one or more coupled reactions. Categories under 20 moves were excluded.
Two limitations worth stating. Category assignment is deterministic but imperfect at the boundary — a launch blog post with a comms push can read as either ship or press, and those cases sit closer to the middle than the headline gap suggests. And this counts public reaction only, so any market that discusses itself somewhere we cannot read is under-represented in both categories, though there is no reason to think it biases one more than the other.
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.