Social cost drives people to abandon voice in public

What it is

At an open-plan desk, a calendar clash, a one-line reply, a meeting in ten minutes are all faster by voice. People type anyway. Public abandonment of voice is a channel choice made before the mouth opens. Product logs then show “weekday voice sessions near zero” and misread it as “nobody wants this skill.” The same people use it at home. In earshot of others they change channel, not need.

Why it happens

Abandonment is a decision, not a recognition failure. Before speaking, people add content exposure to the mark of talking into the air, and compare that sum with “take out the phone and type.” Open-plan norms treat sound as an occupation of the next desk, so the comparison almost always picks the silent channel. Abandonment then feeds itself: public settings never train the phrases or the habit, and voice stays “the thing I do at home,” never entering muscle memory at work. Capabilities that exist only on voice (hands-free scheduling, a reply while walking) are absent at the desk. Household penetration therefore does not extrapolate to “voice-first all day.”

Studying it

Diaries should log intended channel versus actual channel: the moment voice was considered, how many people were present, and whether the person spoke, typed, or dropped the task. The dependent measure is switch rate, not completion — completion drops everyone who never opened a session. On the log side, coarse-split sessions by place (home Wi-Fi / work network / cellular) and look at wake density during work hours; add a counterfactual (“would you have used voice if no one were here”) to pull abandonment apart from “no demand for the feature.” Labs that only measure task success in an empty room will measure abandonment as zero.

Where it stops holding

Driving, full hands, or no screen to type on can push occupancy back over social cost; people may still speak, unwillingly. Talk-as-work settings (warehouse, field radio) recode voice as a legitimate tool and abandonment does not apply. A truly private office or exam room drops the public premise. Writing “they don’t use it at work” as user folly misses that the norm is doing the work. An empty office at night is not the same place as a full bench at noon.

Applying it

• Inventory a skill’s public survival: in an open-plan walkthrough, observers record whether anyone spoke, not whether recognition was accurate. Skills with zero openings should not be marketed as office voice-primary.
• Give desk-frequent goals a silent peer path, and do not scold “please speak up” when voice is not woken.
• Split growth metrics by place. Home activity must not count as workplace adoption. Workplace adoption needs its own bar, and the honest answer may be that voice should not carry those goals there.
• How to check: in a two-week diary, if “thought of voice, typed instead” stably outruns actual openings, the product’s public-voice story is already detached from real channel choice.

Related

• Same group: M4.01.1 Speaking to a dialogue system hands the content to everyone in earshot · M4.01.2 Addressing a device in public is still socially marked · M4.01.4 Bystanders pay a cost for being forced to overhear · M4.01.5 Spoken replies leak more than spoken commands · M4.01.6 Headphones privatize the reply, not the request
• Nearby: M1.01 When voice-first is appropriate · C7.17 Switching between voice and keyboard · M3.05 Voice and screen complementarity
• Search terms: public abandonment of voice · channel switch · social cost of voice