Shared hint budgets need a single priority order across teams

What it is

A dose cap lives on the user’s side as one channel. If several teams each ship bubbles, dots, and update bars, they share that channel’s budget while often counting “we only shipped one this week” separately. Without a unified priority, the budget is spent by whoever shipped first or fights hardest, not by the most useful sentence. This entry is how an organization ranks shared slots. It is not whether a single hint class habituates, and not whether progressive intros need a per-session cap inside one product.

Why it happens

Each team optimizes exposure of its own feature; damage to the channel is on nobody’s books. Locally rational, globally over-fired: growth wants activation, a feature wants adoption, marketing wants a launch, and three “all important” hints arrive the same afternoon. The user lives on one timeline and cannot habituate per team. Unified priority collapses slots into a single queue: one automatic hint at a time, ordered by expected value for this user and this task, not by the release calendar or team politics. Without that key, a pretty budget number still allocates first-come-first-served. Priority must also compare across surfaces—bubbles, bottom bars, in-app mail—as long as they interrupt the same attention, they belong on one table, or “we didn’t use a bubble” becomes an escape.

Where it stops holding

Legal and safety blocks do not enter the teaching-budget ranking; they use their own forced channel, with a different skin so a teaching sequence cannot jump the queue in the name of safety. User-opened help does not spend the automatic budget. Separate products with fully isolated attention (two apps, no overlapping users) may each have a budget; modules inside one shell may not. Mutually exclusive experiment hints still enter the ranking, or the experiment arm fires an extra shot outside the production budget.

Applying it

• Keep a cross-team registry of automatic hints: who ships, to whom, which layer it interrupts, claimed value. Unregistered hints do not ship.
• Issue slots per user session; a single product-owned rule ranks them (for example blocking error > repair of current-task failure > first-relevant core capability > launch promo). Teams may not self-approve a cut.
• Count bubbles, bars, dots, and in-app banners on the same table; a change of surface does not reset the count.
• Verify by listing, for a week of users, which automatic hints fired and which team owned them. Two in one session, or a launch promo ahead of a current-task repair, means ranking is not running.

Related

• Within the group: H2.10.2 High-value hints should displace low-value ones, not stack on them · H2.10.3 Hint fatigue can be read from ignore and dismiss rates · H2.10.4 When the budget is spent, new-feature hints wait in queue
• Adjacent: H2.07 Hint Restraint · H2.03 Progressive onboarding · H5.07 Push frequency
• Search terms: shared hint budget · coaching governance · cross-team priority