Use invite access to widen testing coverage

Invite access can help a trading-agent team hear from a broader range of users if allocation begins with missing experiences. A community drawn entirely from one device, technical background or trading habit may repeatedly find the same problems. We propose a coverage ledger that identifies those gaps without turning user diversity into a points competition.

DXAP's dated Chapter 0 terms describe specific allocation constraints. At the weekly review, signed-up accounts with zero usable invites receive two through a reviewed grant. During the second-week review, at least two new-user redemptions from the current allocation qualify it for restoration to its full recorded size after team verification. Shared but unused codes do not satisfy that condition. These are one-time reviewed mechanisms within the published chapter, whose bonus window ends October 6, 2026 at 00:00 UTC.

Those terms constrain what an organizer can promise. A research invitation could explain the experience being sought and offer information about requesting access. It cannot create an allocation, promise a refill or treat a recruitment spreadsheet as a distribution receipt.

An illustrative coverage problem

Suppose twelve volunteers already cover desktop setup well, while only one has attempted the same task on mobile and none has tested an accessible low-motion workflow. The next useful recruitment question is whether the team can hear from consenting participants in those missing situations. Selecting the loudest referrers would leave the coverage gap unchanged.

A proposed ledger would record the task a volunteer agreed to test, relevant device context and participation status. Avoid collecting sensitive attributes when a functional testing category suffices. “Uses screen magnification” is relevant to a readability task; an unrelated personal profile adds little diagnostic value. Participants should be able to decline a category or withdraw before testing.

Within an equally suitable group, an explicit allocation rule could use a recorded draw or first confirmed availability, subject to the actual authorized invite allowance. The team would publish the selection rationale and constraints, rather than imply that every interested person receives access. A planned invite remains unissued until the DXAP team confirms its grant.

Measure gaps closed, with the missing cases visible

The controls paper companion motivates collecting failures that can be traced to a stage in the harness. A coverage study should therefore count completed task observations and distinct reproducible issues by testing situation. An invite redemption alone does not show that a mobile setup was attempted. A completed attempt that ends in confusion can be more useful than a successful setup with no recorded observations.

We would retain refusals, unredeemed access and abandoned tasks in the coverage report. Their causes may reveal a scheduling problem or an intimidating invitation rather than product behavior. Aggregate summaries can show how many planned situations remain unobserved while keeping individual identities private.

The result would guide the next authorized allocation review: which experience remains missing, what task needs testing and how much support that test requires. This gives invite access a research purpose while preserving the published rules and the DXAP team's review.