# A profile visit “from this post” differs from “since this post”

Original constructed exercise. No real accounts, publications, customers, viewers or experimental results. Keep completed real records private.

## Report a complete constructed observation

P1 — The fictional original post shows 12 profile viewers from that post and 4 followers gained from it at the saved observation time. A separate after-post module shows 9 identifiable entries; the weekly account total is 28. Safe report: “At the recorded observation, the post analytics attributed 12 profile viewers and 4 gained followers to this post. Other modules use different scopes.” Every number here is constructed.

Rejected conclusions are “21 different visitors” from adding 12 and 9, “all 28 visitors came from the post,” and “nine named buyers.” There is no cross-module identity map, common observation window or payment evidence supporting them. A ratio using mismatched modules would create another unsupported claim rather than repair the comparison.

## Preserve missing and net-change cases

P2 — A fictional account’s follower total rises from 100 to 106 while one post reports 4 gained followers. Six is the net account change; four is the platform’s post-linked gain. Other gains and losses may exist, and neither number proves a causal increment. Decision: retain both labeled observations, leave the unexplained difference unresolved and do not invent missing acquisition events.

P3 — A fictional owner has only a temporal after-post viewer list; the direct profile-activity field is unavailable. Safe note: “Some profile visits occurred in the displayed period; direct post-attributed activity was unavailable.” Decision: hold the direct-attribution and customer claims. A visible entry does not establish that the person read the post, wanted contact or bought anything.

## Fill and decide

1. Private original-post/account reference and observation timestamp
2. Module label: direct post activity, temporal list or account total
3. Metric/unit, exact source period and value-or-unknown
4. Account tier/filter/privacy availability; unnecessary identities omitted
5. Other module observations kept separate; no person-sum or inferred mapping
6. Net change versus platform-attributed gain; unreconciled difference
7. Descriptive permitted wording, causal/customer claims held without evidence
8. Proceed/hold and next authorized evidence; no contact inferred from a visit

Optional reproduction checks only fixed example data, arithmetic, timezone conversions and file identities. It does not assess completed private notes, permission, identities, language or effectiveness.