Skip to content

Change a requested window without rewriting earlier evidence

Preserve old and new time scopes when a research question expands or contracts its requested dates.

Guides

Change a requested window without rewriting earlier evidence — what should the operator record?

A request for last week and a request for the last month do not have interchangeable denominators. Even when their result tables overlap, the scope change must remain visible. Record the intended windows and actual timestamp coverage before treating either response as the baseline for later operational work.

Practical operations check
Preserve old and new time scopes when a research question expands or contracts its requested dates.
Store previous_start, previous_end, revised_start and revised_end using an explicit timezone; preserve any uncertain source date precision separately. A local contract might use an inclusive calendar start and exclusive end; express that choice explicitly before comparing anything to offset-bearing publication timestamps.
Use the common eligible interval for an operational comparison and state remaining retrieval differences rather than deriving population growth from returned row counts. If a broader request returns fewer rows, retain that observed result and its coverage notes rather than assuming the transformation lost records or the population shrank.
01 / Guides

Record both intervals

Store previous_start, previous_end, revised_start and revised_end using an explicit timezone; preserve any uncertain source date precision separately. A local contract might use an inclusive calendar start and exclusive end; express that choice explicitly before comparing anything to offset-bearing publication timestamps.

02 / Guides

Name the direction

Label the revision expansion, contraction or shift rather than calling every date edit a refresh of identical research scope. When shifting a window rather than expanding it, neither returned set is necessarily a superset; represent both directions of membership change in the diff.

03 / Guides

Partition the responses

Separate returned rows inside the shared interval from those only eligible under the revised interval; keep missing timestamps in an unresolved bucket. Rows lacking usable dates belong in unknown-time, not automatically inside the common interval just because their other fields resemble eligible evidence from that period.

04 / Guides

Check boundary examples

Inspect records exactly at a start or end boundary using the declared inclusion convention, not the spreadsheet application's display timezone. A date-only record on the boundary may not support an instant-level decision; preserve its lower precision instead of converting it silently to local midnight.

05 / Guides

Choose the comparison scope

Use the common eligible interval for an operational comparison and state remaining retrieval differences rather than deriving population growth from returned row counts. If a broader request returns fewer rows, retain that observed result and its coverage notes rather than assuming the transformation lost records or the population shrank.

Keep the boundaries clear.

  • Preserve old and new time scopes when a research question expands or contracts its requested dates.
  • Operational record: Record both intervals. Context badges do not certify access.

Start with a read-only check: ask your agent to report actual source coverage before collecting records or exporting an audience.

Primary references

Documentation reviewed 2026-11-15. Product capabilities, client availability and plans can change. Third-party marks identify their products, not an affiliation.

Know what social media shows about you.
Start with a question.

Get started
montage