Original practice scenario: A hypothetical research-operations team tracks sites from initial contact through document review and activation. Many sites are still in progress. Define a defensible funnel report.
Read model answer +
Preserve site cohorts and stage history instead of comparing unrelated monthly stage counts. A site still awaiting review should not automatically be classified as a failed activation.
Work through it
Identify each site and onboarding episode, recording stage-entry times, reversals and closure reasons while retaining the original event history.
Build one row per episode with ordered milestone timestamps, validating impossible sequences and handling repeated document reviews without duplicating the episode.
Report stage conversion and elapsed-time measures for comparable entry cohorts, marking incomplete follow-up and separating withdrawals from ongoing cases.
Tradeoffs to discuss
Waiting for mature cohorts improves comparison but delays feedback; an operational work-in-progress view remains useful when clearly distinguished.
Common pitfalls
Do not divide activations this month by contacts this month as though they were the same population.
Reactivated sites may require a new episode rather than overwriting the outcome of the original one.
Original practice scenario: A hypothetical research-literature dashboard groups papers about a therapeutic area. Conference abstracts, later articles and follow-up analyses may describe the same study. Design evidence counting.
Read model answer +
Distinguish publication identity from study identity and analysis identity. Several documents can contribute useful detail while still depending on the same underlying participants or experiment.
Work through it
Retain bibliographic identifiers and study references, proposing publication-family links from explicit trial identifiers and reviewed contextual evidence.
Record which population, time point and analysis each document covers, preserving uncertain relationships rather than merging solely on similar titles.
Produce both document counts and independently defined study counts, exposing linked reports and overlap assumptions for reviewers assessing the evidence.
Tradeoffs to discuss
Conservative grouping preserves distinctions but may overcount related reports; aggressive grouping can collapse genuinely separate studies.
Common pitfalls
Do not treat a follow-up analysis as an independent replication without checking its underlying population.
This data organization supports research review and does not by itself establish treatment benefit or clinical recommendations.
