OutreachGenie / Campaign workbench
LinkedIn Campaign Reply Analyzer
A bigger percentage is not the whole result. Map your campaign CSV, set an equal response window and see which records can support a comparison.
01 / Read a local file
One person per record.
The fictional example uses October 2026 dates. Its counts teach data handling, not campaign performance.
A useful result starts with the right inputs.
Start with the assigned people.
A reply-only export cannot measure response rate. Include everyone assigned to each cohort, the actual confirmed send and the first human reply if there was one. Keep failed or missing sends visible. Define your reply categories before looking at the result.
Distinguish no reply from unknown.
Use positive for a first reply that meets your predeclared definition, other for another first human reply, none only after checking the record, and unknown when evidence is missing. A positive or other reply needs a valid reply date and send date. A positive reply later in an ongoing conversation is a separate metric.
Give both cohorts the same time.
Choose the observation policy before collecting results. This analyzer uses days from assignment, so a delayed send can reduce the time available to respond. People with an unfinished window remain pending. Positive replies divided by mature assigned people describe the observed assignment policy; confirmed sends are a separate delivery diagnostic.
Investigate the difference before acting on it.
Review exclusions and unknown outcomes. Then check whether audience selection, sender, timing or follow-up policy differed. Use the A/B-testing guide to design an experiment; this file comparison cannot repair a biased allocation after the fact.
Before you use the result
Does this tell me which A/B variant won?
No. It reports descriptive cohort counts from your data. It does not verify random assignment, equal audiences, delivery quality, independence or adequate sample size, and it does not calculate significance. A higher observed reply rate alone does not establish that a message caused an improvement.
Which date starts the response window?
Assignment date starts the fixed window. A person is mature once assignment date plus the selected number of days is no later than the analysis date. Replies on the end date are included; later replies are shown separately. Dates must use YYYY-MM-DD. Timestamps must first be converted consistently to a chosen reporting timezone.
How are duplicate and missing records handled?
Repeated identities with the same mapped cohort, dates and outcome count once. Conflicting versions are excluded together and reported. Missing identities, malformed records and invalid date order are excluded. Any excluded record suppresses all percentages until the source is corrected. A blank outcome is unknown, never assumed to mean no reply. An unknown mature outcome suppresses that cohort’s percentage.
Does the CSV have to come from OutreachGenie?
No. Use any allowed CSV you can map to the six required fields. This tool does not promise that a particular product export includes assignment dates or reply classifications. You may need to maintain these in a manual test log. Confirm that the supplied data includes everyone assigned, not only successful sends or replies.
Is the file uploaded or saved?
No. Parsing and calculation run in your browser tab. Downloads are generated locally, with formula-like cells protected for spreadsheet review. Clear data or reload removes the source and results. The tool accepts up to 5 MB, 10,000 data records, 64 columns and 400,000 cells, using the same bounded CSV parser as the CSV checker.