On this page
Key Stats
Key Takeaways
- Retain messages that record a decision, a promise, or a material term.
- Treat scheduling confirmations as temporary unless they support an open question.
- Document the date, folder, and rule before deleting a batch.
- Keep a second copy of materially important messages outside the inbox.
Question and scope
This review asks a narrow operational question: which recruiting email messages need a retention window, and which can be deleted when their task closes? The unit is one message thread tied to one athlete and one program. The review does not attempt to interpret laws for a specific family or school. It looks only at whether a written retention rule can be applied consistently, and it builds on the transfer timing evidence review without repeating its handoff measure.
The scope is fixed before any message is read so the result cannot be steered by a dramatic example. Eligible threads must contain at least one decision, promise, or material term, and at least one later action by the family. A thread with no action is excluded because retention has no operational consequence. A thread with no identifiable program is excluded because it cannot be linked to a recruiting record.
The family file is the object of study because the family often keeps the only continuous copy of a conversation. A program may change staff or systems, and the inbox may be the only place a promise survives. The review therefore treats the family's dated record as the primary evidence and the program's later confirmation as a check. That order prevents a reconstructed memory from replacing the original message.
Population and unit
The population is the set of recruiting email threads a family or support assistant retained during one recruiting cycle. Each thread must name a program, a coach or office, and a topic. Threads that contain only an automated receipt are excluded because they carry no human decision. The same athlete may contribute many threads if the threads belong to different programs or different decisions.
The unit stays the thread, not the message, because a decision often unfolds across replies. A thread that is partly deleted after a review remains in the set under its original key. That rule prevents a cleaned thread from disappearing and making the archive look more complete than it was. Every removal from the set carries a written reason and the date of the decision.
Definitions and measures
Evidence is a message that records a decision, a promise, a material term, or a compliance question. Logistics is a message that only moves a calendar item or confirms a routine arrangement. A thread can contain both, which is why the unit is the thread and not the individual message. The definition focuses on the action a reader could take, not on the tone or length of the message.
The primary measure is the number of days between the close of a recruiting task and the deletion of its thread. A second measure records whether a materially important message has a copy outside the inbox. Both measures are reported with the eligible count and the number of threads that could not be classified. A percentage without those counts would hide the shape of the sample.
Evidence standard
Each thread is tagged durable, temporary, or unresolved using the message that was open at the time of review. A durable thread has a decision or a material term that a later reader might need. A temporary thread has only logistics and no open question. An unresolved thread has a possible decision but no clear record of the outcome. The tag describes the thread, not the family's confidence.
When two readers disagree, both tags stay in the file until a later message or a written clarification settles the question. A verbal explanation given after the review is recorded separately and cannot retroactively change the tag. The reviewer records the thread, the retrieval date, and the specific line that supports the tag. An empty subject line is never treated as proof that a thread is unimportant.
A retention decision must be reproducible from the record. The reviewer notes the rule applied, the date, and the person who approved the deletion. This is the same discipline used in the family document retention plan, applied to a narrower unit. A deletion without a note is treated as a gap rather than a completed action.
Collection
The sample is drawn from threads that entered an active recruiting file during the declared window. Collection uses a fixed interval and a written stopping point so the reviewer cannot keep adding threads until a pattern appears. Threads are included whether or not the family considered them important, because a routine message can become evidence when a question opens. The note states how many threads were eligible, how many were read, and where reading stopped.
Late additions are refused after the results become visible. A thread added at that point would be selected with knowledge of the outcome and would weaken the comparison. If a thread is discovered after the stop, it is logged as out of scope and kept for the next window. The reviewer also records any program that sent only automated messages, because that absence is itself an operational finding.
Every program with no durable thread is listed in the collection note even when no deletion was tested. That list shows whether the sample is limited by the family's filing or by the program's communication. A reviewer who cannot find a durable thread should not silently drop the program from the frame. Listing the gap keeps the denominator visible and gives the next cycle a concrete place to start.
Retention test
The retention test asks three questions of each thread. First, would a later reader need this message to explain a decision? Second, is the message available somewhere other than the inbox? Third, does an open question still depend on it? A thread that answers yes to the first or third question stays, and a thread that answers no to all three can follow the temporary rule.
The test is applied before the deletion, not after. Applying it afterward invites the reviewer to justify a choice already made. The order matters because a deleted thread cannot be re-read without a backup. The reviewer records the answers and the date so a second reader can see the sequence.
A thread that fails the test is moved to the temporary bucket rather than deleted immediately. The family then has one review cycle to notice a missed dependency before the message disappears. The bucket is reported with the main result instead of being hidden in an appendix. That delay is a control, not indecision.
Privacy and access
Study identifiers replace athlete names in the working set, and the crosswalk stays inside the approved recruiting system. Only the fields needed to test the retention rule are opened, and temporary access is removed when the review ends. Messages from coaches are summarized rather than copied into the report. The report lists what was withheld or unavailable instead of quietly leaving it out.
Family privacy obligations shape the handling of the threads, but they do not make the evidence reliable on their own. A thread can be private and still be incomplete. The review separates the access question from the accuracy question so one does not substitute for the other. That separation is especially important when a family shares a file with a new assistant.
Limits and next step
The conclusion is tied to the cycle, the mix of programs, and the retention rules the family already used. A shorter window or a set of threads concentrated in one sport may produce a different result. Re-running the review after the next recruiting season shows whether the pattern holds. Where the second run differs, the report shows both outcomes and names the operational decision that depends on the choice.
The output supports one process fix and a scheduled retest, not a legal opinion about any specific record. A reasonable next step is to define the durable, temporary, and unresolved buckets in writing before the next cleanup. A second step is to save materially important messages outside the inbox. Neither step requires a new platform, and both can be checked without collecting more personal data.
A conclusion would change if the review window overlapped a major staff change or if most threads came from one program. A larger sample with more programs would test whether the pattern is local or general. A second reviewer with less recruiting experience would test whether the definitions depend on insider knowledge. The report names those conditions instead of treating the current result as final.