How to Read Alumni Participation Metrics

Alumni participation needs a definition before it needs a percentage. A recorded sign-in, an accepted connection and a graduate finding a job answer different questions. This note sets out a proposed method for a future, approved participation report. It is not a report of current Alumnia results or a reconstruction of a historical customer result.

We have removed historical participation figures from our public proof copy because the complete original calculation, cohort rules and exclusions could not be verified from the retained evidence. We will publish measured figures only with a reproducible method and approval for that specific report.

Freeze the group and the reporting period

For the proposed method, take one UTC snapshot of one community. Let N be the unique people who have joined and verified their membership at that snapshot. An imported contact or a pending invitation does not count as a joined member. A person appears once, even if they have duplicate records.

Before calculating anything, record the snapshot timestamp, the community boundary, how verification was established, the deduplication rule and the exact exclusions. The proposed policy includes staff and administrators who are also verified members, and excludes identified test or demo accounts, deceased members, and deleted or removed memberships. If an approved report uses a different policy, state it beside the result. These are proposed rules; they do not establish which filters a historical report used.

Retain an access-controlled source export or query and its version for verification. Do not publish member names, contact details, a customer's identity or absolute cohort sizes in an anonymous public note. If the group is too small to report safely, omit the result. Privacy takes precedence over a percentage.

Recorded sign-ins in the previous 90 days

Let L be the unique people in the frozen cohort with a recorded authentication sign-in in the previous 90 days. Use an inclusive start timestamp and an exclusive snapshot timestamp, both in UTC. The share is 100 × L ÷ N.

This measures recorded authentication sign-ins. It is not daily use, monthly active users, time spent, email opens or attendance at an event. A person with a persistent session may use the product without creating a new authentication sign-in. An application activity timestamp such as last_active_at measures another event and must not be substituted silently for an authentication timestamp such as last_sign_in_at.

A future report must name the actual field and event it uses. If the needed history is unavailable, report that limitation or omit the figure rather than infer a past sign-in from today's state.

Accepted connections at the snapshot

Let E be the unique accepted pairs at the snapshot where both people belong to the frozen cohort. Treat a connection as an undirected pair: A–B and B–A are one pair. Exclude pending and rejected requests, self-connections, duplicate pairs and connections to people outside the cohort.

Let C be the distinct members participating in at least one such pair. The connected-member share is 100 × C ÷ N. The average connections per connected member is 2 × E ÷ C, because each accepted pair touches two people.

These are cumulative accepted connections at the snapshot. They are not necessarily connections created within the sign-in reporting window. Record the status source, snapshot boundary and pair deduplication rule. If N or C is zero, mark the relevant metric not calculated; do not present a missing denominator as zero participation.

Illustrative report — invented values

The following values are invented to explain a report. They are not observed results, current Alumnia benchmarks or a promise to a customer:

• 30% recorded sign-ins: the share of the defined cohort with an authentication sign-in in the stated 90-day window.
• 20% connected members: the share participating in an accepted pair at the snapshot.
• 2.4 connections per connected member: accepted connections averaged only over members who have at least one.

For an approved measured report, publish its real UTC snapshot and window, the cohort definition, exclusions, source field, connection rules and rounding policy beside these labels. Calculate from unrounded counts; round percentages to whole percentage points and averages to one decimal place. Do not attach a real-looking observation date to invented values. Product screenshots containing demo data are also illustrations, not outcome evidence.

Participation is not a career outcome

Recorded sign-ins show one form of access. Accepted connections show that people agreed to connect. Neither proves hiring, mentoring completion, donations or admissions. They also do not show whether a relationship was useful or whether Alumnia caused an outcome.

Evaluate those outcomes separately, with the participants' consent, a defined question and observation period, and a stated response denominator. A single community's participation cannot establish a representative benchmark for other institutions.

Put the definitions to work

Use these distinctions when selecting alumni engagement software or planning a university alumni network. For a mentoring workflow, the alumni mentorship program guide explains why follow-up about the relationship matters beyond a connection count.

Our public product overview links back to this method so readers can distinguish product capabilities, illustrative screens and approved measured evidence. This note describes a proposed reporting method, not a claim that every existing dashboard already implements it.