Know which student is slipping in week three, not at the exam board.
Attendance, assessment, engagement and finance signals become one explainable risk score, then an alert with an owner and a due date. Every intervention is recorded, so you can prove which ones actually worked.
The problem
Every institution knows who withdrew. Almost none can say who was about to.
The signals arrive on time and separately. Attendance sits in a register, marks in a gradebook, engagement in the learning environment, an unpaid instalment in finance, and a tutor's quiet concern in nobody's system at all. Each one is unremarkable alone. Together they describe a student who will not be here next term, and no single person is looking at all five.
Three things go wrong every term. Risk is discovered late, when the withdrawal form is already signed. Concern is raised without an owner, so a flagged student is everyone's worry and nobody's task. And intervention is never measured, so an institution repeats an outreach programme for years without knowing whether it changed a single outcome.
The operating model
One student record, and the signals reach a person.
Four design choices turn retention from an annual report into an operating layer: signals are read live rather than uploaded, every score explains itself, every alert has an owner, and every intervention is measured against what happened next.
Signals are read, not uploaded
Attendance, assessment, engagement, registration and finance are already on the same model, so risk reflects the register marked this morning rather than a file that landed overnight.
A score that cannot explain itself is not used
Every risk figure opens onto its contributing signals, their weights, the threshold crossed and the model version in force. An advisor sees why a student surfaced before deciding what to do about it.
An alert is a task, not a notification
Raised against a named student, routed to a named owner, with a due date and a state. It closes when an action is recorded, not when someone reads it.
Intervention is measured, not assumed
Because the alert, the action and the subsequent attendance, marks and progression sit on one record, the institution can compare outcomes for students who were reached against comparable students who were not.
What it does
Five things a success team needs. One record.
Signals and risk, early alerts, advising and case management, support and referrals, and the analytics that show what worked. Pick one to see what sits inside it.
Click any block to explore
Live inputs, one score, and the reason it moved.
- Attendance decline against course and programme thresholds
- Missed, late and failed assessments, and grade trajectory
- Engagement with the learning environment and campus services
- Outstanding fees, instalment default and financial holds
- Prior-term performance, repeat attempts and credit shortfall
- Institution-owned weighting model, versioned with effective dates
- Factor-level explanation on every score, no black box
Raised, routed, owned, and closed with an outcome.
- Threshold and rule-based alerts per programme and cohort
- Manual flags raised by tutors and lecturers from the class list
- Automatic routing to the responsible advisor, tutor or service
- Severity, due date and escalation when an alert ages
- Alert states from open to acknowledged, actioned and closed
- Student and guardian notification templates with delivery tracking
- Bulk actions for cohort-wide interventions
The advisor's day, on the record rather than in an inbox.
- Advisor caseloads with assignment by programme, cohort or campus
- Appointment scheduling, availability and attendance of the meeting
- Case notes with confidentiality levels and access control
- Success plans with agreed actions, owners and review dates
- Academic standing, probation and progression review workflows
- Full interaction history across advisors and terms
- Student view of their own plan, appointments and next steps
A referral that is tracked, not an email that is hoped for.
- Referrals to counselling, learning support, disability services and finance
- Service acknowledgement, appointment and closure recorded
- Consent and confidentiality handling per service and per note
- Accommodation records carried into teaching and examinations
- Peer mentoring, tutoring and workshop enrolment
- Hardship and hold resolution linked to the finance record
- Service load and waiting-time reporting
Whether it worked, not whether it happened.
- Retention and progression rates by cohort, programme and campus
- Intervention outcomes against a comparable non-intervened group
- Alert volume, response time and closure rate by advisor and service
- Attrition drivers ranked by contribution, term on term
- Early-warning accuracy and model performance review
- Board and regulator reporting from live records
- Evidence of student support for quality review and accreditation
Inside the product
The whole success cycle, in one place.
This is the application, not a diagram. Move through the surfaces your advisors and success team work in: the risk board, a student's risk detail, the alert queue, the advising session, referrals, and retention analytics.
| Student | Trigger | Owner and due date | State |
|---|---|---|---|
| Zahra Khatun | Attendance 54% | L. Varghese · due todayauto-routed, severity high | Open |
| Lucas Chen | Failed CA1 | L. Varghese · 24 Julmeeting scheduled | Actioned |
| Noah Miller | Financial hold | Finance office · 26 Julreferred, acknowledged | Referred |
| Ava Johnson | Tutor flag | P. Raman · overdue 2 daysescalated to HoD | Escalated |
Lifecycle or product
The lifecycle is the operating model. This is the product that runs it.
Two pages, two jobs. Read the lifecycle to understand how the student record connects registration, attendance, degree audit and graduation; read this page to see the application your advisors open on Monday morning.
One of the institutional lifecycles: the governed model for matriculation, registration, records, degree audit and graduation.
The application that runs retention: risk, alerts, advising, referrals and intervention analytics.
How does the student record connect registration, attendance, progression and graduation?
What do advisors, tutors and the success office actually see and do in each screen?
The full record model, including registration, transcripts, degree audit and graduation clearance.
Signal to outcome: risk scoring, alerts, advising, case notes, referrals and retention analytics.
You are designing the operating model or comparing architectures.
See the lifecycle →
You are replacing a retention dashboard or a spreadsheet of at-risk students.
See inside the product →
Displacement
A retention dashboard tells you the rate. It cannot tell you who to call.
Analytics tools that sit beside the systems of record start from an overnight extract. They can rank a cohort, but they cannot open a case, route it to an advisor, record what was agreed or see whether the student turned up next week. The insight arrives without a mechanism, so the dashboard is admired and the student still leaves.
An overnight extract, already stale.
Read live from attendance, marks, engagement and finance.
A ranked list somebody must action.
An alert with an owner, a due date and a state.
Nobody can say.
Measured against comparable students who were not reached.
Good at charts. Disconnected from the advisor, the case and what happened next.
Also part of the operating system
Student Success & Retention is also part of Student Success & Records.
Connect it and risk reads the register marked this morning, the mark entered yesterday and the hold placed by finance, so an alert reflects the institution as it is rather than as it was last night.
Built for every role
One record. Five views of the same student.
Opens a caseload ranked by risk, and knows why each student is on it before the meeting.
Raises a concern from the class list and it becomes a task with an owner, not an email.
Sees open alerts, response times and which cohorts are drifting, during the term.
Referrals arrive with consent and context, and close with a recorded result.
Retention reported with evidence of what the intervention changed, not just the rate.
What academic teams say about running student success on Creatrix.
Creatrix Campus has helped us centralise all student information in one reliable and intuitive system, improving efficiency and reducing errors. We use it to manage student data, schedule sessions, and automatically generate transcripts, which has simplified many of our academic processes.
What I like best is its user-friendly and intuitive design. It makes managing tasks, attendance and academics seamless and efficient.
Go deeper
From the Creatrix library.
Blogs, whitepapers and case studies for the people who carry the caseload: advisors, tutors, heads of department and student services.
Frequently asked
Plain answers about student success and retention.
What signals does the risk score use?
Attendance decline, missed or failed assessments, engagement with the learning environment, outstanding fees and financial holds, prior-term performance, and any programme-specific factor you configure.
Signals are weighted by a model your institution owns and versions, and every score shows which factors contributed and by how much.
Does AI decide who is at risk?
No. The model ranks and explains risk and suggests a next-best action, but it never places a hold, contacts a student or closes a case on its own.
An advisor decides every intervention, and the decision, the owner and the outcome are recorded.
How is an alert different from a report?
An alert has an owner, a due date and a state. It is raised against a named student when a threshold is crossed, routed to the advisor or tutor responsible, and it stays open until an action is recorded.
A report tells you what happened. An alert asks someone to do something about it, and escalates if they do not.
Can faculty raise a concern directly?
Yes. A tutor or lecturer raises a manual flag from the class list with a category and a note, and it enters the same queue as a system-generated alert, with the same routing, ownership and closure rules.
How does advising work in practice?
Each advisor has a caseload showing their students, current risk, open alerts and upcoming appointments. Appointments are scheduled and logged, and case notes are recorded with confidentiality levels.
Referrals to counselling, learning support, disability services or finance are tracked to closure rather than sent as an email and hoped for.
Can we tell whether an intervention actually worked?
Yes. Because the alert, the action taken and the subsequent attendance, assessment and progression data sit on one record, outcomes for students who received an intervention are compared against comparable students who did not.
Results break down by cohort, programme and intervention type, so the success office can stop funding outreach that never moved a number.
Can we run it standalone?
Yes. It runs on its own against attendance, assessment and enrolment data imported from your existing systems, and it connects into the Academic Operating System when you are ready, so risk reads live signals from registration, attendance, examinations and finance without an overnight file.
SEE IT ON YOUR OWN COHORT
Bring one cohort and last year's withdrawals. We will show you when they were visible.
A tailored proof session uses your attendance, assessment and engagement signals, and your own advising structure, not a canned demo. Leave with a named, dated next step.