A regional education provider wanted to identify learners needing support without turning a risk score into a label. Discovery showed that retention, service navigation and the governance of AI-assisted decisions had to be designed together. Marketways connected early support with student agency, confidentiality and evidence that an intervention helped.
The engagement objective
Through the initial discovery, Marketways defined the objective: coordinate retention, support, service and AI controls.
How Marketways translated the problem
We began with a practical question: Which students need which support, and when? The first analysis used attendance, learning, assessment, finance, support, engagement and outcomes.
That evidence could not be read in isolation. Risk prediction must not stigmatise students or confuse correlation with cause. Students encounter one institution while academic and administrative ownership is fragmented.
A risk score could identify a learner without explaining what support would help. We kept prediction, cause and intervention distinct so analytics did not become a label or substitute for educational judgement.
We did not judge each component by its isolated KPI. We examined how people, assets, decisions and constraints affected one another, then used the evidence to test whether an apparent improvement would strengthen the complete system or merely move cost, pressure or risk elsewhere.
How the engagement developed
The initial work on student retention and support exposed dependencies with learner experience and service redesign, education ai and assessment risk. Treating them as separate recommendations would have left the operating trade-offs unresolved.
- Student retention and support: Act earlier while preserving student agency and confidentiality.
- Learner experience and service redesign: Make services easier to navigate and resolve.
- Education AI and assessment risk: Set appropriate evidence, disclosure and oversight for each use.
Evidence we examined
- Attendance, learning, assessment, finance, support, engagement and outcomes.
- Journeys, contacts, complaints, cases, service levels and student context.
- Use case, assessment design, representative tests, overrides, appeals and outcomes.
Industry conditions we accounted for
- Risk prediction must not stigmatise students or confuse correlation with cause.
- Students encounter one institution while academic and administrative ownership is fragmented.
- The acceptable error and explanation differ for tutoring, marking, admissions and misconduct detection.
How our engagement contributed to business impact
We connected every method to a decision and a business measure. The organisation could assess the engagement through operating results as well as model performance.
- Student retention and support
- Method: Machine Learning & Predictive Analytics, Research & Evidence Collection.
- Evidence: Attendance, learning, assessment, finance, support, engagement and outcomes.
- Decision supported: Act earlier while preserving student agency and confidentiality.
- Impact measure: Retention, support uptake, attainment and fairness.
- Learner experience and service redesign
- Method: Process, Workflow & Systems, Market, Customer & Behavioural Analytics.
- Evidence: Journeys, contacts, complaints, cases, service levels and student context.
- Decision supported: Make services easier to navigate and resolve.
- Impact measure: Resolution, effort, repeat contact and trust.
- Education AI and assessment risk
- Method: Machine Learning & Predictive Analytics, Process, Workflow & Systems.
- Evidence: Use case, assessment design, representative tests, overrides, appeals and outcomes.
- Decision supported: Set appropriate evidence, disclosure and oversight for each use.
- Impact measure: Validity, subgroup reliability, appeal outcome and workload.
Implementation
The engagement was structured as pilot with outcome evaluation. We connected the analysis to the decisions, operating constraints and measures that the organisation would continue to use.
How success was assessed
The overall assessment considered timely support, retention, fairness and trust. The supporting measures included:
- Retention, support uptake, attainment and fairness.
- Resolution, effort, repeat contact and trust.
- Validity, subgroup reliability, appeal outcome and workload.
Services and methods used
Services: Risk Detection, Customer Satisfaction & Experience Research, Process & Workflow Analysis & Redesign, AI & Model Risk, Business Systems Design & Architecture.
Methods: Machine Learning & Predictive Analytics, Research & Evidence Collection, Process, Workflow & Systems, Market, Customer & Behavioural Analytics.