Adaptive Recommendations

The GritCert platform analyzes each student’s quiz performance and automatically recommends targeted practice activities. This page explains how the recommendation engine works.

How Topics Are Prioritized

After each quiz attempt, the platform computes a per-topic priority tier for every topic covered by that lesson’s questions. The tier determines which activities are surfaced first and how urgently.

Priority Tiers

Needs Practice — topic correct rate below 70%

The student has insufficient mastery of this topic. Activities targeting this topic are surfaced immediately as the highest priority. Fires on any attempt, including the first.

Watch Out — topic is declining

The student previously showed stronger performance on this topic but has regressed. Triggered when all three conditions are true:

• Current correct rate is lower than the prior attempt
• The drop is 15 or more percentage points, OR the current rate is 20% or more below the prior rate (relative decline)
• Current correct rate is still below 75% (strong topics are not flagged for minor variance)

Reinforcement — topic is at or above 70% and stable or improving

The student is developing mastery. Activities are still recommended to lock in retention, with spacing intervals based on sub-tier:

• Emerging (70–79%): high-frequency spacing, activities every 2–3 days
• Solid (80–89%): moderate spacing, activities every 5–7 days
• Mastery (90%+): light spacing, activities every 14–21 days

Priority Order

When multiple topics need attention, the platform surfaces activities in this order:

1. Needs Practice (highest urgency)
2. Watch Out (declining topics)
3. Reinforcement (emerging and solid mastery)

Why These Thresholds

The platform is calibrated for high-stakes certification contexts where the exam pass threshold is 80%. The 70% floor for Needs Practice reflects research showing that students below 70% on a topic have roughly a 40% forgetting rate within two weeks — insufficient for reliable exam performance. The Watch Out tier uses relative decline rather than a fixed point drop to filter out random variance on small question draws (5–10 questions per lesson).

Activity Matching

Activities are matched to priority topics via topic tags stored on each activity. When the platform computes a student’s priority topics, it queries all available activities whose topic tags overlap with those topics, then ranks them by priority tier. Tenant admins can view, tag, enable, and disable activities from the admin console.