Understanding Engagement Scores

Engagement scores give you a single number that reflects how actively each client is using Policy Stack. Understanding what goes into the score — and what the thresholds mean in practice — helps you use it as a useful attention signal rather than a performance grade.

What Factors Into the Score

The engagement score is a weighted composite of five activity signals. Each signal is measured over the trailing 30 days and weighted by recency — activity from the last 7 days counts more than activity from 3 weeks ago.

| Signal | Weight | What It Measures | |--------|--------|-----------------| | Login frequency | High | How often the client logs in — daily vs. weekly vs. sporadic | | Policy snapshots recorded | High | Whether the client is keeping their banking system data current | | Features explored | Medium | How many distinct areas of the platform the client has used | | Repayment schedules updated | Medium | Whether the client is actively maintaining their repayment plans | | AI Chat sessions | Low | Usage of the AI assistant feature |

Why Login Frequency and Snapshots Are Weighted Highest

A client who logs in regularly and keeps their snapshot data current is getting genuine value from the platform — they can see their banking system in context. A client who never logs in or whose policy data is months old can't effectively track their system regardless of what other actions they take.

Why AI Chat Is Weighted Low

AI Chat usage is a signal of engagement, not a prerequisite for it. Many clients who use Policy Stack primarily for tracking and visualization never use the chat feature. Its low weight ensures it doesn't distort scores for clients who engage meaningfully in other ways.

Score Thresholds and What They Mean

| Score | Band | What It Typically Reflects | |-------|------|---------------------------| | 75–100 | Active | Logging in at least weekly, snapshot data up to date, using multiple features | | 40–74 | Moderate | Occasional logins (weekly to monthly), some features explored | | 10–39 | Low | Infrequent logins, limited feature usage, possibly stale snapshot data | | 0–9 | Inactive | No meaningful platform activity in the past 30 days |

A Note on New Clients

New clients (those who joined in the last 30 days) will have lower scores simply because they haven't had time to accumulate activity signals. A new client with a score of 30 is not necessarily disengaged — they may have logged in several times, entered their first policy, and are just getting started. Watch the trend over their first 60 days rather than the absolute score in the first few weeks.

How Scores Update

Scores are recalculated nightly. A client who logs in and records a snapshot today will see their score reflect that activity in tomorrow's calculation. The 30-day trailing window means scores can drop naturally as past activity falls off — a client who was very active 3 weeks ago but hasn't logged in since will see their score decline even without doing anything different.

Using Scores as an Attention Signal

The engagement score is most useful for directing your attention, not for judging outcomes. Common patterns and what they suggest:

High score client — minimal outreach needed. Clients tracking themselves well don't need you to prompt them for data. They're using the platform as intended.

Moderate score client — periodic check-in. These clients are engaged but not maximally so. A quarterly check-in call, an update about a new feature, or a reminder to record a recent snapshot often re-activates engagement.

Low or Inactive score client — proactive outreach. A client who has drifted into low or inactive territory may have questions, may have lost the habit of logging in, or may be in a period where they're not actively thinking about their banking system. A personal note or call — not a system-generated message — is usually the most effective way to re-engage.