Performance Analytics turns raw Decision Engine activity into two sets of signals: Operational Health and Portfolio Health. Together they tell you whether your lending operation is running efficiently and whether the loans you're making are performing.
This guide has two parts. Part 1 sets out the business case behind each report: what it tells you and why it matters at management and board level.
Part 2 explains how to use each metric in practice: what to look for, how to interpret what you see, and how to respond when something looks off.
Part 1 - The business case
Why Performance Analytics
Lending managers review application volumes, turnaround times, and portfolio health every week. But raw numbers don't tell you whether to act.
Performance Analytics organises Decision Engine outcomes into two connected themes so that every metric becomes a signal you can act on. As a result, lending managers act with confidence rather than guesswork, and CEOs and boards have clear evidence of operational and lending performance.
The two themes answer different questions, but the real value is in reading them together: operational metrics tell you how you're lending; portfolio metrics tell you whether those decisions are paying off.
Operational Health - how efficiently are we running?
Turnaround rate
Turnaround rate is the headline growth metric. How fast you return decisions directly affects member experience and conversion. Applicants who get fast answers stay and become members; applicants forced to wait often go elsewhere.
Automation rate
Shows how much of your lending the Decision Engine is handling without manual review. A rising automation rate alongside stable delinquency is a signal that your credit policies are well-calibrated and your team's time is focused on the applications that genuinely need human judgement.
Override rate
Shows how often loan officers manually change an automated recommendation. Viewed alongside automation rate and early delinquency, it tells you whether human intervention is adding value or introducing inconsistency. And whether your ruleset is aligned with how your credit union actually lends.
Withdrawal rate
Shows how many applicants abandon the process before a decision. Every withdrawal is a lost lending opportunity, and one that's invisible without a dedicated measure.
Open Banking consent and utilisation rates
These tell you how many members are connecting their bank data and what proportion of those connections actually inform decisions. Decisions made with real transaction data produce more accurate affordability assessments and are associated with lower early delinquency.
Portfolio Health - are we growing within our risk appetite?
Portfolio Health is not just about how much you're lending, but how effective that lending is in terms of arrears. These metrics provide the regulatory-aligned figures boards need.
Portfolio Status
Breaks your loan book down by repayment status:
On Time,
Early Arrears,
Bad Loan,
and Default
as a board-ready donut chart you can present directly, without manual preparation.
Time to Delinquency
Goes deeper by showing when loans first fell into arrears relative to origination, it distinguishes two fundamentally different problems:
early-cohort delinquency points to assessment quality,
while later-cohort delinquency typically reflects life events that couldn't have been predicted at the point of decision.
The two require completely different responses. One means reviewing your lending criteria, the other means reviewing intervention strategies for existing borrowers.
Volume, value, and average loan value
These metrics provide the context needed to interpret portfolio shifts correctly, distinguishing a genuine change in quality from a change in lending volume.
Part 2 - How to use it
Note: Metrics refresh overnight after midnight every Sunday. Review Performance Analytics first thing Monday morning as part of your weekly lending review.
Operational Health metrics are available to all NestEgg users as standard. Open Banking metrics are only visible if your credit union uses the NestEgg Open Banking solution. Portfolio Health metrics require a monthly CAIS/INSIGHT file upload.
Understanding control boundaries
Every chart shows three reference lines:
Average (dotted blue line) - your own historical average, calculated on a rolling 28-week basis.
Upper boundary (dotted red line) - the upper edge of your normal range.
Lower boundary (dotted red line) - the lower edge of your normal range.
The boundaries are not fixed targets set by NestEgg. They adapt automatically as your data accumulates, reflecting what is normal for your credit union specifically.
A value within the boundaries means performance is consistent with recent history. Whereas a value outside the boundaries, in either direction, signals a meaningful shift that warrants attention.
Note: Boundaries update on the same 28-week rolling cycle as the rest of the metrics. In the early weeks, the boundaries will be wide and will narrow as more data builds up.
Operational Health
Review these every Monday morning. The metrics refresh overnight on Sunday, so Monday is your first opportunity to act on the latest data.
Metric | What it measures | Cadence |
Turnaround rate | % of decisions returned within 24h / 48h / 72h and still unfinalised | Weekly |
Automation rate | % of decisions made without manual review | Weekly |
Override rate | % of automated recommendations manually changed by a loan officer | Weekly |
Withdrawal rate | % of applicants who withdrew before receiving a decision | Weekly |
Open Banking consent rate | % of applicants invited to connect Open Banking who did so | Weekly |
Open Banking utilisation rate | % of decisions where OB data was available and actually applied | Weekly |
Turnaround rate
Your primary weekly health indicator.
What you see
Four tiles showing the percentage of decisions finalised within 0–24h, 0–48h, and 0–72h, plus an unfinalised figure for decisions still in progress. Select a tile to update the trend chart below.
How to read it
Start with the 0–24h tile: a high percentage means members are getting answers quickly. A falling 0–24h figure with a rising Unfinalised figure is your clearest signal that a backlog is building.
The buckets are cumulative: the 0–48h figure includes everything in the 0–24h figure. Use the shorter buckets to understand pace, and the Unfinalised figure to quantify any queue.
What to do when something changes
Unexpected drop in 0–24h rate: investigate whether staffing, system access, or application volumes changed that week.
Rising Unfinalised figure: this is a backlog building in real time. Prioritise clearing it before it affects member experience.
Improvement following a process change: a rising 0–24h rate after a change (e.g. enabling auto-accepts for a product) confirms the change is working.
Good practice: Check turnaround rate every Monday morning before reviewing anything else. A deteriorating trend here has the most direct impact on member experience and conversion.
Automation rate
Growth in automation coupled with steady portfolio performance indicates your configuration is effective and the Decision Engine is processing applications without manual intervention. When the automation rate decreases, more applications require manual intervention. This can stem from:
Strategic adjustments - the decline may be a planned result of recent changes to your decision rules.
Pool evolution - it may signal a change in the applicant demographic, where more cases fall into ambiguous “grey areas” that current rules can't automatically resolve.
Marketing efficacy - it can be a positive sign that marketing is attracting your target audience, producing more “refer” outcomes instead of outright declines.
Related metrics
Automation rate is most useful alongside override rate and time to delinquency. A high automation rate only has value if the automated decisions are also performing well in the portfolio.
Good practice: Track any change in automation following ruleset updates, as this may be an anticipated outcome. If the rate changes unexpectedly without a known trigger, evaluate whether the applicant demographic has evolved or your rules need adjustment.
Override rate
This metric shows the percentage of automated recommendations manually changed by a loan officer. For example, where the Decision Engine returned a Decline and the officer overrode it to an Accept, or vice versa.
What a good override rate looks like
The target is 5% or below, meaning at least 95% of automated recommendations are accepted without change. A rate above 5% should be investigated.
A high override rate means loan officers are frequently disagreeing with the Decision Engine. This isn't necessarily wrong, officers apply contextual judgement the engine can't, but at scale it usually means the ruleset is misaligned with how your credit union actually lends:
Too many Decline rules are triggering on applications your team would normally approve.
Too few Refer rules are triggering, so borderline applications pass through as Accepts when they should be escalated.
Watch out for
Sample size matters. With fewer than 50 decisions in the period, even 3 overrides will put you above 5%. Don't change rule settings based on a small sample. And always review Portfolio Health first — if overridden decisions are performing worse, turning off a Decline rule would increase future losses, not reduce friction.
How to investigate a high override rate
Identify the direction - are officers overriding Declines to Accept, or Accepts to Decline? The direction tells you which rules are misaligned.
Review with your underwriting team - walk through a sample. Is there a clear, consistent reason, or inconsistency between officers?
Check portfolio performance - look at Time to Delinquency and Portfolio Status before drawing conclusions. If overridden decisions perform in line with the book, rules may need recalibrating. If worse, the current rules may be protecting the portfolio.
Engage NestEgg if rule changes are needed - don't adjust rule settings unilaterally. Bring your findings and we can assess the appropriate change together.
Withdrawal rate
Shows the percentage of applicants who started an application but withdrew before receiving a decision. Every withdrawal represents a lost lending opportunity and, in most cases, a member who went elsewhere.
How to read it
A downward trend is positive: more applicants are staying in the process long enough to receive a decision. A rising withdrawal rate can indicate friction in the journey, delays long enough to prompt abandonment, or both:
Sudden spike: check whether a system issue occurred that week.
Gradual upward trend: investigate whether average turnaround time has increased. Members who wait too long often withdraw.
Step change: if withdrawal rate jumps following a product or journey change, the change may have introduced unexpected friction.
Good practice: Review turnaround rate and withdrawal rate together. A rising withdrawal rate alongside a falling turnaround rate is a strong signal that speed is the underlying cause.
Open Banking consent rate
Only visible if your credit union uses the NestEgg Open Banking solution.
Measures the proportion of applicants who successfully linked their bank accounts after being invited to do so. Activating Open Banking is only the first step, its effectiveness depends on applicant participation.
How to read it
A consent rate that falls outside the lower control boundary suggests a growing proportion of applicants are declining to share data. This could mean:
The member-facing explanation needs to be clearer or more reassuring.
Certain segments are opting out at higher rates. For example, members applying for smaller amounts may feel the request is disproportionate.
The invitation is appearing at a point in the journey where members aren't ready to engage with it.
A low consent rate reduces the proportion of decisions enriched with Open Banking data, which in turn affects the quality and accuracy of affordability assessment.
Open Banking utilisation rate
Only visible if your credit union uses the NestEgg Open Banking solution.
Of all applicants asked to connect their Open Banking data, the utilisation rate shows the proportion whose data was successfully retrieved and used to inform the decision. This differs from consent rate because consent alone doesn't guarantee data reaches the decision. Several factors can prevent a consented connection from contributing usable data:
Technical issues during the data exchange process.
Bank coverage - the applicant's bank may not be available for data sharing (check current availability in the supported banks and providers article).
Applicant friction - the applicant may have dropped off during connection after giving initial consent.
Connectivity barriers - the applicant may be on a device without their mobile banking app, or may have forgotten their login credentials.
How to read it
The gap between consent rate and utilisation rate is your data quality signal. If consent is high but utilisation is materially lower, applicants are starting the Open Banking journey but not completing it. A temporary dip may indicate a technical issue on the day; a sustained downward trend warrants investigation. Contact NestEgg support if the utilisation rate falls outside the lower control boundary for more than two consecutive weeks without an obvious explanation.
Portfolio Health
Portfolio Health metrics show how the loans you've made are performing over time. They are driven by your CAIS/INSIGHT file and should be reviewed monthly, not weekly.
Portfolio Health is only available once you are sharing a monthly CAIS/INSIGHT file with NestEgg. To upload your file, visit: dashboard.nestegg.ai/settings/upload-CAIS
The 70% match rate threshold
All Portfolio Health metrics require your decisions to be matched against your CAIS data. If the match rate falls below 70%, all Portfolio Health metrics are suppressed and a notice appears. If you see this notice:
Check for gaps in your CAIS uploads - missed or incomplete uploads lower the match rate because decisions can't be paired with repayment data.
Check for unfinalised decisions - decisions not finalised in the dashboard can't be matched. Every loan should have a corresponding Accept decision recorded.
Contact NestEgg if the issue persists - if uploads are complete and decisions finalised but the notice remains, contact support.
Portfolio Status
A complete view of your loan book by repayment status:
On Time, Early Arrears (1–2 missed payments),
Bad Loan (3–6 missed payments),
and Default.
Presented as a board-ready donut chart you can share directly.
How to use it
Review monthly alongside your other CAIS data. Look for shifts in the proportion of Early Arrears as this is your earliest warning of deteriorating quality, before loans progress to Bad Loan or Default.
If Early Arrears increases significantly month on month, investigate whether it's concentrated in a particular product, origination period, or applicant segment, and cross-reference with Time to Delinquency to understand the timing.
Time to Delinquency
Shows when loans first fell into arrears relative to origination, broken into four cohorts by time since origination.
Why it matters
Early-cohort delinquency (within the first 108 days) is a signal about assessment quality; if a disproportionate number of loans go bad quickly, the original decision may have been made with incomplete or inaccurate information.
Later-cohort delinquency is harder to predict at origination; loans that perform well initially but deteriorate later typically reflect life events, redundancy, health issues, relationship breakdown, rather than a failure of credit assessment.
Volume, value, and average loan value
These context metrics show lending activity over the period. They aren't performance indicators in themselves, but they're essential context for interpreting Portfolio Status and Time to Delinquency. Use them to distinguish between scenarios that can look similar on a donut chart:
Early Arrears proportion rises because portfolio quality has deteriorated.
Early Arrears proportion rises because lending volume dropped, so the higher-risk segment disproportionately affects the arrears rate.
Early Arrears proportion falls because lending volume increased and a new cohort of recent loans hasn't yet had time to demonstrate performance.
If lending volume increases sharply in the same period arrears proportions are rising, treat Portfolio Status with caution and allow more time before drawing conclusions about credit quality.
Good practice: A growing loan book with stable arrears proportions is a positive signal — you're scaling without a quality trade-off. A flat or shrinking book with rising arrears is a clearer signal of genuine deterioration that warrants investigation.
Monthly review checklist
Use this checklist when reviewing Performance Analytics at your monthly lending committee or board meeting.
| Review item | Section |
☐ | Turnaround rate: is the 0–24h rate stable or improving? Is Unfinalised growing? | Op. Health |
☐ | Automation rate: is it stable? Has it changed following a ruleset update? | Op. Health |
☐ | Override rate: is it at or below 5%? If above, review with the underwriting team. | Op. Health |
☐ | Withdrawal rate: is it stable or improving? Cross-reference with turnaround rate. | Op. Health |
☐ | OB consent rate: is it within normal range? (OB clients only) | Op. Health |
☐ | OB utilisation rate: is the gap between consent and utilisation acceptable? (OB clients only) | Op. Health |
☐ | CAIS file uploaded for the current month | Port. Health |
☐ | Portfolio Status: is Early Arrears proportion stable or improving? | Port. Health |
☐ | Time to Delinquency: is early-cohort delinquency within normal range? | Port. Health |
☐ | Volume / value: does lending activity explain any shifts in portfolio proportions? | Port. Health |