Nudge performance reports
How to read per-nudge-type stats, the engagement funnel, and A/B test results in the elNudge dashboard.
The Nudges panel in Analytics shows how each nudge type is performing — how often nudges are shown, how visitors respond, and how many of those interactions lead to a purchase.
Per-nudge-type stats
elNudge fires different nudge types depending on visitor behaviour and intent score:
| Nudge type | When it typically fires |
|---|---|
| Chat bubble | Persistent; appears after a short delay on product pages |
| Toast | Brief pop-up for high-intent moments (e.g. returning to a product page) |
| Exit-intent modal | Triggered when the visitor shows exit signals |
| Sticky bar | Subtle banner, often used for low-intent or browsing-stage visitors |
For each nudge type, the panel shows:
| Metric | What it measures |
|---|---|
| Shown | Number of times this nudge type was displayed to a visitor |
| Click rate | % of shown instances where the visitor clicked the nudge CTA |
| Dismiss rate | % of shown instances where the visitor dismissed the nudge |
| Chat open rate | % of shown instances where the visitor opened the chat panel |
| Conversion rate | % of sessions where this nudge was shown that ended in a purchase |
Reading the engagement funnel
Each nudge type has its own mini-funnel:
Shown → Clicked → Chat opened → PurchaseWork through the funnel top to bottom to identify where drop-off is happening:
- Low click rate — the nudge message is not compelling, or it is firing too early (visitor is not ready). Try adjusting tone or reviewing your agent's greeting style.
- High click rate, low chat open rate — visitors are engaging with the nudge but not committing to chat. Consider whether your CTA language sets the right expectation.
- High chat open rate, low conversion rate — visitors are chatting but not buying. Review recent chat transcripts (Dashboard → Conversations) to see what questions are going unanswered or where conversations stall. This is usually a knowledge base gap.
- High dismiss rate — the nudge may be interrupting at the wrong moment. Check your quiet zone settings (Dashboard → Settings → Quiet Zones) to exclude pages or timing windows that are not conducive to nudging.
Identifying underperforming nudge types
Sort the Nudges panel by conversion rate to identify nudge types that are showing often but not contributing to purchases. Common causes and fixes:
| Symptom | What to try |
|---|---|
| Toast has low click rate | Message too generic — use Custom AI Instructions to make it more product-specific |
| Exit-intent modal high dismiss | Offer value in the modal message — ask a question rather than making a statement |
| Chat bubble rarely opened | Check if it is appearing on pages where visitors are not ready to engage (e.g. home page, blog) |
A/B test results (Growth+ plans)
If you are running an A/B test (configured in Dashboard → A/B Tests), test results appear in a variant comparison table below the main Nudges panel.
The table shows, for each variant:
| Column | What it shows |
|---|---|
| Variant | A (control) or B (challenger) — with the config change displayed |
| Sessions | Total sessions in this variant group |
| Nudge CVR | Conversion rate of sessions where a nudge was shown |
| Baseline CVR | Conversion rate of sessions with no nudge interaction in this variant |
| Lift | Difference in CVR between nudge-influenced and baseline for this variant |
| Statistical confidence | How confident the system is that the result is not random (shown as a percentage) |
Wait until confidence reaches 95% before drawing conclusions. Ending a test early on promising-looking results is a common cause of false positives.