Audience: Marketers, lifecycle leads, and campaign analysts using AID to monitor and improve campaign performance.
Prerequisites: Set up an Agent →
By the end, you’ll be able to:
- Navigate the four Insights views
- Understand performance metrics like lift, conversion, and engagement
- Optimize based on real data and segment-level patterns
Insights dashboard
When to use Insights
Use the Insights dashboard:
- After launching a campaign through an agent
- During experimentation with variants, timing, or channels
- Before making changes to content or targeting
- For reviews and reporting, including lift analysis and performance deep dives
What is Insights?
Insights is your analytics hub for AI Decisioning. It shows not just whether your campaign is working—but why, and how to improve it. You’ll see:
- Campaign performance vs. control (holdout) groups
- Top and bottom performing actions and variants
- Segment-level trends (e.g. by location, device, or loyalty tier)
- Optimal send windows by day and time
How to access Insights
- Go to AI Decisioning > Insights
- Select the agent you want to evaluate
- Choose a view:
- User breakdowns
- Personalization insights
- Creative overview
- Timing
Each view helps answer different campaign questions.
Optimization loop: See → understand → act
| View | See | Understand | Act |
|---|---|---|---|
| User breakdowns | Which segments are converting? | Where your campaign resonates—or doesn’t | Adjust content or targeting for each segment |
| Personalization insights | Which variants work best? | Why certain content wins with different users | A/B test versions, adjust tone or creative |
| Creative overview | Which messages underperform? | Signs of fatigue or misalignment | Pause poor variants, promote winners |
| Timing | When users are engaging | Ideal send windows by time zone | Refine delivery times in guardrails |
Insights view deep dives
1. User breakdowns
See how conversion rates vary by:
- Region or location
- Loyalty tier
- Age group
- Device or browser
Examples:
| What you see | What it might mean | What to do |
|---|---|---|
| CA users convert 40% higher | Strong regional resonance | Test localized content for other high-value states |
| Gen Z underperforms | Tone may feel too formal | Test emoji-heavy or casual subject lines |
| Desktop users engage more | Mobile layout issues | Optimize template design for mobile |
2. Personalization insights
See which variants resonate with which user groups. Patterns often emerge across:
- Subject lines
- Images or layout types
- Offers or recommendations
Examples:
| What you see | What it might mean | What to do |
|---|---|---|
| Variant A wins with users < 25 | Fun, visual-first content works | Test informal language and short layouts |
| Grid layout fails in Midwest | Cultural or visual mismatch | Try simplified or image-light formats |
| SMS beats push for urban users | Better accessibility or reach | Reallocate sends by region/channel |
3. Creative overview
Compare all message actions side-by-side using metrics like:
- Sends
- Click and conversion rate
- Unsubscribe rate
Examples:
| What you see | What it might mean | What to do |
|---|---|---|
| High unsubscribes on one variant | Off-tone or overexposed message | Pause it and test a softer CTA |
| Many sends, low conversions | Weak CTA or poor offer match | Refresh content, clarify incentive |
| Clicks high, conversions low | Broken link or poor landing page | QA destination and A/B test design |
4. Timing
View heatmaps of engagement by:
- Day of week
- Time of day
- Local time zones
Examples:
| What you see | What it might mean | What to do |
|---|---|---|
| Evenings convert better on West Coast | Better user availability | Extend send windows to 6–9 PM PST |
| Low Monday open rates | Inbox clutter or timing issue | Shift campaigns to midweek or Sunday |
Marketer cheat sheet — turn data into action
Use insights to evolve your campaign strategy
Here’s how marketers use Insights in practice:
- Swap out underperforming messages after checking conversion or unsubscribe rates
- Reallocate budget to high-performing user segments
- Adjust send windows based on timing heatmap trends
- Plan your next agent around what's working (e.g. double down on a winning subject line or segment)
Insights loop at a glance
| View | Use it to... | Then try... |
|---|---|---|
| User breakdowns | Spot strong segments | Tailor creative per group |
| Personalization insights | Match variants to user traits | Add targeted variants |
| Creative overview | Compare performance side-by-side | Pause weak messages, scale winners |
| Timing | See conversion timing trends | Adjust send windows in guardrails |
Metrics to watch
| Metric | What it tells you | Watch out for |
|---|---|---|
| Conversion rate | Is the message effective? | Low = rethink CTA, timing, offer |
| Unsubscribe rate | Are users opting out? | High = message fatigue or bad fit |
| Click rate | Is your content engaging? | High click + low conversion = check destination |
| Lift vs. holdout | Is AID driving real value? | Flat lift = experiment with new creative or audience |
Quick wins by use case
| Goal | Action |
|---|---|
| Boost conversions | Focus on top segments from user breakdowns |
| Cut unsubscribes | Remove overused messages or adjust cadence |
| Improve targeting | Use breakdowns and variants together |
| Optimize send times | Use heatmap to tune guardrails |
| Report success | Pull performance and lift insights for stakeholders |
Weekly optimization checklist
✔️ Review top and bottom performing actions ✔️ Scan unsubscribe and fatigue signals ✔️ Check lift vs. holdout for key audiences ✔️ Identify 1–2 test ideas (e.g. variant, time, audience) ✔️ Share an insight win with your team
Tips
- Use tags to group variants by tone or format
- Stack user breakdown + creative overview to find hidden wins
- Start small: one meaningful test per week leads to sustained growth