
Segments that power Analytics reports and Engage sends
Define behavioral segments once in Rheo and use them for Analytics cohort analysis, Convert targeting, and Engage broadcasts and automations: without CSV exports.
Segments are the handoff point between understanding users and acting on that understanding. When segments live only in a analytics vendor, growth exports CSVs. When they live only in email tools, product cannot reproduce the audience in a funnel chart. Rheo defines segments on the unified customer profile so Analytics, Convert, and Engage read the same rules.
This post explains segment types, authoring patterns, consent gates, and operational hygiene: so your next win-back send targets exactly the users who stalled on the paywall last week.
Why one segment definition matters
| Problem with duplicate definitions | Consequence |
|---|---|
| “Power users” means 5 sessions in analytics, 10 in CRM | Wrong send size |
| Funnel segment not available in Engage | Delayed campaigns |
| Manual refresh | Stale audiences |
Rheo segments are declarative: traits, events, and time windows evaluated against the profile.
Building blocks
| Building block | Example |
|---|---|
| Trait | plan = free, marketing_email_consent = true |
| Event count | paywall_viewed ≥ 2 in last 7 days |
| Event sequence | trial_started without purchase_completed in 14 days |
| Revenue | lifetime_revenue > 0 from billing integration |
| Flow state | Started onboarding v12, did not complete |
| Platform | platform = ios |
Combine with AND/OR; nest for advanced cases per your workspace features.
Lifecycle of a segment definition
Segments are not static labels: they move through a simple lifecycle:
| Stage | Action |
|---|---|
| Draft | Product proposes rule on paper; validate event exists in staging |
| Validate | Compare segment size to manual SQL or legacy tool export |
| Active | Used in dashboards and at least one Engage or Convert surface |
| Monitor | Size and conversion impact tracked weekly |
| Archive | Rule deprecated; automations detached |
Skipping validate is how teams send email to zero users or to the entire database.
Segment patterns for subscription apps
Acquisition quality
| Segment name | Rule sketch | Analytics | Engage |
|---|---|---|---|
| Paid social trialists | source = facebook + trial_started 7d | Cohort retention | Educate on day 3 |
| Organic high intent | Organic + paywall_viewed | Funnel compare | Optional nudge |
Onboarding and Convert
| Segment name | Rule sketch | Analytics | Engage |
|---|---|---|---|
| Stuck step 4 | Flow step reached, not complete 48h | Drop-off confirm | Email with deep link |
| Experiment control holdout | Variant = control | Lift analysis | Exclude from promo |
Tie to onboarding funnel drop-off.
Revenue and churn
| Segment name | Rule sketch | Analytics | Engage |
|---|---|---|---|
| Trial ending | trial_ends_at within 48h | Conversion rate | Push + email |
| Lapsed paid | Was subscriber, not active | LTV post-mortem | Win-back journey |
Billing traits sync from RevenueCat/Stripe; verify latency in docs.
Consent-safe marketing
Always intersect promotional audiences:
marketing_email_consent = true
AND NOT unsubscribed
AND <your behavioral rule>
See marketing consent.
Analytics: segments as lenses
In Analytics reports:
- Filter funnels and retention by segment.
- Compare completion rates: stuck users vs completers.
- Attribute revenue: which segment drove MRR this month?
Export segment definitions in team wiki: not one-off screenshots: so Engage authors trust the numbers.
Engage: segments as audiences
| Send type | Segment use |
|---|---|
| Broadcast | Snapshot audience at send time |
| Automation entry | User enters when segment becomes true (or on event + segment) |
Dynamic membership: user who fixes onboarding exits “stuck” segment automatically: avoid sending “finish setup” after completion.
Convert: targeting experiments
Use segments to restrict who sees a channel or experiment arm:
- iOS-only fix for layout bug.
- Locale-specific flow for DE users.
- Exclude employees via
is_internaltrait.
Experiment analysis still segments by variantId inside targeted population.
Operational hygiene
| Practice | Frequency |
|---|---|
Name segments with owner prefix growth_trial_ending | Always |
| Document definition in description field | On create |
| Review size before broadcast | Each send |
| Archive unused segments | Quarterly |
| Test with internal profiles | Each new rule |
| Red flag | Action |
|---|---|
| Segment size 0 | Rule typo or missing events |
| Segment = entire MAU | Missing consent filter |
| Sudden 10x jump | Billing sync or event duplicate |
Segments vs tags vs lists
| Concept | Rheo approach |
|---|---|
| Static email list | Prefer dynamic segment |
| One-off CSV upload | Last resort; breaks loop |
| Real-time event trigger | Automation entry event |
| Slow-changing trait | Segment on trait |
Advanced: warehouse-enriched traits
If you compute scores in Snowflake, push traits via API on schedule:
churn_risk_score updated nightly → segment churn_risk > 0.8 for Engage.
Keep operational segments in Rheo; keep model features in warehouse with thin sync.
Connect to the growth loop
- Analyze funnel → identify population.
- Segment define rule.
- Convert fix in-app for subset OR Engage message subset.
- Analyze again: did segment shrink?
Hub post: Analyze → segment → convert → engage.
Worked example: paywall stall end-to-end
Monday: Analyze: Funnel on live onboarding pin shows 58% loss between personalization and paywall on Android only.
Tuesday: Segment: Create growth_android_paywall_stall = reached paywall_viewed without trial_started within 72h, Android, live pin 19.
Wednesday: Convert: Publish variant B with one fewer screen before paywall; 50/50 experiment on Android channel.
Thursday: Engage: Automation entry on same segment (exclude experiment holdout if policy requires): wait 6h, email with deep link to channel.
Friday: Analyze: Compare trial start rate: segment size shrinking, experiment variant B + message arm vs control.
Document the segment ID in the experiment ticket so nobody re-exports a CSV “just for marketing.”
Collaboration norms
| Role | Owns |
|---|---|
| Product | Segment naming, hypothesis |
| Growth | Engage copy and send calendar |
| Engineering | Trait correctness, billing sync |
| Data | Warehouse traits pushed to profile |
Weekly 15-minute segment review in shared doc: new segments, archived segments, size anomalies.
Bottom line
Segments are not a marketing feature: they are the contract between analytics and action. Rheo teams that treat them as shared infrastructure ship campaigns that match what the dashboard already proved.