Your Metrics Look Fine… Until You Segment Them
A 8% conversion rate sounds fine. Until you segment: desktop converts at 14%, mobile at 3%. A 40% retention rate sounds okay. Until you segment: organic retains at 55%, paid at 18%. US teams that segment their metrics discover that averages hide both wins and leaks.**
Your dashboard shows solid numbers. Conversion: 8%. Retention: 40%. Engagement: good. You report them. You track them. You assume they're representative.
Then someone segments. By device. By source. By cohort. By behavior.
The averages collapse into a range. Some segments perform great. Others perform terribly. The "fine" metrics were hiding both. You've been optimizing for the average while the tails were pulling in opposite directions.
The Segmentation Truth
Conversion
| Segment | Conversion Rate | |---------|-----------------| | Overall | 8% | | Desktop, organic | 18% | | Desktop, paid | 9% | | Mobile, organic | 7% | | Mobile, paid | 2% |
The "2%" segment might be 30% of your signups. You're celebrating 8% while a third of your users convert at 2%. The average hid the leak.
Retention
| Segment | Week 4 Retention | |---------|------------------| | Overall | 40% | | Activated users | 65% | | Non-activated | 8% | | Organic | 52% | | Paid | 22% |
The 8% non-activated segment might be half your signups. They're dragging the average down. Fix activation, and overall retention jumps. The average hid the opportunity.
Engagement
| Segment | Actions/Week | |---------|--------------| | Overall | 12 | | Power users | 45 | | Casual users | 4 | | At-risk | 0.5 |
The "at-risk" segment might be 20% of users. They're one step from churn. The average engagement looked okay. The at-risk segment was invisible.
How to Segment Everything
The Core Dimensions
- Device: Desktop vs. mobile vs. tablet
- Source: Organic, paid, direct, referral
- Cohort: Signup week, signup month
- Activation: Hit first value or not
- Behavior: Power user, casual, at-risk
Segment every key metric by these dimensions. Conversion. Retention. Engagement. Revenue. The patterns will emerge.
The Tools
You need analytics that supports filtering. Funnels filtered by source. Retention filtered by cohort and activation. Conversion filtered by device.
SingleAnalytics supports segmentation across funnels, retention, and events. Filter by any property. Compare segments. Find the leaks and the wins.
What to Do With the Truth
Fix the Weakest Segment
The segment with the worst numbers is your highest-leverage fix. If mobile paid converts at 2%, either fix the mobile experience or stop buying mobile paid traffic. Improving that segment from 2% to 5% might mean a 50% overall conversion lift.
Scale the Strongest Segment
The segment with the best numbers is your template. If organic activated users convert at 25%, figure out how to get more of them. Can you drive more organic? Can you improve activation? Double down on what works.
Stop Optimizing for the Average
A/B tests on the landing page might move the average 1%. Fixing the mobile experience might move mobile conversion 10%. The latter could have a bigger impact. Segment to find the highest-leverage opportunity.
Real Impact
A US SaaS company had 7% conversion. They segmented. They found:
- Desktop organic: 16%
- Mobile paid: 1.8%
Mobile paid was 25% of signups. They were paying for users who almost never converted. They fixed the mobile checkout flow. Mobile paid conversion went from 1.8% to 6%. Overall conversion went from 7% to 9.5%. One segment. One fix. Massive impact.
Ready to see what your metrics are hiding? Segment everything with SingleAnalytics and find the leaks, and the wins.