Predictive Power of Visitor Metrics
Quick Summary
Not all columns in the Visitors table are equal predictors. Engagement Status, Secondary Conversions, and Session Count are the strongest predictors of conversion likelihood, while Location and Device are informational only.
Core Insight
Predictive Power Ranking
| Rank | Metric | Predictive Value | Why |
|---|---|---|---|
| 🥇 | Engagement Status | Very High | Directly reflects behavioral intensity |
| 🥈 | Conversions (secondary) | Very High | Past action predicts future action |
| 🥉 | Activity (sessions) | High | Return visits show commitment |
| 4 | Activity (time) | High | Time investment = interest level |
| 5 | Channels | Medium-High | Source quality varies significantly |
| 6 | Activity (pages) | Medium | Exploration depth indicates interest |
| 7 | Value Score | Medium | Historical value, not predictive |
| 8 | Location | Low | Informational only |
| 9 | Device | Low | Informational only |
Top Predictors Explained
1. Engagement Status (Strongest Predictor)
Why it predicts: Combines multiple behavioral signals into one score.
| Status | Conversion Likelihood |
|---|---|
| Hot Lead 🚀 | Very High |
| Engaged 🔥 | High |
| Returning 🔄 | Medium |
| New 💎 | Low |
How to use: Sort by Engagement. Focus outreach on Hot Lead and Engaged visitors.
2. Secondary Conversions (Strong Predictor)
Why it predicts: Visitors who take one action are much more likely to take another.
| Secondary Conversions | What It Signals |
|---|---|
| 0 | Passive browser |
| 1 | Willing to engage |
| 2+ | Highly motivated |
How to use: Any visitor with Conversions > 0 (even secondary) is a priority.
3. Session Count (Strong Predictor)
Why it predicts: Return visits indicate your product stays on their mind.
| Sessions | Conversion Likelihood |
|---|---|
| 1 | Low — just browsing |
| 2-3 | Medium — evaluating |
| 4+ | High — seriously considering |
How to use: Check the Activity column. Prioritize visitors with 2+ sessions.
4. Time on Site (Moderate-Strong Predictor)
Why it predicts: Time investment shows genuine interest, not accidental clicks.
| Time Spent | Conversion Likelihood |
|---|---|
| < 30 seconds | Very Low — bounced |
| 30s - 2 min | Low — casual browse |
| 2 - 5 min | Medium — interested |
| 5+ min | High — evaluating seriously |
How to use: In Activity column, look for time values above 2 minutes.
5. Channel Source (Moderate Predictor)
Why it predicts: Traffic quality varies by source. Some channels bring buyers, others bring browsers.
| Typically Higher Intent | Typically Lower Intent |
|---|---|
| Organic Search | Social (casual browsing) |
| Paid Search (branded) | Display ads |
| Direct | Referral (varies) |
| Paid Social (varies) |
How to use: Compare conversion rates by channel in your data. Double down on high-converting sources.
Metrics That Don't Predict Well
| Metric | Why It's Not Predictive |
|---|---|
| Location | Where they are doesn't indicate intent |
| Device | Desktop vs. mobile doesn't predict conversion (usually) |
| UTM Parameters | Useful for attribution, not prediction |
| Period/Date | When they visited doesn't predict if they'll convert |
These columns are valuable for understanding visitors, but not for predicting behavior.
Practical Prediction Formula
High conversion likelihood =
Hot Lead OR Engaged status + 2 or more sessions + 2+ minutes time spent + Any secondary conversion
If a visitor matches 3+ of these criteria, they're a strong conversion candidate.
Why It Matters
Behavior predicts behavior.
The best predictors are all action-based:
- How often they return (sessions)
- How long they stay (time)
- What they do (conversions)
- How active they are (engagement status)
Demographics (location, device) tell you WHO. Behavior tells you WHAT THEY'LL DO.