Best Practices for LENS AI
Get more accurate, actionable insights by following these best practices when asking questions.
Writing Better Questions
Be Specific About Time
❌ Vague: "How is traffic?"
✅ Better: "How did traffic change last week compared to the week before?"
✅ Best: "Compare daily traffic for Jan 8-14 vs Jan 1-7, broken down by source"
Include Context
❌ Missing context: "Why did conversions drop?"
✅ With context: "Why did conversions drop on January 10th? We launched a new landing page that day."
Specify Metrics
❌ Ambiguous: "How are we performing?"
✅ Clear: "What's our conversion rate, average order value, and total revenue for this month?"
Name Your Segments
❌ Unclear: "How's mobile doing?"
✅ Precise: "What's the conversion rate for mobile users from paid search in the US?"
Question Templates
Traffic Analysis
"Show me [metric] for [time period], broken down by [dimension]"
Examples:
• "Show me sessions for last 30 days, broken down by source"
• "Show me pageviews for this week, broken down by device"
• "Show me unique users for January, broken down by country"
Conversion Analysis
"What's the [conversion type] rate for [segment] over [time period]?"
Examples:
• "What's the purchase rate for mobile users over the last 7 days?"
• "What's the signup rate for UK visitors over this quarter?"
• "What's the add-to-cart rate for organic traffic over Black Friday week?"
Comparison Questions
"Compare [metric] for [segment A] vs [segment B] during [time period]"
Examples:
• "Compare conversion rate for new vs returning visitors during January"
• "Compare revenue from Google vs Facebook campaigns last month"
• "Compare mobile vs desktop bounce rate this week"
Trend Analysis
"How has [metric] trended over [time period]? Highlight any [anomalies/changes]"
Examples:
• "How has conversion rate trended over the last 90 days? Highlight any drops"
• "How has traffic from Google trended this year? Note any significant changes"
• "How has average order value trended since we launched the new pricing?"
Investigation Questions
"Why did [metric] [increase/decrease] on [date/period]? Check [potential factors]"
Examples:
• "Why did traffic spike on March 15? Check for referral sources"
• "Why did conversion rate drop last Tuesday? Check by device and source"
• "Why did revenue increase 40% yesterday? Break down by product category"
Conversation Techniques
Build on Previous Answers
LENS remembers context. Use it:
You: What was our conversion rate last week?
LENS: 2.3%, down from 2.8% the week before...
You: Break that down by traffic source
LENS: [Uses same time period automatically]
Google Ads: 2.8%
Organic: 1.9%
...
You: Why is organic lower?
LENS: [Investigates organic specifically]
Organic conversion rate is lower because...
Ask for Recommendations
You: Which campaigns should I scale based on last month's performance?
LENS: Based on ROAS and volume, I recommend:
1. Scale: "Winter Sale - Retargeting" (ROAS 4.2, €12k spend)
2. Scale: "Brand Search" (ROAS 6.8, €5k spend)
3. Test increase: "Lookalike Audiences" (ROAS 2.8, €3k spend)
Campaigns to reduce:
• "Cold Audiences - Interest" (ROAS 0.8)
Request Specific Formats
You: Give me a table comparing all traffic sources for last month,
sorted by conversion rate, including sessions, conversions,
and revenue columns
LENS:
| Source | Sessions | Conversions | CR | Revenue |
|-------------|----------|-------------|-------|----------|
| Email | 12,450 | 498 | 4.0% | €24,900 |
| Google Ads | 45,230 | 1,130 | 2.5% | €45,200 |
| Organic | 34,120 | 682 | 2.0% | €27,280 |
...
Common Mistakes to Avoid
Don't Ask Multiple Unrelated Questions
❌ "What's our traffic? Also, what's the best performing campaign? And can you show me revenue by country?"
✅ Ask one topic at a time, then follow up.
Don't Assume LENS Knows Your Goals
❌ "Is this good?"
✅ "Is a 2.3% conversion rate good compared to our historical average?"
✅ "How does our 2.3% conversion rate compare to last year?"
Don't Forget Time Zones
If your data spans multiple time zones:
✅ "What was traffic on January 10th PST?"
✅ "Show me hourly traffic in UTC"
Don't Overlook Segments
Generic questions get generic answers:
❌ "Why is bounce rate high?"
✅ "Why is bounce rate high for mobile users from Facebook ads landing on the homepage?"
Getting Help with LENS
When LENS Isn't Sure
LENS will tell you when confidence is low:
LENS: I found a correlation between the traffic drop and your
Google Ads pause, but I'm only 60% confident this is the
main cause. Other factors might include:
• Server response time increased 200ms that day
• It was a holiday in Germany (your second-largest market)
Want me to investigate any of these further?
When Data Is Insufficient
LENS: I don't have enough data to answer this reliably.
You asked about conversion rate by city, but:
• 73% of sessions have no city data
• Only 12 cities have >100 sessions
Would you like me to analyze by country instead?
Providing Feedback
Help LENS improve:
- 👍 Thumbs up - Answer was helpful
- 👎 Thumbs down - Answer wasn't useful
- 💬 Comment - Explain what was wrong
Your feedback improves LENS for everyone.
Power User Tips
Save Useful Questions
Bookmark questions you ask frequently:
- After getting an answer, click ⭐ Save
- Access saved questions from LENS → Saved
- Re-run with one click
Create Question Templates
For repeated analyses with different parameters:
Template: "Weekly Performance Review"
Questions:
1. "Summary of traffic, conversions, revenue for last week vs previous week"
2. "Top 5 performing and bottom 5 performing campaigns by ROAS"
3. "Any anomalies or issues I should know about?"
Export Answers
Export LENS answers for presentations:
- Copy as Markdown - For docs and notes
- Copy as Image - For presentations
- Export to Sheets - For further analysis
Set Up Automated Questions
Have LENS answer questions automatically:
- Go to LENS → Automation
- Click Create Scheduled Question
- Enter your question
- Set schedule (daily, weekly, monthly)
- Choose delivery (email, Slack, dashboard)
Scheduled Question: "Weekly Performance Summary"
Question: "Give me a detailed summary of this week's performance
including traffic, conversions, revenue, top campaigns,
and any issues detected."
Schedule: Every Monday at 8:00 AM
Deliver to: team-marketing@company.com