Asking Questions
Qluent translates your questions into SQL queries. The better your question, the better the answer.
Your admin has configured Qluent with context about your business—which tables exist, what columns mean, how metrics are calculated, and common query patterns. When your question aligns with this context, Qluent gives accurate answers. When there's a gap, Qluent has to guess.
Not sure what to ask?
Start by asking Qluent what's available:
Discover your data
- "What tables do you have access to?"
- "What data do you have about orders?"
- "What customer information is available?"
Find out what's possible
- "What questions can I ask about sales?"
- "Give me some example questions about revenue"
- "What metrics are defined?"
This helps you discover what's possible and learn the terms that work best with your data.
Example questions
- "Total revenue from completed orders by country for Q4 2024"
- "Customer count by signup source for the last 90 days"
- "Average order value for returning customers by month"
Do's
Be specific about time
| Instead of... | Say... |
|---|---|
| "Recent sales" | "Sales in the last 30 days" |
| "This quarter" | "Q4 2024" or "October to December 2024" |
| "Year over year" | "Compare 2024 vs 2023" |
Name the measure you want
| Instead of... | Say... |
|---|---|
| "Top products" | "Top products by revenue" |
| "Best customers" | "Customers with highest order count" |
| "Performance" | "Conversion rate" |
Specify how you want results grouped
| Instead of... | Say... |
|---|---|
| "Sales breakdown" | "Sales by country" |
| "Customer analysis" | "Customers grouped by signup month" |
Use terms that match your data
If your data has "orders", ask about "orders" (not "purchases" or "transactions"). If metrics are defined as "MRR" and "ARR", use those terms.
Don'ts
Don't assume Qluent knows abbreviations
| Risky | Safer |
|---|---|
| "What's the AOV?" | "What's the average order value?" |
| "Show me the CR" | "Show me the conversion rate" |
| "LTV by cohort" | "Customer lifetime value by signup month" |
Unless your admin has defined these abbreviations, spell them out.
Don't skip the "by what" and "of what"
| Vague | Clear |
|---|---|
| "Show me the top 10" | "Top 10 products by revenue" |
| "What's trending?" | "Which products have increasing sales?" |
| "Compare performance" | "Compare revenue between Q3 and Q4" |
Don't use ambiguous references
| Ambiguous | Clear |
|---|---|
| "How did we do?" | "What was our total revenue last month?" |
| "The usual report" | "Monthly revenue by region for December" |
Each question should be self-contained.
Understanding responses
The verification badge
When you see a verification badge, Qluent used a pre-defined metric. This is the highest confidence answer—your admin has validated this exact calculation.
"No data" responses
This can mean:
- Filter too narrow → Try a different date range or fewer filters
- Value doesn't exist → Check spelling, try variations
- Data genuinely doesn't exist → Your database has no matching records
If results look wrong
- View assumptions → Did Qluent interpret correctly?
- Rephrase → Be more specific
- Escalate → If the issue persists, let your admin know
If things go wrong
Rephrase your question
Try being more specific:
- Add explicit filters
- Use exact column/value names
- Break compound questions into parts
Escalate
If rephrasing doesn't help, escalate. This notifies your admin team and helps improve Qluent for everyone. Don't feel bad about escalating—every escalation makes Qluent smarter.