Tips & Best Practices

Writing Better Queries

The quality of your results depends on the quality of your questions. These techniques will help you get precise, actionable answers from Presso.

Be Specific with Date Ranges

Always include an explicit time frame. Without one, Presso has to guess what period you mean, which can lead to unexpected results.

Use natural language for dates -- Presso understands phrases like "last 7 days", "this month", "Q4 2025", and "January 1 to January 31".

Ask for Comparisons to Surface Insights

Raw numbers are hard to evaluate without context. Always ask for a comparison period so you can see whether metrics are improving or declining.

Show me this week's revenue compared to last week

Compare my Google Ads ROAS this month to last month

How does my conversion rate for the past 7 days compare to the
30-day average?

Comparison types that work well:

• Period over period: this week vs last week, this month vs last month
• Platform vs platform: Google Ads vs Meta Ads performance
• Segment vs segment: mobile vs desktop, new vs returning customers
• Channel vs channel: organic vs paid traffic conversion rates

Request Visualizations

Charts make trends and patterns immediately visible. Ask for them when reviewing time-series data or comparing categories.

Create a line chart of daily revenue for the past 30 days

Show a bar chart comparing ad spend by campaign

Plot conversion rate by traffic source as a chart

Charts are especially useful for spotting trends that are hard to see in tables, such as gradual declines or seasonal patterns.

Include the Metrics You Care About

Name the specific metrics you want to see. This avoids getting back a generic overview when you need particular numbers.

Show me top 10 products by revenue, including units sold,
average selling price, and return rate

List my Google Ads campaigns with spend, clicks, conversions,
CPA, and ROAS for the past 30 days

Cross-Platform Analysis

Presso's greatest strength is combining data from multiple sources. Use this to get a holistic view that would normally require switching between several dashboards.

Correlate Data Across Sources

Ask questions that span multiple platforms to uncover relationships:

Show me Shopify revenue alongside GA4 sessions for the past 30 days.
Do traffic trends match revenue trends?

Compare the revenue attributed by Google Ads to the actual Shopify
revenue from orders tagged with Google Ads UTM parameters

Which of my top Search Console keywords drive traffic to my
best-selling product pages?

Full-Funnel Analysis

Combine top-of-funnel (ads, search) with bottom-of-funnel (orders, revenue) data:

For my top 5 Google Ads campaigns, show me the full funnel:
impressions, clicks, sessions in GA4, add-to-carts, and Shopify orders

Which traffic sources from GA4 have the highest sessions but lowest
conversion rate in Shopify? Where am I losing potential customers?

Attribution Comparison

Compare the conversion numbers reported by Google Ads, Meta Ads,
and GA4 for last month. How much do they differ?

Understanding attribution differences across platforms helps you make better budget decisions.

Chaining Questions for Deeper Analysis

Do not try to get everything in a single question. Instead, use a sequence of questions that progressively narrow down to the insight you need.

The Funnel Approach

Start broad, then go deep:

1. Overview: "How did the store perform last week?"
2. Identify: "Which metric showed the biggest change?"
3. Investigate: "Break down that metric by traffic source"
4. Root cause: "What changed in that specific traffic source?"
5. Action: "What should I do to fix this?"

The Comparison Approach

Compare first, then explain:

1. Compare: "Compare this week to last week across all key metrics"
2. Isolate: "Revenue dropped -- break it down by product category"
3. Explain: "The electronics category fell. Show me individual products"
4. Verify: "Is this a stock issue? Check inventory levels for those products"

Follow-Up Refinements

After any answer, you can refine without restating context:

• "Filter that to just the last 7 days"
• "Sort by conversion rate instead"
• "Add a column for week-over-week change"
• "Show me the same data but for Meta Ads"
• "Now chart that"

Building Routines

Establish regular check-in routines so you catch problems early and make data-driven decisions consistently.

Suggested Cadence

Daily Check Template

Start each day with a single prompt:

Morning check: Yesterday's revenue and order count compared to the
same day last week. Any products below 10 units in stock that sold
more than 3 units yesterday? Any high-risk fraud orders?

Weekly Review Template

Weekly review for last week:
1. Revenue, orders, and AOV with WoW change
2. Daily revenue chart
3. Top 5 products by revenue
4. Google Ads and Meta Ads spend and ROAS
5. Top 3 things that improved and top 3 that declined

Performance Tips

Optimize Your Queries

• Ask for specific data rather than open-ended questions. "Show me revenue by day for the last 7 days" is faster than "Tell me everything about my store".
• Limit result sets when you only need the highlights. "Top 10 products" is faster than "All products sorted by revenue".
• Avoid unnecessary date ranges. Pulling 12 months of daily data takes longer than pulling 30 days. Request only the granularity you need.

Connection Priority

If you experience delays, check which data sources are connected. Each query to an external platform adds response time. When speed matters:

• Ask about one data source at a time rather than combining all sources in one query
• Break large multi-platform questions into sequential focused questions
• Start with the most important data source first

Managing Context

• Keep sessions focused on one topic for best results
• If a conversation becomes long with many data requests, start a new session to avoid context limitations
• Summarize findings before starting a new line of questioning

Security Best Practices

Data Access

• Presso connects to your data sources using OAuth tokens. These tokens are scoped to read-only access where possible.
• Presso does not store your e-commerce data. Data is fetched in real time when you ask a question and is not persisted after the session.
• Each data source connection is independent. Connecting Shopify does not grant access to your Google Ads data, and vice versa.

Authentication Hygiene

• Periodically review which data sources are connected in your Presso account settings
• If a team member leaves, revoke their Presso access and reconnect data sources with updated credentials if needed
• Use strong, unique passwords for each connected platform

Sharing Results

• When sharing Presso outputs with others, be mindful that they may contain sensitive business data such as revenue figures, customer details, or ad spend
• Exported CSV files and reports contain raw data -- store them securely
• Avoid sharing screenshots that include customer personally identifiable information

Access Control

• Only connect data sources that you need for your regular workflows
• Review connected integrations periodically and disconnect any that are no longer in use
• If you suspect unauthorized access, disconnect all data sources and reconnect them to generate new tokens