Basic Usage - Presso Docs

How to Ask Questions

Presso connects your e-commerce data sources to Claude Code through the MCP (Model Context Protocol). You interact with it by typing natural language questions directly into Claude Code -- no special syntax or query language required.

Write Like You're Talking to an Analyst

The best way to use Presso is to ask questions the same way you would ask a data analyst sitting next to you. Be specific about what you want, and include any relevant context.

Show me total revenue and order count for the last 7 days

What were my top 10 best-selling products last month by units sold?

Compare my Google Ads and Meta Ads ROAS for the past 30 days

Be Specific About What You Want

Vague questions produce vague answers. The more detail you provide, the more useful the response will be.

Instead of...	Try this
"How are sales?"	"Show me daily revenue for the past 14 days with week-over-week comparison"
"Show me ads data"	"What are my top 5 Google Ads campaigns by conversions this month?"
"Any issues?"	"Are there any products with fewer than 10 units in stock that sold more than 20 units last week?"

Specify Date Ranges

Always include a time frame in your question. Presso supports natural language date references:

Relative dates: "last 7 days", "past 30 days", "this month", "last quarter"
Specific dates: "from January 1 to January 31", "on February 14"
Comparisons: "this week vs last week", "February vs January"

Show me daily sessions from GA4 for the last 14 days compared to the previous 14 days

Understanding Responses

Presso returns data in several formats depending on what you ask for. Claude Code will choose the most appropriate format automatically, but you can also request a specific format.

Data Tables

For structured data like product lists, order summaries, or campaign metrics, you will receive formatted tables:

Show me my top 5 products by revenue this month

This returns a table with columns such as product name, units sold, revenue, and average order value.

Analysis and Insights

When you ask "why" or "what happened" questions, Presso pulls the relevant data and Claude Code provides analysis:

Why did my conversion rate drop last week?

Claude Code will examine traffic sources, landing page performance, device breakdown, and other factors to identify likely causes.

Charts and Visualizations

You can request visual representations of your data. Claude Code can generate charts directly in supported environments:

Show me a chart of daily revenue for the past 30 days

Create a bar chart comparing ad spend across all my campaigns this month

Querying Patterns

Presso supports several common query patterns that map to the kinds of questions e-commerce operators ask every day.

Time-Based Queries

Pull metrics for a specific period or track trends over time.

What was my total revenue yesterday?

Show me weekly order counts for the past 3 months

What is the trend in my average order value over the last 90 days?

Comparison Queries

Compare metrics across time periods, platforms, campaigns, or segments.

Compare this month's revenue to last month

How does my Google Ads CPA compare to Meta Ads CPA for the last 30 days?

Compare conversion rates between mobile and desktop traffic this week

Top/Bottom Queries

Find your best and worst performers across any dimension.

What are my top 10 products by revenue this month?

Which 5 ad campaigns have the lowest ROAS this quarter?

What are the bottom 10 search queries by click-through rate in Search Console?

Threshold and Alert Queries

Identify items that meet specific criteria.

Which products have fewer than 5 units in stock?

Are there any ad campaigns spending more than $100/day with ROAS below 1.0?

Which products had zero sales in the last 30 days?

Working with Multiple Data Sources

One of Presso's most powerful features is the ability to query across multiple data sources in a single conversation. This lets you correlate information that normally lives in separate dashboards.

Cross-Platform Questions

Show me total ad spend and ROAS from both Google Ads and Meta Ads for last month, side by side

Compare my Shopify revenue to the revenue reported in GA4 for the past 7 days

Which of my top-selling Shopify products get the most organic search traffic?

Multi-Step Analysis

You can build a complete picture by asking a sequence of related questions:

Start broad: "How did my store perform last week?"
Identify an area of interest: "Break down revenue by traffic source"
Drill into the detail: "Show me the Google Ads campaigns driving that paid search traffic"
Take action: "Which of those campaigns should I increase budget on based on ROAS?"

Each follow-up question benefits from the context of previous answers in the same conversation.

Following Up and Drilling Down

After receiving an initial answer, you can ask follow-up questions to explore the data further. Claude Code maintains context throughout the conversation, so you do not need to repeat yourself.

Drill Down

→ "Show me revenue by category for last month"
→ "Break down the Electronics category by individual product"
→ "What's the return rate for that top product?"

Pivot

→ "Show me top traffic sources by sessions this week"
→ "Now show me the conversion rate for each of those sources"
→ "What's the average order value for the organic search segment?"

Refine

→ "Show me all orders from last week"
→ "Filter that to only orders over $100"
→ "Of those, how many were from repeat customers?"

Claude Code can format data for easy export and sharing.

Request Specific Formats

Export last month's sales data as a CSV table

Summarize this week's performance in a format I can paste into an email

Create a markdown table of my top campaigns that I can add to a report

Save to Files

When working in Claude Code, you can ask for results to be written directly to files:

Save the monthly revenue report as a CSV file at ./reports/monthly-revenue.csv

Session Tips

Start a new conversation for new topics. If you switch from analyzing ad performance to investigating inventory, starting fresh avoids confusion from prior context.
Keep related questions in one session. Follow-up questions and drill-downs work best when the prior context is available.
State your goal up front. Saying "I want to understand why conversions dropped last Tuesday" gives Claude Code context to pull the right data from the start.
Ask for summaries. After a deep-dive analysis, ask Claude Code to summarize the key findings and recommended actions.