Docs
Your store
Affinsy runs inside your Shopify admin, on your store’s own orders. This page covers how orders arrive from Shopify, reading the figures at the top of the Reports page, and what the data quality pips mean for your reports.
Affinsy in your Shopify admin
Installing Affinsy from the Shopify App Store sets it up for your store. There is nothing to configure and no separate login: open it from Apps in your Shopify admin.
- Reports is the home page: your store’s figures at the top, your analyses below.
- Audiences holds the customer lists you saved from a finding, with their results.
- Data shows the sync with Shopify and the data health.
Your plan sets how many orders Affinsy keeps for the store.
How orders arrive
Affinsy reads your orders from Shopify itself. There is no file to export and nothing to copy between the two.
- Order history
- Loads in the background once the plan's trial starts. A few minutes on most stores; the Reports page shows the progress.
- New orders
- Arrive as they are placed. Edits, returns and cancellations update the order here too, and an hourly check catches anything missed.
- What is read
- Order dates, line items, prices and the customer's Shopify ID. No names, emails or addresses leave Shopify.
How orders are counted
- A customer’s spend adds up each order’s total, shipping and taxes included, after discounts and returns: the same figure as Amount spent on the customer’s page in Shopify.
- Gift cards stay on their order but are not counted in the customer’s spend, as in Shopify.
- Test orders and cancelled orders are left out.
The store figures
Five figures sit at the top of the Reports page, each with a 12-month sparkline. Hover the sparkline for any month’s value and change, and hover a figure’s name for how it is calculated.
- Orders
- Orders placed in the last 30 days of data, against the 30 days before.
- Avg order value
- Each order's subtotal after discounts, before shipping and taxes, divided by orders. Same window.
- New customers
- Customers whose first order fell in the last 30 days of data.
- Returning customers
- Share of the last 30 days' buyers who had ordered before. Amber when few customers ever come back: segmentation needs a second order to work with.
- Customer LTV
- What customers have spent to date, divided by customers. Spend is each order's total, shipping and taxes included.
A dash means the figure cannot be computed yet, with the reason next to it: for example no orders in the last 30 days, or no orders placed by an identified customer.
How to read it
“Last 30 days” means the last 30 days of your data, not of the calendar. The Last order label next to the date range tells you how fresh the data is. It turns amber once the newest order is more than a week old.
Data quality
The Data quality pips grade how much each analysis will have to go on in your store: Bought together, Customer segments, Repeat-purchase patterns, Reorder timing, Audiences and Revenue figures. Dark green is strong, teal fair, amber weak, grey unavailable. Hover the pips for the reason behind each grade.
Weak does not mean broken: the analysis still runs, with fewer or less certain patterns. Unavailable means the orders lack what the analysis needs, usually a customer on the order.
Each analysis takes the grade of its weakest input:
- Multi-item orders
- Share of orders with 2+ products. Strong from 35%, fair from 15%. Drives Bought together.
- Orders
- Strong from 2,000, fair from 300. Drives Bought together.
- Repeat rate
- Strong from 25%, fair from 10%. Drives Customer segments, Repeat-purchase patterns and Reorder timing.
- History
- Strong from 180 days, fair from 60. Drives Repeat-purchase patterns and Reorder timing.
- Customer ID coverage
- Strong from 70% of orders. Without any, everything except Bought together and Revenue figures is unavailable.
- Price coverage
- Share of products with a price. Strong from 90%, fair from 50%. Drives Revenue figures; without any there are no revenue figures: no campaign values, AOV or LTV.
What to do with it
- Low multi-item share: basket rules will be thin. Start with a customer segments report and reorder timing instead.
- Short history: reorder cycles need at least a few cycles to be visible, so a young store sees them fill in as orders arrive.
- Many guest orders: an order without a customer cannot be tied to a person, so customer segments and reorder timing count identified customers only.
Analyses
Every analysis you run appears as a card on the Reports page, with its runs listed underneath: when each ran, the period it covers and its headline result.
Hover a card for its actions:
- Re-run on fresh data once new orders have arrived. Earlier runs stay available in the report’s run picker.
- Schedule automatic re-runs, so the report stays current.
- Copy link: always opens the latest run.
- Share, Pin to top, Delete.
Read what each report shows in the market basket and customer segments guides.
Chat
Chatin the header opens an assistant that works on your store’s data. It reads your reports, looks up products, runs new analyses when a question needs one, and builds audiences. Its answers come from your numbers, not guesses.
- Chats are private to you.
- The same tools are available in Claude, ChatGPT and other clients over MCP. See Merchandising Engineer.
Team and settings
Everyone Shopify admits to the app on your store sees the same analyses and reports. The store owner manages the plan on the Billing page.
- Owner
- The plan and billing, and who has access.
- Member
- Everything else: analyses, audiences and reports.
The Datapage (side menu, under Store) holds the store’s sync with Shopify, when orders last arrived, and the data health with Preview rows.