Chat analytics overview
Everything on Analytics is computed live from your store's actual conversation data — there's no separate analytics pipeline running behind it, and no delay while a batch job catches up. This page walks through where the numbers come from and what each one means.
Where the data comes from
Every chat conversation is stored as a thread — one browsing session's worth of chat — made up of individual messages. The dashboard reads straight from these:
| Metric | How it's calculated |
|---|---|
| Unique users | Distinct visitors, identified by ShopGuide's own first-party visitor id (aligned with Shopify's visitor id; a logged-in customer keeps a stable identity) |
| Sessions | Distinct browsing sessions, a 30-minute rolling window, aligned with Shopify's session id |
| Interactions | Sessions that included at least one customer message |
| User messages | Total messages sent by customers |
| Chat orders / revenue | Orders where a webhook matched the purchasing customer back to a prior chat session |
Visitor and session identity use ShopGuide's own first-party ids, which follow Shopify's own visitor and session ids where Shopify provides them, so the numbers stay comparable to Shopify Analytics. A logged-in customer keeps a stable identity. Nothing is stored or counted until the visitor accepts analytics tracking, and the ids are removed if they withdraw consent.
The funnel cards
Same underlying data, read as ratios:
- Chat View Rate — sessions where the widget was opened, over total sessions
- Session Interactions — sessions with a real exchange, over total sessions
- User Interactions — unique visitors who engaged, over all unique visitors
- Chat Conversion — chat sessions that ended in a purchase, over all chat sessions
- Session to Order — total orders over total sessions
- Chat Share of Orders — chat-attributed orders over all store orders
- Sessions per User — how many sessions the average visitor racks up
- Avg. Interactions / Session — average messages exchanged per session
Shopify sessions vs. ShopGuide sessions
Under Performance Ratios a toggle sets how "sessions" are counted:
- Shopify Sessions (the default): "Session metrics use Shopify Analytics sessions without bots, only from the countries in your Shopify Markets." Bots are always excluded. It needs the
read_reportsscope and Shopify's Protected Customer Data approval; without those, session-based ratios shown/ainstead of a misleading number. - ShopGuide Sessions: "Session metrics use every session ShopGuide tracked (page views included), bots not filtered." Always available.
The country filter is a setting: in Settings, under Dashboard Sessions, the checkbox Count only sessions from countries you sell to is on by default. Its help text: with it on, Shopify sessions also exclude visits from countries that aren't in your Shopify Markets, since those visitors can't buy from you. Turn it off to count visits from all countries.
The dashboard also has an Include Path Tracker-only sessions checkbox for sessions that only triggered page-view tracking without any chat message. It affects the chart line and the unique-user and message series.
Chat Conversion and Chat Share of Orders don't move when you flip this toggle — they don't use total sessions as a denominator in the first place.
Breakdowns and date range
Beyond the funnel cards, the dashboard also shows a device breakdown (mobile vs. desktop), an input-method breakdown (button clicks vs. typed text vs. voice), and a date-range picker (Today, Yesterday, Last 7/30/90/180/365 days, or a custom range) so any of this can be viewed for a specific window.
What's genuinely not here
No sentiment scoring, no churn prediction, no usage heat maps on this page. NPS is real but separate: if you turn it on in Settings → AI Behavior, a short two-question survey goes out ten days after a chat-assisted order, and responses land on their own dedicated NPS page — it's off by default and doesn't show up on the main analytics dashboard. For individual conversations and where customers actually hit dead ends, see Conversation insights.