Every sale in your dashboard traces back to a real conversation.
No estimates, no modelled guesses. When an order comes in, Linra ties it to the exact chat that earned it, or it counts nothing at all.
Revenue traced to a conversation
One example store: 8,240 across 143 orders where the visitor chatted first, each order matched to its conversation by a concrete signal.
The path of one traced sale
Attribution here is a real chain, not a score. Four steps run in order, and every one leaves a record you can open.
The shopper chats
A visitor opens the copilot, asks a question, and the copilot answers from your catalogue and pages. That conversation is stored with a first-party visitor ID.
The chat leads to checkout
The visitor adds to cart and buys. The visitor ID rides along on the cart as a cart attribute, so the order itself carries proof of who chatted.
The order is matched
When the order arrives, the resolver looks for the most recent chat interaction inside your attribution window and ties the two together by one concrete identifier.
It shows as traced revenue
The match is written down with the order value, the items bought, and what the copilot discussed. It appears in your dashboard, and it is one of only two things you are ever billed on.
A single order, traced back
Every attributed order names the signal it was matched on and the message it began at. Chat to order, first touch on Jun 22, held next to the order number and its value, so the number is proven, not asserted.
Shopify
Revenue you can point to
The three signals, in order of confidence
A match is only ever made on a real signal, and every match names the exact one it used. The resolver tries three in strict order, and the first that hits wins.
Cart attribute
Highest confidenceA visitor ID written straight onto the cart during the chat. It rides along with the order itself, so it is the strongest proof the buyer is the person who chatted.
Visitor cookie
One step belowThe first-party ID Linra's own widget set in the browser, reported back with the order. Same person, but it depends on the client reporting it, so it sits under the cart attribute.
The buyer's email matched to an email seen in a chat. It proves the same person, not the same browsing session, so it is the weakest and comes last.
What you see behind every charge
Open the Usage tab and every billable event is one row. Click a row and the chat that led to the order opens beside the order details, the full transcript, so you read the actual conversation behind the charge. If a chat did not earn it, one dispute button strikes the charge from your bill. In one example store, 600 conversations become 120 closed cases and 90 traced sales, and you can open the transcript behind every single one.
What attribution never claims
No view-through
Being on the page or seeing the copilot earns nothing. A match needs a real interaction tied by a concrete identifier.
No cross-device guessing
The keys are a first-party visitor ID or an email. If a buyer switches device and nothing shared comes back, there is no match.
No match, no record
If none of the three signals hit, nothing is attributed. Orders outside the window are dropped, and each order counts once.
Read on
This is the deep version of how revenue is traced. The billing model it feeds, one for a closed case and one for a chat-sourced sale, lives on the pricing page, the short map is in the docs, and the feature itself sits under revenue attribution.