See where buyers stall in your catalog — then call the ones who did
Behaviour records what shoppers actually do on a Shopify storefront today — clicks, scroll depth, add to cart, checkout clicks, rage clicks — scores each session for intent and shows the heatmap, the replay and the funnel step where it stopped. The B2B twin will measure the wholesale equivalents: the catalog, the spec sheet, the MOQ and price-tier table and the quote form, with a quote funnel from catalog view to submitted quote. It is on the roadmap, with no date announced.
- Roadmap product — no date announced, no B2B numbers on this page
- Built on the shipped event pipeline, funnel builder, heatmaps and replays
- Privacy rules carry over unchanged: no input values, never the raw IP
Price list opened → Quote form abandoned
5m 10s · 7 pages
Compared 4 SKUs → Left in the catalog
6m 02s · 9 pages
Spec sheet downloaded → Quote submitted
2m 44s · 4 pages
~1 min
From embed to first data today
Enable the Comver Behaviour app embed, visit the storefront — click and scroll data appears within a minute
4 steps
Default Shopify Funnel today
Product view → Add to cart → Cart opened → Checkout clicked, measured live from the store’s own events — the pattern the quote funnel follows
0
Input values recorded
Forms log the submit, never what was typed; replays blank every field and the raw IP is never stored. This rule does not change for quote forms
30 days
Default replay retention
Recordings 1–365 days per store, events 90 days by default, enforced by a daily purge
These figures describe Comver AI Commerce for Shopify DTC as it ships today. The B2B twin is a roadmap product — nothing here is a B2B measurement, screenshot or price, and the mock-ups are concept previews with illustrative figures.
Every catalog event, named — and only what you switched on
One theme app embed records page and product views, clicks, scroll depth, cart and checkout events today, batched every five seconds or 25 events and swallowed on failure so the storefront never breaks. The B2B twin adds the surfaces a wholesale catalog actually has, on the same pipeline and the same switches.
Catalog & category views
Page type, referrer and UTM parameters, device and viewport; SKU, title and the price tier the buyer was shown on catalog pages.
Price list & tier tables
The tier table opened, which volume break was hovered or clicked, and whether the buyer’s own tier was visible without scrolling.
Spec & safety sheets
Specification, safety and warranty documents opened or downloaded — the technical objection, measured rather than assumed.
SKU comparison
Two or more SKUs in a category opened in one session, or the comparison table used, with the SKUs that were held against each other.
MOQ & quantity
Quantity steppers, pallet and case selectors, and attempts below the minimum order quantity — the moment a buyer discovers the MOQ.
Quote form
Opened, part-filled, abandoned or submitted — field names only, never what was typed. The B2B abandoned checkout.
Order pad & reorder
Quick-order bar entries, reorder-from-history clicks and standing-order edits, each carrying the module that produced them.
Friction
Rage clicks — three or more within 600 ms in a ~40 px radius — on tier tables, lead-time blocks and quote fields, reported as the element they hit.
Catalog → price list → quote started → quote submitted
The shipped Default Shopify Funnel counts sessions from product view to checkout click, names the step with the biggest drop-off and opens a prefilled experiment aimed at it. The B2B funnel keeps that shape and changes the steps, because a wholesale catalog does not end at a checkout button: it ends at a submitted quote or a raised purchase order. Every percentage will stay clickable — behind it there is a buyer, an account and a session you can open.
- Steps: Catalog viewed → Price list opened → Quote started → Quote submitted, filtered by device, period and account stage
- The biggest stall is named in a sentence, not left to be spotted — and hands straight to a Callio campaign on the buyers who stalled there
- Purchase orders and payment stay in Shopify, so the funnel ends where the measurable storefront ends and Accounts carries on from there
Quote funnel · where buyers stall
ComingCatalog → price list → quote → PO, measured from buyer sessions the way Behaviour measures product → cart → checkout today.
Bar lengths are illustrative. Nothing will be shown until real buyer sessions arrive.
The spec sheet, the MOQ table, the quote form
DTC stalls are cart hesitation and shipping cost. Wholesale stalls are different and duller: a minimum order quantity discovered too late, a lead time buried under an accordion, a price tier the buyer cannot find their own line in, a quote form that asks for a VAT number the buyer does not have to hand. Each of those is an event, so each can be counted, ranked and named in the morning brief — with the buyers it happened to attached, ready to call.
- Page heatmaps over a real captured catalog page: click density and scroll depth per device, and the top clicked elements with their selectors
- Compare all buyers with high-intent buyers, quoting buyers or the previous period, with one plain-language sentence about the key difference
- Copilot’s brief gains the account view: which price-list pages leak buyers, which quotes are ageing, which key accounts are slipping
Account-health brief
ComingRefreshed every morning, like the Copilot brief today
Reasons only over connected sources. Missing inputs are shown, not replaced with assumptions.
The behaviour engine that ships today for DTC
These are the shipped Comver AI Commerce screens the catalog view will be built from — the same capture, the same heat layer, the same sanitised replays and the same funnel builder, pointed at a storefront rather than a catalog.
Click density and scroll depth over a real captured page
One click captures a screenshot of any tracked page; Comver draws click density and scroll reach over it per device, and lists the top clicked elements with their text and CSS selector — so you can see whether the hottest thing on the page is the control you wanted.
Page heatmap
Where visitors click and how far they scroll on each tracked page.
Top clicked elements
Comver AI
Add to cart and the second thumbnail take most clicks; the size picker draws repeat clicks. Test an explicit selected size above the CTA.
Track the behaviour, not the buyer’s paperwork
The privacy rules are not re-opened for B2B — they are the shipped ones, and quote forms are exactly where they matter most. The script never reads what is typed, replays are sanitised in the browser before they leave it, and the server keeps coarse location and a salted hash rather than the raw address.
0
Input values read
Forms log the submit with id, name and action only — VAT numbers, delivery addresses and order references are never read
100%
Replays sanitised
Scripts, iframes, noscript, templates and inline handlers removed; every field and placeholder blanked before the snapshot leaves the browser
Never
The raw IP
Country, city and region plus a salted IP hash and user agent — the raw address is not stored
30 days
Default retention
Recordings 1–365 days per store, events 90 days by default; a daily purge enforces both
Logged-in customers pass their id, name and email through the embed today so their sessions can be matched to a record — so this is behaviour analytics with names attached where they exist, not anonymous analytics. That will be true of buyers too, and the account portal login is what will make it work.
From a stall to a segment, a campaign, an experiment
Behaviour is not a report you export. Today every stall becomes a segment, a journey, an on-site experience or a prefilled A/B test in one button, and each lands in the Comver tool that runs it. The B2B twin keeps the buttons and changes what they aim at.
- 1
Name the stall
The funnel says where buyers stop and the heatmap says what they clicked instead — “opened the price list, never started a quote”, “hit the MOQ at the quantity stepper”, “abandoned the quote form at the VAT field”.
- 2
Turn it into an audience
The detected stall becomes a saved segment with the accounts and buyers behind it, reusable in Accounts, Callio campaigns and catalog personalization — one definition, not three drifting copies.
- 3
Point Callio at it
Call the buyers who stalled, text when the call cannot connect, email when the phone cannot reach at all. Anything the agent should not settle goes to the rep who owns that account, transcript attached.
- 4
Fix the page and prove it
Open a template A/B test prefilled with the hypothesis — the tier table above the fold, the lead time out of the accordion, three fields fewer in the quote form — measured on real events with statistical confidence.
Frequently asked questions
Partly. The shipped Behaviour app embed will record page and product views, clicks, scroll depth, form submits and rage clicks on any Shopify storefront, including a catalog — so heatmaps, replays and journeys already work. What does not exist yet is the B2B vocabulary: price-tier tables, spec sheets, MOQ and quote forms as first-class events, and a quote funnel to put them in.
Catalog viewed → Price list opened → Quote started → Quote submitted, counted from buyer sessions the way the Default Shopify Funnel counts product view to checkout click today, with sessions, step conversion and drop-off per step, and the biggest stall named in a sentence you can hand to a campaign.
No — and that rule is not being relaxed for B2B. Form tracking records that a form was submitted, with its id, name and action, never field values. Session replays are sanitised DOM snapshots with every input, textarea and select value blanked before the snapshot leaves the browser.
They already do — a heatmap is drawn over a real captured screenshot of any tracked page, with click density, scroll reach per device and the top clicked elements listed with their selectors. What is planned is reading them in B2B terms: whether the buyer found their own price tier, whether the lead time was reached, whether the MOQ was discovered before or after the quantity was set.
No. Today the last measured step is “checkout clicked”: Behaviour measures the storefront and does not attribute revenue. The B2B twin will end at the submitted quote for the same reason — purchase orders, payment and terms live in Shopify, and attributed revenue is counted on the account in Accounts and Callio.
No. Intent scores, friction detection, funnels, journeys and replay briefs are rule-based. Copilot summaries and the AI notes on heatmaps and funnels are added on top when an AI provider is configured — and the product says “summary unavailable” honestly when it is not, rather than inventing a description.
Recordings default to 30 days and are adjustable from 1 to 365 per store; events default to 90 days. A daily purge enforces both, and recording is capped at 50 sessions a day by default (adjustable 0–1,000). These settings live in Comver, not in the theme.
No. Every B2B mock-up on this page is a concept preview with illustrative figures, and says so. The proof band describes Comver AI Commerce for Shopify DTC as it ships today. Nothing is shown in the product until real visits arrive.
Keep exploring
Find the step that costs you the quote
Catalog and quote-form behaviour is on the Comver AI B2B Commerce roadmap. Join the waitlist to help decide what ships first — or start today on the behaviour engine it will be built from.
14-day free trial · Live on Shopify in minutes · Checkout stays in Shopify