Every ID your buyer left, on one ring. Every touch, on one trail
Keyring works out which records across your tools are the same person, and asks you when it can't be sure. Paper Trail lays out what that person did, from the first ad click to the payment.
We'll match your five latest sales to the ads that made them, or you don't pay.
An animation in five steps. Six tools each hold a different record of the same buyer. Keyring joins them into one person, asking for approval on the weak match. Paper Trail lays out her touches in order, from the first ad click to the payment. The payment is credited to channels under six attribution models. Ranked by cash, not leads, the ad that looked weakest turns out to pay best.
Every tool met your buyer under a different name
Meta has a click ID. Your site has a visitor cookie. Typeform and iClosed have her email, HubSpot has her phone, Stripe has a customer ID. Keyring joins them into one person, and a match made on a phone number alone waits for your yes.
Six tools each hold a different record of the same buyer: a Meta click ID, a site visitor ID, her email in Typeform and iClosed, her phone in HubSpot and a Stripe customer ID. Keyring joins them into one person, Sarah P. The match made on the phone number alone waits for approval, then is approved.
Six kinds of record, joined into one buyer
- Ad clicksUTM tags and nine kinds of ad-platform click ID, caught the moment a visitor lands.
- Site visitsPages, scroll depth, video progress and clicks, kept as one visitor across your domains.
- FormsQuiz and application answers, with the visitor ID carried in as a hidden field.
- Bookings and callsBookings from iClosed or Calendly, plus every dial and meeting logged in your CRM.
- PipelineLead status changes and opportunities, in the order your team made them.
- PaymentsCharges, refunds and failed payments from Stripe.
One buyer, with every touch and every payment on the record
Built from the full history of each tool, and kept current as they sync.
Her browsing joins her the moment she fills in the form
The Atlas pixel follows one visitor across your domains, from the landing page to the booking page. When she submits the form, every page she read before she was a name becomes her history.
One visitor clicks a Meta ad and reads three pages across two domains while anonymous. When she submits the quiz she becomes Sarah P., and the three earlier visits join her history. She then books a call.
One buyer, every touch, in order
Each touch carries the marketing signal it arrived with: the UTM tags and the ad-platform click ID. Forms, bookings, calls and the payment sit on the same line, so the sale reads from first click to cash.
Sarah P.'s trail in order: she clicked a Meta ad for Quiz v3 on March 3 and read six pages on two domains, watched the case study on YouTube and took the quiz on March 4, clicked the follow-up email on March 5, clicked a brand search ad on Google and booked a call on March 6, and was closed by Dana on call 4 on March 9, paying $4,320.
Six attribution models, side by side
Every sale is credited under all six at once: first click, last click, linear, time decay, position based and data-driven. Switching is a click, so pick one for the board and keep the others for the argument.
One $4,320 sale credited to Meta ads, YouTube, email and Google Ads under six attribution models: first click, last click, linear, time decay, position based and data-driven. Each model splits the same sale differently.
Track every sale, then send it back to Meta and Google
Meta counts what its own pixel saw. Atlas counts the sales that closed later, on a call, in your CRM or in Stripe, credits each one to the ad that started it, and sends it back to Meta and Google as a conversion.
- Every sale countedCalls, invoices and payments days after the click still land on the right ad.
- Sent back as conversionsMeta and Google receive every sale Atlas attributed, matched to the click.
- Corrected ROASEach ad's spend set against the cash its buyers paid, beside the ROAS Meta reports.
Three ads over the last 30 days, out of 497 sales Atlas attributed. Meta reported Webinar cold as the best, at a ROAS of 2.6, then Quiz v3 at 1.4 and Founder story at 1.1. Corrected on the cash their buyers paid, Quiz v3 returned 6.0 times its $14,400 of spend, Founder story 3.0 times its $10,400 and Webinar cold 0.5 times its $19,600, so Meta's best ad lost money. Atlas then sends the 497 attributed sales back to Meta and Google as conversions, matched to the click.
Your funnel, built from real events
Name it or start from a template, add the steps you track, then pick the tool and filter behind each one: visited this page, submitted this form, booked a call, paid. Atlas counts who cleared each step and recomputes every hour.
Building a funnel named Quiz to close from the Quiz → call template: five steps, each read from one tool with a filter. Visitors from the Atlas pixel on any page, quiz completed from Typeform where the form is Quiz v3, calls booked from iClosed for the Strategy call, calls held from HubSpot where the outcome is Completed, and closed from Stripe where the payment succeeded. Over the last 30 days it flows from 62,480 visitors to 1,236 quiz completions, 98 calls booked, 70 calls held and 27 closed: 61,244 skipped the quiz, 1,138 never booked, 28 did not show and 43 had the call without buying.