Attribution, six models
The rule-based models GA4 removed, side by side and worked out from each visitor's full path. Under every model each purchase's credit adds up to exactly one.
Keep your tag, your property and your BigQuery export. Siftraq adds the reports GA4 doesn't have, reading your data where it already lives. Nothing to migrate, and nothing to hand over.
| Channel | Share | Orders credited |
|---|
Siftraq works from the pieces you already have and leaves each one where it is.
| Piece | Where it is | Who controls it |
|---|---|---|
| Tracking tag | Your website, unchanged | You |
| GA4 property | Your Google account | You |
| Raw event history | Your BigQuery, in your Google Cloud project | You |
| Query costs | Your Google Cloud bill, at Google's rates | You, with a ceiling per query |
| Reports | Siftraq, worked out on request | Your team's sign-ins |
Each report is worked out from the full, unsampled event export, the same data GA4 itself collects, read with the care of an analytics engineer.
The rule-based models GA4 removed, side by side and worked out from each visitor's full path. Under every model each purchase's credit adds up to exactly one.
Every visit in order, event by event, identified only by GA4's own anonymous IDs.
Sessions, engagement and the purchase funnel, counted from the raw export.
Where visitors leave, which GA4's own reports no longer make easy.
/checkout/shipping/kitsEvents as they reach GA4's streaming export, checked every 30 seconds.
Connect Claude or ChatGPT and ask about your sessions, landing pages and exit pages. Siftraq works out the answer from your own export.
Sample data. Also: ecommerce and products, and audience by country, device and language.
For owners, marketers and the agencies that help them: anyone who relies on GA4 but doesn't write SQL.
Attribution, the purchase funnel and the visits behind each order. Shopify stores can also check GA4 against their real orders.
Sessions, landing pages, bounce and exit pages for sites where a form, not a cart, is the goal.
Engagement and the paths visitors take before they sign up or get in touch.
One sign-in for every client, each client's data in that client's own Google Cloud project.
Session, engagement and exit reports work for any site that exports GA4 to BigQuery. Attribution reports work on purchases today.
You'll need a Google account that can see your GA4 property, and a Google Cloud project for the BigQuery export. Siftraq walks you through it.
Only your name and email are shared at this step. Siftraq asks for anything more one permission at a time, when a step needs it.
Siftraq lists the GA4 properties you can see and shows which ones already export to BigQuery. If none does, Siftraq can create the export for you, once, if you ask.
Reports are worked out from your export as you open them. GA4's export also builds your history from that day on, for as long as you keep it.
The BigQuery export is a free feature of GA4. Storing the data and querying it are billed by Google to your own project, at Google's published rates.
BigQuery includes a free monthly allowance for storage and for queries. A small business's GA4 export often stays within it. Google's own pricing page has the current figures.
Every query Siftraq runs carries a size ceiling and reads only the days it needs, so a mistake can't turn into a large bill. Reports never write to your project.
If you have a Google Cloud project, Siftraq creates the GA4 export in it with one click. If you don't, Siftraq shows you the few steps to make one, and checks each one for you.
Siftraq will keep your GA4 export in a BigQuery project it runs and bills for you. The history in it stays yours: when you'd like it in your own Google Cloud, Siftraq copies it over and points your GA4 export there.
The reports work from what GA4 already collects. Next, Siftraq takes over the collection itself, and sends what it sees on to GA4, Google Ads and Meta.
Siftraq sets up your server-side tagging and keeps it running, in your own Google Cloud project at Google's own rates. There's no separate tag-hosting service to sign up for or pay, and no cloud engineering on your side. Collection moves to your own domain, where ad blockers and browser limits lose far less.
From the same server, purchases and leads go to Google Ads and to Meta's Conversions API, so their bidding learns from the conversions the browser missed.
Orders whose thank-you page never loaded, sent to GA4 from your store's records and checked so none is counted twice.
A plain-language check that your consent banner and tags fire in the right order.
Ready-made tracking for lead forms, bookings and sign-ups, not just online shops.
If you'd rather not depend on GA4 or run any of this yourself, the Hosted Edition is Siftraq for Shopify stores, with its own pixel and nothing to set up in Google Cloud.
Siftraq is opening to a small group of businesses and agencies first. Leave your email and we'll write when there's a place for you, with a short call to set it up together if you'd like one.