Usermaven With Causality Engine: Running Both on One Store

The short answer: add up Usermaven's Influenced column for a month and the channels claim about twice the orders Shopify took. The column is built that way, and it is not the number to budget on. Usermaven, the attribution and analytics platform with its own tracking script, credits a conversion goal by one of seven built-in models over a lookback you set between 30 and 180 days (attribution guide). Keep it, and keep GA4 running as well: Causality Engine, a causal attribution tool, reads only GA4's Attribution paths export, never Usermaven's data. Running both costs one extra file and nothing else to start, since the first finding is free on the start page.

Disclosure: I founded Causality Engine and own it, and this page recommends it. Every claim about it links to Causality Engine's own pages.

Before the two tools sit in one report:

1. Divide the Influenced total by the goal count. For example, Influenced conversions summing to 3,261 across channels over 1,540 goal conversions is 3,261 ÷ 1,540 = 2.12 channel credits per order. Any figure built on that column double-counts by design.
2. Leave the GA4 tag where it is. The read uses the GA4 history the store already has; Usermaven's script sits beside it in theme.liquid without replacing it.
3. Agree one model before comparing. Pick Last Touch Non-Direct in Usermaven for the side-by-side, since GA4's session channels follow a last-click rule too (scopes page), then let the read add what neither rule can.

Two attribution layers, one store

What Usermaven counts as an order, and who gets it

Nothing is attributed until a conversion goal exists. The attribution guide walks through Create Conversion Goal: a name, the event that counts, a static value or a dynamic one (dynamic only for custom events), and a box for unique conversions. Set the default currency under Workspace settings, Miscellaneous first.

On Shopify the event comes from two installs, per the Shopify guide. The tracking script goes before the closing head tag in theme.liquid, with autocapture on. The purchase needs a second step: a custom pixel under Settings, Customer events, built from your workspace API key and server token, that sends Checkout Completed with total revenue, products and quantities to Usermaven's server endpoint. The guide warns that Shopify's sandbox can make some page reports show a /web-pixels@ path instead of a storefront URL.

Credit is then shared by the model you pick, over the lookback you pick. Three report parts read the result differently:

• Influenced counts every visitor from a channel who later converted, with their conversions and revenue. Each channel a buyer touched counts that buyer.
• The model columns split each conversion once, by rule.
• Conversion Paths and Days to Convert show the sequences and the delays behind the credit.

The guide also charts channel performance under several models side by side. That chart is the tool's own evidence that credit is a convention.

Where GA4 fits when Usermaven is installed

Usermaven sends nothing to GA4, and GA4 sends nothing back. Both read the same UTM tags on arrival, and each files the visit by its own channel rules: Usermaven by its Custom Channel Mappings, GA4 by Google's default channel group. If a campaign shows up as one channel in Usermaven and Unassigned in GA4, the tag is the problem; GA4 uses Unassigned when no rule matches.

That GA4 file is what makes the read possible. Its start page gives the export as Advertising, Attribution paths, a date range of 12 months or more, your purchase key event, then Share this report, Download file, CSV.

Worked example: 1,612 orders, three readings

For example, with figures made up to show the arithmetic and no shop behind them: Shopify books 1,612 orders in September. Usermaven's Checkout Completed goal records 1,540, and GA4 1,498 purchases.

The last-touch column adds up: 431 + 389 + 371 + 262 + 87 = 1,540. Influenced reaches 3,261, which is 3,261 - 1,612 = 1,649 more than the shop booked. Paid Social's Influenced 702 against its last-touch 431 says 702 - 431 = 271 buyers touched Paid Social and were credited elsewhere. GA4 puts 288 orders under Direct, against Usermaven's 87, because the rules for direct visits differ.

All three columns describe the same 1,612 orders. None of them says what would have sold without Paid Social.

The read starts from GA4's 288 under Direct. Per its how-it-works page, it splits that row back to the channels that sent those buyers, sets each channel's caused share beside a last-click rebuild, grades it with a data-health score from 0 to 100 and adds a next step to test in one region. Where its Paid Social share sits well below Usermaven's 431, the disagreement is the finding: test that channel in one region before the budget moves.

Free first: Usermaven's own chart, then a pause

The first check is inside the tool you pay for already. In Attribution, Channel/Source, select First Touch, Last Touch and Linear together and read the performance chart: a channel whose bar changes most between models is the one whose credit is most a matter of rule. Then test it by hand, since the guide documents no holdout. Pause Paid Social in one region for four weeks and compare Shopify orders there with the regions that kept it.

When that leaves more than one channel open, Causality Engine, which I founded, reads the GA4 export from the start page: the preview costs nothing and is computed in your browser, and the full read is €99 once, refundable for 30 days (pricing, checked on 11 November 2026).

Where running both stops paying

• Without GA4 there is no read. Usermaven's events cannot be uploaded; a store that removed GA4 has no file.
• Ads, content and paths stay in Usermaven. The read works per channel and never names an ad, a page or a sequence.
• Thin history or spend. The one-off read page says not to buy on thin analytics history, and a channel under 1% of credit on both methods is called too small to act on (agents page).

Usermaven's ecommerce page opens on a sample that reads First Touch over the last 30 days, sets its attributed ROAS beside a platform's own and recommends which channel to scale, while its pricing page sends every buyer to a demo for a figure. The tool is sold as the steadier figure beside the platform's, and a figure picked from seven rules is still a rule's answer, not a test's.

See also

• Usermaven: models, privacy modes and the missing price list
• How to use Usermaven: HubSpot deals traced to paid campaigns
• Usermaven on the marketplace: Usermaven's affiliate program terms
• Causality Engine: the read, its price, its refusals
• How to use Causality Engine: from the CSV to a regional test
• Causality Engine on the marketplace: affiliate terms, given as the owner's
• Causality Engine vs Usermaven: choosing one instead of running both
• Usermaven vs Google Analytics: its script against GA4's

FAQ

Does Usermaven connect to Causality Engine?

No. Neither tool sends anything to the other. The read takes the CSV you download from GA4, as its integrations page describes, and Usermaven keeps its own events.

Which Usermaven model should I set beside the read?

My recommendation: Last Touch Non-Direct, because its rows add up to the goal total and it sits closest to GA4's last-click view. Then compare First Touch for the channels where the read and Usermaven disagree (attribution guide).

Can Usermaven see Shopify revenue without the custom pixel?

The Shopify guide puts checkout and revenue tracking in a separate custom pixel under Customer events. The theme script alone records visits and clicks.