Kit With Causality Engine: Check the Email Attribution Table

The short answer: Kit says the cart sequence recovered the sale. Meta says its ad did. Shopify shows one order. If you run Kit, the creator email platform that used to be ConvertKit, on a store with purchases in GA4, read Kit's figures as claims made under Kit's rules. Its abandoned checkout template waits 1 hour, sends 3 emails, waits 1 day, and tags everyone who then buys "Abandoned checkout recovered" (Kit + Shopify). Causality Engine, a causal attribution tool that works from a GA4 export, sets that claim beside every other channel's, and its first finding is free at app.causalityengine.ai/start.

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

What to settle before the tag count is quoted:

1. The share the tag describes. For example, 500 abandoned checkouts enter the template and 40 leave with the recovered tag: 40 ÷ 500 = 8%. That 8% bought after an email. How many bought because of one is a different number, and the tag cannot give it.
2. The switch you already own. One toggle in Kit's Advanced settings, "Automatically append UTM parameters to email links", marks every Broadcast and Email Sequence link so GA4 can file the visit (Automatic UTM parameters).
3. Three figures for one week. Kit's count, GA4's last-click count for Email, and the orders the store booked, in one row.

Kit and the GA4 read in one table

Three ways Kit scores itself

Clicks. The Broadcasts dashboard defines Clicked as the number of recipients who clicked at least one link. Kit's troubleshooting page adds that strict mail security can click links before any person does (automatic opens and clicks).

The recovered tag. In the Shopify app's template, a Product purchased node pulls in anyone who buys a Shopify product and applies "Abandoned checkout recovered"; people who finish the sequence without buying get "Abandoned checkout lost". The Kit + Shopify article says the two tags let you "identify how many sales you've recovered". The rule is purchase after email. It asks for no click, and the waits are steps you can edit.

Email Attribution. Insights, a Creator Pro feature, adds a table that works "through the last-click attribution model", which the Engagement dashboard article defines as what a subscriber last interacted with before buying. The Sales dashboard beside it reports Revenue by Source on a first-touch model. Two tables, two rules, and neither article names a time window or a setting that changes the model. Insights needs its own pixel on the site (Setting up Insights), and the Sales article describes "sales sold through Kit", so confirm Shopify orders appear there before quoting it.

Getting Kit's clicks filed as Email

Go to Settings, then Advanced. Kit lists three tracking options: Subscriber ID, UTM codes and Hashed subscriber email (Global advanced tracking). With UTM codes on, source and medium are fixed at kit and email. Google files a session under Email when source or medium is email, e-mail, e_mail or e mail (default channel group), so Kit's defaults pass untouched.

What Kit's article warns about:

• Send time decides. Parameters are added when the email goes out. A Broadcast sent before the toggle was saved carries none.
• Preview emails are bare. To inspect a link, send the real email to yourself.
• A hand-typed parameter wins. Kit fills in only the missing ones. A link you once tagged utm_medium=newsletter still says newsletter, and Google has no Email rule for that word.

With the toggle off, Google's page on (direct) / (none) traffic describes the result: links without UTM parameters lose their traffic source information.

One launch week, five counts: a worked example

For example, with invented numbers that illustrate the arithmetic and describe no real store: a skincare shop sends a five-email launch Sequence through Kit and books 180 orders that week.

The three sellers of attention claim 54 + 88 + 67 = 209 orders. The store shipped 180, so at least 209 - 180 = 29 claims double up. Kit and GA4 disagree with each other too: 54 - 31 = 23 orders where Kit saw an email click and GA4 saw a later click from somewhere else.

None of the four is lying. Each applies its rule to what it can see. Beside a table like this, a causal read places each channel's share of what it caused next to a last-click rebuild from the same file, a data-health score from 0 to 100 for every channel, and a next step (how it works). If the read puts Email above or below Kit's 54, move no budget on either figure yet. Hold back a segment first.

Where the read stops and Kit carries on

• The read needs a purchase key event in GA4, and the export steps on the start page ask for 12 months of history or more.
• A small list can fall under the line. A channel below 1% of credit on both methods is called too small to act on (what causal attribution costs), and where spend is too small for a signal the home page says the product declines.
• The read is per channel. It will not rank email three of a Sequence against email five, and it holds no purchase history per subscriber. Kit does both (Purchases).

Insights answers "Which emails drive the most sales?" only on Creator Pro, and it answers by last click, the rule under which the closing email of any launch looks like the hero. The report kindest to email is the one sold on the higher plan.

The free check: a held-out Tag, then the export

Kit documents the exclusion, not the shuffle. Export your subscribers, randomise them in a spreadsheet, and tag one in ten "holdout". On each promotional Broadcast, add a filter group for that Tag and set it to "Matching: none" (excluding subscribers from a Broadcast). After four sends, count buyers in both groups from the store's orders by customer email.

For example, 8,100 mailed subscribers produce 162 buyers and 900 held-out subscribers produce 12: 162 ÷ 8,100 = 2.0% against 12 ÷ 900 = 1.33%. The emails added (2.0% - 1.33%) × 8,100 = 54 orders, a third of the 162 they could claim.

Kit's A/B test is no substitute. It sends two versions to two 15% groups and picks the higher click rate, so everyone still gets an email (A/B testing email content).

The holdout costs nothing and settles the email question on its own. For the question across channels, download GA4's Attribution paths report and take the free first finding on the start page; the preview runs in your browser. Causality Engine, which I founded, then runs the full read on that file for €99 once, excluding VAT, refundable within 30 days (pricing, checked on 11 November 2026).

See also

• Kit: what Kit is, who it suits, the complaints
• How to use Kit: routing product signups into sequences
• Kit on the marketplace: Kit's affiliate program terms
• Causality Engine: the product, its prices, its limits
• How to use Causality Engine: export, free finding, full read
• Causality Engine on the marketplace: its affiliate program terms
• Kit vs beehiiv: which newsletter tool fits a seller
• How to measure email incrementality: the holdout method in full

FAQ

Is there a Kit app or connector for Causality Engine?

No. Nothing is installed in either product. The read starts from a CSV you download from GA4, and the integrations page says it covers every channel that export records. Kit appears there as Email once its links carry utm_medium=email.

Does the "Abandoned checkout recovered" tag prove the email saved the sale?

It proves the purchase came after the person entered the automation. The template tags anyone who buys a Shopify product after receiving one of the three emails, clicked or not. Only a holdout shows how many would have bought anyway.

Is Kit's automatic UTM tagging on by default?

The article reads as off: its first heading is "How to enable automatic UTM parameters", and it tells you to switch the toggle on in Advanced settings and save. Check yours before the next send.