← Back to Help Center

Understanding your numbers · Attribution

Shopify Attribution Models Explained

Last updated: 1 July 2026 · Applies to the Attribution dashboard in Juicy

Shopify doesn't tell you which ad drove a sale — it only records the last source. Juicy tracks the full click journey and lets you switch between six attribution models instantly, so you can see the same Shopify orders through different lenses and answer different questions. The orders never change; only how credit is distributed changes.

How Juicy builds the click journey for each Shopify order

Before any model runs, Juicy works out the real list of ad clicks behind each Shopify order.

1. The pixel on your Shopify store records every visit that arrives carrying an ad's tracking tags.
2. When a Shopify order is placed, Juicy looks at the visits tied to that shopper in the run-up to the order.
3. Repeated loads of the same ad within 30 minutes count as one click. Visits with no ad information are set aside.
4. Each remaining click that maps to an ad or platform becomes a touchpoint on the order.
5. Your chosen model then decides how the order's revenue and profit are shared across those touchpoints.

Juicy attributes on clicks only. Ad views and impressions receive no credit.

Try each model — see how credit shifts

The journey below shows three ad clicks before a purchase. Switch models to see how credit is distributed in real time.

Single-click models

Single-click models give 100% of an order's credit to one touchpoint. Simple and decisive.

Single First click

All credit goes to the very first ad the shopper clicked. In the example above, Meta takes the whole order.

• Use when you want to see which ads introduce new customers and start journeys.
• Strength — surfaces top-of-funnel demand creation that last-click reporting routinely undervalues.
• Watch out — ignores everything after the first click, including the ad that actually closed the sale.

Single Last click

All credit goes to the final ad before purchase. Google takes the full order in the example.

• Use when you have short purchase cycles or want to know what converts.
• Strength — focuses on the ads that finish sales; useful for lower-funnel budget decisions.
• Watch out — over-credits demand-capture channels like Search and hides the ads that first brought the customer in.

Single Last click per platform

Each platform receives full credit for its own most recent click. In the example, Meta, TikTok, and Google are all credited with the order — because each was the last click on its own platform.

• Use when you want to compare each platform against its own native dashboard (Ads Manager, Google Ads, etc.).
• Strength — every platform sees its true last-click contribution; no platform is starved by another closing the sale.
• Watch out — one order can count under multiple platforms, so totals across platforms can exceed your real order count. Use a single-credit model when you need a total that reconciles with your shop's actual orders.

Weighted models

Weighted models share an order's credit across every click in the journey, each according to a different rule.

Weighted Linear

Credit is split equally across every click. With three clicks, each gets one third.

• Use when you want a neutral, unbiased baseline that respects the whole journey.
• Strength — no single click is over-valued or under-valued.
• Watch out — spreading credit evenly can flatten the difference between an ad that did real work and one that was just along for the ride.

Weighted Position based

The first and last clicks each get 40%. The remaining 20% is split equally across any clicks in between. With just two clicks it is 50/50.

• Use when you believe the ads that open and close a journey matter more than the middle.
• Strength — captures both discovery and conversion in a single view.
• Watch out — the 40/20/40 split is a fixed rule of thumb, not a measurement of true influence. Treat it as a lens, not a verdict.

Weighted Time decay

More credit goes to clicks that happened closer to the purchase, on a roughly seven-day half-life. A click the day before the purchase carries more weight than one from a week ago.

• Use when recent touches tend to drive your sales, or purchase cycles are short.
• Strength — reflects momentum, giving the closing stretch its due without ignoring earlier clicks entirely.
• Watch out — deliberately underweights top-of-funnel ads. Pair it with first click if you also want to see what starts journeys.

Which model should you use?

There is no single correct model — only the right model for the question you are asking right now.

Your question	Best model
Which ads bring in new customers?	First click — shows which ads open journeys and create demand.
Which ads close sales?	Last click — shows what converts; good for bottom-of-funnel work.
How does this platform compare to its own dashboard?	Last click per platform — lines up most closely with each network's native reporting. Read per platform, not as a shop-wide total.
A balanced, neutral read across the full journey?	Linear — treats all clicks equally. A good everyday baseline.
Which ads open and close — but not the middle?	Position based — leans on the first and last clicks.
Which recent ads pushed the sale over the line?	Time decay — favours the clicks nearest to purchase.

A practical habit: keep one reconciling model (first click, last click, or linear) as your everyday view, and switch to last click per platform only when benchmarking a single platform against its native reporting.

How direct and organic traffic behave

Not every visit arrives from an ad. Someone might type your store address directly, follow an organic search result, or click a link with no tracking tags. Juicy cannot tie those visits to a specific ad.

In every model, a visit only takes part if Juicy can match it to an ad or an ad platform. Visits with no ad information are set aside when the journey is built. If an order has no ad clicks at all, it falls back to the order's own recorded source — a purely organic order is not quietly handed to a paid channel.

If a channel you expected to see is missing or lower than expected, the usual cause is missing tracking tags on the ad links. Check that your links carry their tracking parameters.

Frequently asked questions