Why your attribution tools disagree
Run the same date range through Consequential, Google Analytics, Triple Whale, and Shopify and you get four different pictures of where your sales came from. One tool calls an order paid, another calls it organic, and the revenue each assigns to Meta or Google rarely matches. This page explains why that happens, why it doesn’t mean one tool is broken, and how to compare them without chasing numbers that were never meant to agree.
Each tool measures a different thing
Section titled “Each tool measures a different thing”The tools disagree because they answer different questions from different data:
- Shopify is the source of truth for what happened: which orders were placed, for how much, what was refunded, and your true cost of goods. It isn’t an attribution engine. Shopify records a
referring_siteand a landing source for an order, but a shopper who clicks an ad on Monday and returns directly to buy on Thursday often lands in Shopify as Direct or Organic, because that’s how they arrived at checkout — not how they first discovered you. - Google Analytics measures sessions and the paths between them. It attributes an order to the channel of the session it credits, under its own model (GA4 defaults to a data-driven, last-non-direct view). If GA4 saw a click on a paid link, it can call the order paid even when Shopify calls it organic.
- Triple Whale and other attribution tools reconstruct the journey from pixel data and their own modeling. Each vendor’s model is different, so each reconstructs a different answer.
- Consequential classifies every order into a channel from the click signals on the visit that led to it — the UTMs on the landing URL, the click IDs that Google and Meta add to ad links (
gclidandfbclid), and the visit source Shopify resolves for the order — then assigns credit under the attribution model you select. It works from click signals only. It doesn’t use raw referrer hostnames or view-through data.
None of these is wrong. They’re measuring different events with different rules, so they produce different numbers for the same orders.
The reasons the numbers diverge
Section titled “The reasons the numbers diverge”Attribution model
Section titled “Attribution model”The model is the rule that decides which touchpoint gets credit. Last-click gives all credit to the final touch before purchase; first-click gives it to the first. Switch a report from last-click to first-click and the same orders move between channels without a single order changing. Consequential lets you pick the model on the Acquisition Dashboard; GA4, Triple Whale, and each ad platform apply their own defaults. Comparing a last-click number in one tool against a data-driven number in another compares two models, not two tools.
Click credit versus view credit
Section titled “Click credit versus view credit”Ad platforms count view-through conversions — someone who saw an ad, didn’t click, and bought later. Click-based attribution doesn’t. Consequential, like most independent tools, credits clicks only. Meta’s own reported revenue is almost always higher than any click-based tool because it’s claiming credit for views that other tools never counted.
Attribution window (in the other tools)
Section titled “Attribution window (in the other tools)”Ad platforms, GA4, and Triple Whale each credit a touchpoint only if the purchase falls inside their lookback window, and those windows differ — a 7-day window and a 28-day window assign the same order to different channels. Consequential doesn’t apply a configurable lookback window: it matches an order to a campaign by the click IDs and UTMs carried on the visit. So a difference here is almost always the other tool’s window at work.
Deduplication and match rate
Section titled “Deduplication and match rate”Browser pixels and server-side events both report the same purchase, and each tool deduplicates differently. Some orders can’t be matched back to any click or view data at all. On the Acquisition Dashboard, Match rate shows the share of orders a given ad provider could tie to its own data — the unmatched remainder is exactly where tools diverge most.
Tagging gaps
Section titled “Tagging gaps”When a link reaches your site without clean UTMs, each tool guesses the source differently from the referrer, and those guesses rarely agree. Missing or inconsistent tags are the most common reason one tool sees paid where another sees direct or organic. See Why you see (not set) in your Google Analytics reports.
How to compare tools correctly
Section titled “How to compare tools correctly”- Compare like for like. Line up the same attribution model, the same attribution window, and the same date range before you call a gap a problem. Mixing a last-click tool with a data-driven one guarantees a mismatch.
- Compare whole periods, not partial ones. A Consequential window ends yesterday on your store’s timezone; a tool showing today’s partial data against a full day elsewhere always looks off. See Date ranges and presets.
- Anchor totals to Shopify. Use Shopify for total sales, refunds, and margin — the facts. Use Consequential, GA4, or Triple Whale for how those sales happened — the attribution. Expect the totals to tie out and the channel splits not to. Consequential reconciles every platform-claimed conversion against the orders that actually settled in Shopify, so a platform cannot credit itself with a sale that never happened.
- Read platform-reported numbers as a separate column. On the Acquisition Dashboard, every revenue and ROAS figure comes in two forms: what the ad platform reports, and what Consequential attributes under your model. Treat them as two measurements, not as one number and an error.
Why Consequential’s number is the one to trust
Section titled “Why Consequential’s number is the one to trust”The tools all measure something real, but they aren’t equally trustworthy for the decision you care about — how much revenue a channel actually drove. Consequential answers that one question:
- It reconciles to settled orders. Every platform-claimed conversion is checked against the orders that actually settled in Shopify. Ad platforms grade their own homework and report the number that flatters their spend; Consequential only counts a conversion that ties to a real, settled order, so a channel can’t inflate its own results.
- It’s independent. Consequential doesn’t have ad budget to defend, so it has no reason to over-credit Meta or Google the way each platform over-credits itself.
- It uses hard click evidence. Credit follows the click IDs and UTMs carried on the visit (
gclid,fbclid, and your tags), not a modeled guess about views a shopper may never have seen. That makes the number conservative and repeatable. - It covers paid and organic in one place. The Acquisition Dashboard classifies every order — paid, email, organic, direct, and the rest — against the same settled-order base, so you compare channels on one consistent ruler instead of stitching four tools together.
Treat Shopify as the truth for what sold and Consequential as the truth for how it sold. The other tools are useful cross-checks, not the number you optimize against.
What you can do about it
Section titled “What you can do about it”Most of the gap between tools comes from signal you can improve. Work through these in order:
- Tag every external link. Give every ad, email, and social link a
utm_source,utm_medium, andutm_campaign, using one consistent vocabulary. This is the single biggest lever on attribution quality. See Understanding UTM tracking. - Fix the tagging gaps you already have. Check how much revenue is arriving with no source attached, and resolve the
(not set)and unattributed causes at the source. - Turn on order-side attribution. Under Settings → Acquisition, enable order-side attribution so the dashboard shows attributed revenue, CAC, and the new-versus-returning split instead of platform-only numbers.
- Reduce signal loss with the Pixel. Set up server-side tracking so purchases reach the ad platforms even when browser pixels are blocked — this tightens both your reporting and the platforms’ optimization.
- Keep every source connected. A lapsed connection or a broken pixel shows up as a falling Match rate. Check your connected sources when numbers drift.
- Pick one model and compare consistently. Choose last-click or first-click, use the same model and date range across every tool, and judge trends over whole periods rather than single days.
When divergence is a real problem
Section titled “When divergence is a real problem”Expected divergence has a cause you can name — a different model, window, or view-through policy. Investigate further when:
- Shopify’s total sales do not tie out to the tools (a data or connection issue, not an attribution one).
- A large share of orders lands in Unattributed or Direct across every tool at once — usually a tagging or signal problem at the source.
- The Match rate for a provider drops suddenly, which points to a broken pixel or a lapsed connection rather than a modeling difference.
For these, start with your tagging and connections, not with the attribution model.
Related
Section titled “Related”- Signal loss, and why it wastes ad spend — the broken feedback loop underneath most reporting gaps.
- Attribution models — how each model assigns credit.
- Understand the Acquisition Dashboard — where attributed and platform-reported numbers sit side by side.
- Understanding UTM tracking — the one acquisition signal you control.
- Why you see (not set) in your Google Analytics reports — what tagging gaps look like in GA4.