Every Surface Claims the Credit. The Budget Is Still Due.
Shopify, GA4, Google Ads, Meta, and your email platform each report a different number for the same month. You already know they will never match. That is not the problem. The problem is the budget meeting, where each surface claims credit, the claims sum past 100 percent, and next quarter's allocation has to be decided anyway.
The budget meeting, not the tracking problem
There is a version of this article for someone who just discovered their dashboards disagree. You are past that. You have accepted the disagreement, probably for a while now, and you are stuck one step later: the meeting where money gets assigned to channels. Google Ads reports one revenue figure. Meta reports another. The email platform claims a share of its own. GA4 splits everything a fourth way, and none of it sums to what Shopify says actually happened.
The instinct in that meeting is to keep reconciling. Find the adjustment, the deduplication setting, the attribution model that makes the numbers converge, and then allocate against the reconciled truth. That instinct is the trap, because reconciliation is not available. Not because your setup is broken. Because of what each surface is.
Why the claims sum past 100 percent
Each platform is scored by its own referee. Google Ads reports conversions using Google's attribution, inside Google's window, from Google's vantage point in the journey. Meta does the same with Meta's. Your email platform counts anyone who touched an email inside whatever lookback it prefers. Nobody in this arrangement is lying. Each surface is answering the question "what did I contribute" from a position where its own contribution is the most visible thing in the frame.
A customer who clicks a Meta ad, comes back through branded search, and converts off an email will legitimately appear in three victory columns. Three surfaces each observed a real touch. Each claims the sale under its own rules. The sum passes 100 percent not because someone made an error but because the claims were never drawn from a shared ledger in the first place.
The late-position advantage
The claims are not just overlapping. They are systematically tilted toward whoever stands closest to the sale. A surface positioned late in the journey harvests credit for demand created earlier. Email sits at the end of the journey almost by definition: the address was captured by a purchase or a signup that something else drove. Retargeting is late by design; it only exists downstream of a first visit. Branded search is late by nature; nobody searches your brand name before they have heard of you. All three will report strong numbers, and all three are partly reporting the harvest of work done upstream.
Here is that mechanism in real data. A DTC brand I work with hears from its email platform that email drives 25 to 30 percent of revenue. The figure arrives verbally, in monthly meetings, with no window, no denominator, and no stated attribution model. Neither I nor the client can audit it; it is a vendor self-report, unauditable by design. That is not a weakness in this article. It is your actual situation with most platform claims.
GA4, for the same store, attributes 7.4 percent of GA4-recorded revenue to email, pooled across sixteen months. Month by month, the GA4-attributed email share ranges from 0.0 to 23.9 percent: zero in January and February 2025, peaking at 23.9 percent in November 2025, over BFCM. A surface whose share swings from zero to nearly a quarter by month is not a stable input to an allocation decision.
Now look at the two numbers together. The email peak of 23.9 percent, reached in the single highest-demand month of the year, when paid media was working hardest, nearly equals the platform's own year-round claim of 25 to 30 percent. That is the late-position advantage made visible. Email looks largest exactly when the most demand is being created upstream of it, because that is when there is the most credit to harvest.
One footnote that strengthens the point rather than undermining it: GA4 captured roughly 81.5 percent of platform-of-record revenue across that window. Even the surface I am using to check the claim is not the ledger.
Stop reconciling. Assign jobs.
You cannot make the numbers agree. You can decide what each number is for.
The way out is not a better attribution model. It is a role assignment. Shopify is the platform of record: it says what happened, full stop, and it is never the attribution layer. GA4 is the cross-channel comparison layer: directional, useful for relative movement between channels, and never quoted for a count without its capture rate attached. Google Ads and Meta are optimization instruments: they exist to steer their own bidding, they are good at it, and they are the wrong input for a cross-channel budget decision, because each one is scored by its own referee.
Under this assignment, nothing gets discarded. The email platform's claim, the Meta ROAS, the Google conversion count all remain useful as directional evidence about their own surface. What changes is that none of them gets to be the number the budget is decided against.
How you actually decide
The allocation decision needs a measure the platforms cannot inflate, which means a measure that contains no attribution at all. Blended spend against total revenue from the platform of record, watched over time. Total paid spend across every channel, divided into what Shopify says the store actually did, tracked month over month. Nothing in that ratio can double-count, because nothing in it makes a claim about which touch caused which sale.
The honest cost: this is slower and less satisfying. It will never tell you that a specific dollar produced a specific sale, and if you shift budget between channels, the verdict shows up over months in the blended ratio rather than instantly in a dashboard. What it tells you is whether the whole machine is getting more or less efficient as you change the mix, which is what the budget meeting is actually asking. The meeting was never really asking which channel deserves credit. It was asking whether next quarter's mix should look like last quarter's.
This piece answers the funding half of a question that starts with what revenue is made of. If a large share of your revenue has to be re-acquired every month, as I showed in revenue you have to rebuy, then the budget you are allocating is not discretionary growth spend. It is partly the cost of standing still, and deciding it against surfaces that all flatter themselves is how the cost gets hidden. The stores that get this right are not the ones with the cleanest attribution. They are the ones that stopped asking their dashboards to agree and started asking their ledger whether the machine is improving.
If your budget meeting runs on five disagreeing dashboards, this is the work behind analytics and reporting engagements, and it starts with the ledger. Get in touch.