Skip to content

Premium fashion (DTC)Client work

Three Tools, Three Answers: Reconciling Triple Whale, Meta and Google for a Multi-Currency Store

How a multi-currency EU fashion brand reconciled Triple Whale, Meta and Google, and got 37 of 37 orders to Meta once, with the right amount and currency.

The brief

Context

A premium fashion and footwear brand in the EU ships worldwide from a multi-currency Shopify store. Its leadership team used Triple Whale for attribution, Meta for paid social, and Google for search, and the three tools told three different stories. The brand became a client in September 2026 and is on a monthly plan.

Sounds familiar?

Symptoms

The situation Premium fashion (DTC)
Tick what applies to you

Tick the lines that describe your case.

Describe my task

The investigation

Diagnosis

I ran a gated audit: a preview edition first, the full edition after payment, and a corrected v2 once Meta and Google access was granted.

  1. The audit covered the nine web pixels, the theme app blocks, consent gating per sender, the Meta dataset quality, and the source of Google’s conversions. The findings went into a log with an evidence file per claim.

  2. The root finding was that Google’s only primary conversion came from a Triple Whale click-conversion upload, not from the storefront. Google was not “taking all the sales” because the store reported them; it was taking one third-party upload at face value.

  3. The audit also produced an app-pixel ownership map (a Google channel app, a third-party Meta Conversions API app, Triple Whale, Klaviyo and replay tools) and a browser-versus-server value and currency parity check.

  4. One number looked like a defect and was not: Meta showed “35 of 37” purchases. I dug in and closed it as an aggregation-view difference — the same orders, counted by source view and by event view.

Handover

What I built

Delivery noteDelivered Sep 2026

  1. One server route for Meta. I installed Fixel Pixel as the Meta server leg through collaborator access, switching the app embed on with a small, backed-up theme change. Meta Conversions API went live with value and currency matching the store’s presentment currency.
  2. A dedup key in place of a hope. Browser fbp and click fbc coverage are now tracked as numbers over time, not assumed, and the server leg shares one identity so an order is counted once.
  3. A watch on the uploads. An always-on host job checks that Google’s conversion upload stays fresh and alerts on any gap; a second hourly job checks the storefront and alerts if a theme push drops the embed.
  4. Order-level reconciliation. Every order is matched against what each platform received, with the amount and currency checked, so a mismatch shows as a named order rather than a monthly argument.

Before → after

Results

Client work Done for real clients. Client details are anonymised.

  • Orders delivered to Meta exactly once, right amount and currency

    Sep 2026

    37/37

    Before
    disputed — Triple Whale credited Meta with more sales than Meta showed
    After
    37 of 37 orders → 37 purchases, across 9 currencies

  • Click id coverage (fbc) on InitiateCheckout

    Sep 2026

    11.8% to 78.9%

    Before
    11.8%
    After
    78.9%

  • Browser id dedup key (fbp) on PageView

    Sep 2026

    0.06% to 65.9%

    Before
    0.06%
    After
    65.9%

  • Meta Purchase Event Match Quality

    Sep 2026

    7.7 to 8.1

    Before
    7.7
    After
    8.1
The full results table
What changedBeforeAfterRead on
Orders delivered to Meta once, amount and currency matchingdisputed37 of 37, across 9 currenciesSep 2026
Click id (fbc) on InitiateCheckout11.8%78.9%Sep 2026
Browser dedup key (fbp) on PageView0.06%65.9%Sep 2026
Meta Purchase Event Match Quality7.78.1Sep 2026

Notes on the numbers

The first preview edition contained a known error, which I corrected in the full edition. That is worth saying plainly: an audit that never corrects itself is not being checked.

A note from Daniilbefore you decide

What this means for a similar business

When an attribution tool disagrees with the platform it reports on, do not start by rebuilding everything. Start by finding who actually feeds the conversion. In this case the tool that claimed to report on Google was also the only thing uploading Google’s conversion, so the number was circular.

Two habits keep this honest. First, track identity coverage as a number — click id and browser id per event — so a drop is visible before it becomes a story. Second, check value and currency parity on a multi-currency store, because a correct event with the wrong currency matches nothing.

— Daniil

Situations this case answers

All situations
Read another case

Silent Zeros: When a Reporting Pipeline Says Success and Delivers Nothing

I built a lead-gen reporting cockpit, repaired a silent n8n failure, and separately built scheduled Google Ads imports from call-tracking data.
Daniil Maximkin

Hi, I’m Daniil.

I work with you from defining the problem to implementation and handover. You talk to the person who does the work. I work in English and Russian.

My story My work

Have a similar task?

Describe your task

The first answer is free, within one working day. Or write directly: next@taskfordaniel.com