Case study, an international design brand

Many stores, many tools. One version of the truth.

A design brand selling to consumers and trade customers across several countries had its numbers spread over a dozen tools that did not agree. We brought them into one dashboard, and the harder part: we decided which source to trust for each number.

  • Shopify stores
  • Email accounts
  • Google Analytics
  • Meta Ads
  • Forecast sheet
  • Currencies

One dashboard

  • Orders counted from the stores
  • Visits counted from Analytics
  • Every figure in one currency
  • Bot traffic filtered out
  • Weekly actuals filled in
Before: each answer meant opening every tool. After: one place, and each number comes from the source that actually knows it.

The problem

A dozen tabs for one answer

The brand sells through several Shopify stores, for consumers and for trade customers, in several currencies. Each store had its own sales figures and its own email marketing account. Website traffic sat in Google Analytics, ad spend in Meta, and the forecast in a spreadsheet kept by hand.

A simple question like "how did we do last week?" meant opening every tool, converting currencies by hand and stitching the answer together. Even then, the tools disagreed.

The gap

Same week, two answers

The store was headless: the website people browse is built separately, with Shopify running the catalogue and checkout behind it. Sales reached Google Analytics two ways, through Shopify's own Google integration and through the tracking tags on the website, and the two did not always agree.

In one week the stores recorded seventeen times the orders Google Analytics showed. Any report built on Analytics was wrong before it started.

Google Analytics
1x
The stores
17x

Why it was hard

Deciding what to trust

Connecting the data was the easy part. The work was choosing, for every number, the one source that knows it, and saying so when a number cannot be known.

  • Orders come from the storesAnalytics keeps only what the stores cannot see, visits, product views and add-to-carts.
  • Bot traffic is filtered outAt its peak, automated visits made up more than nine in ten engaged sessions and made engagement look better than ever.
  • One currency for every totalEach figure is converted before it is added up, and the dashboard says so when it uses an estimated rate.
  • Missing is shown as missingA visit without tracking consent reads "not measured", never a misleading zero.

What we built

One private dashboard

One signed-in app that brings together every store, every email account, Google Analytics and Meta Ads, with every figure in one currency.

  • An assistant that answers questionsIn plain words, across every source, and every answer comes from the real data.
  • The forecast fills itself inEach week the actual results land in the team's forecasting spreadsheet beside the plan, without overwriting anything.

The result

One screen, trusted numbers

What used to take an afternoon of copying between tabs is now one screen, and the team can trust every number on it.

  • Every store in one viewRevenue, orders and average basket for every store, market and channel, in one currency, compared with any past period in one click.
  • Questions answered in secondsThe team asks in plain words, like which market slowed down last month, and gets a chart drawn live from the stores, Analytics, email and ads. No answer is ever guessed.
  • Ads measured against real salesMeta's spend, claimed purchases and return on ad spend, per campaign and per ad, sit beside the store orders that actually came from Facebook and Instagram. The team sees what the ads really brought in, not only what Meta says they did.
  • The forecast fills itself inEvery week the finished week's actuals land in the forecasting sheet beside the plan, without anyone typing them.
  • Nothing made upEvery figure is measured or marked as missing, with the source it needs named on screen. There is no placeholder data anywhere.
  • Yours to keepLike everything we build, the code and the data belong to the client.

Is your data spread over too many tools?

Tell us where your numbers live today. We'll show you how they could sit in one place you own.

In Person or Remote. We reply within four hours.