40% → 99%
marketing-data completeness, rebuilt from the ground up
Case study · Shipping platform · Sweden
How Sendify went from 40% to 99% trustworthy data
Sendify is a Swedish shipping platform - digital-first, growth driven by marketing. The tracking worked. The problem was that nobody believed the numbers it produced. Here's what was broken, what it was quietly costing them, and what good looks like - so you can spot the same thing in your own setup.
The problem
They weren't short on data. They were short on data anyone trusted
For years Sendify had paid specialist after specialist to improve measurement. Website, backend, ad platforms and analytics each collected on their own, and each reported a different number. Conversions went missing. Customer journeys broke halfway through. When we audited it, completeness on business-critical data sat at roughly 40%.
Here's what that costs and most teams never put a price on it: when the marketing team can't trust a report, they stop deciding from it. Hours go into re-checking numbers instead of acting on them, and the big budget calls quietly fall back to gut feeling. Half-true data is worse than no data, because you still act on it.
What good looks like
The test isn't how much you collect. It's whether every team reports from one number
Most teams treat tracking as a pile of tools - GA4, Tag Manager, the Meta pixel, server-side, ad-platform APIs. Those are only destinations. A foundation you can trust has three properties, and you can check yours against them today:
- Every event enters once, through one point of entry - not collected independently by five tools that each see a different slice.
- Every event is the same shape - consistent definitions, so "a lead" means the same thing in GA4, in Meta and in the boardroom deck.
- Quality is measured, not assumed - you can say your data is 95% complete because something checks it, not because it looks fine.
If two of your tools report two different revenue numbers and nobody can say which is right, you don't have a tracking problem. You have a trust problem - and no amount of new tools fixes it.
What we did
We rebuilt the foundation so every event is collected once, then validated, enriched and trusted
We didn't patch the isolated issues. We rebuilt the data foundation around one rule: every marketing event enters the ecosystem exactly once. Four moves carried the result:
- 01
One point of entry. Website, backend, payments, CRM and analytics now feed one common collection point through a real-time webhook setup - so the data is predictable and every new source plugs in the same way instead of becoming another integration to babysit.
- 02
Enrich and stitch identity once. Before anything reaches a report or an ad platform, each event is enriched (campaign, customer and business context) and identity-stitched - anonymous visitor to logged-in user to backend activity, joined into one continuous journey, privacy-conscious throughout. Done once, not re-done badly by every tool.
- 03
Validate before Tag Manager, not after. Events are cleaned and standardised before they hit GTM server-side, so the marketing team keeps managing tags in the interface they know while every downstream destination receives high-quality, consistent data.
- 04
Monitor quality automatically. Missing parameters, event completeness, volume drops and delivery failures are checked continuously - so broken tracking is caught on arrival, not six weeks later in a report. Quality became measurable instead of a feeling.
On top of that foundation sits a marketing data warehouse Sendify owns: every validated event stored once, as the single source every report, ROAS calc and executive dashboard reads from. Plus custom multi-touch attribution (credit across the journey, not just the last click) and privacy-first server-side activation that feeds the ad algorithms cleaner conversion signals than browser tracking ever could.
The result
From 40% completeness to 95-99% - and a team that decides from it
Business-critical datasets that held about 40% of the expected information now run between 95% and 99% complete, depending on the source. One trusted collection point. One warehouse Sendify owns as the single source of truth, so every department reports from the same number. Automated monitoring that flags issues immediately. Cleaner signals going back to the ad platforms. And because every validated event sits in one structured place, the data is AI-ready by design - the team can ask questions like "which campaigns drive the highest lifetime value?" without exporting and stitching spreadsheets.
- Data completeness up from ~40% to 95-99%
- One collection point for every marketing event, website and backend
- A marketing data warehouse Sendify owns as the single source of truth
- Automated quality monitoring - issues caught on arrival
- Custom multi-touch attribution and privacy-first ad activation
- AI-ready marketing data, faster reporting, decisions made on facts
The takeaway
Most teams think they have a tracking problem. They have a trust problem
Tools don't fix inconsistent data - a foundation does, where every event is collected once, validated continuously and made trustworthy for the whole business. If you can't name the one number your team would bet next quarter's budget on, that's the place to start. It's usually closer to 40% than you'd like.