tl;dr
Klar runs a first-party tracking script under your own domain — so it can't be blocked by ad blockers or browser privacy changes
To stitch sessions across devices and browsers, Klar uses 25 different identifier dimensions — and maps them probabilistically rather than doing a hard match on unreliable signals like IP addresses alone
Zero-party data from discount codes and post-purchase surveys is layered on top to enrich the journey further
The result is a set of high-probability user journeys, not 60% guesses — fed directly into your profitability and retention data in Klar
All data stays in the EU — Klar is certified to ISO/IEC 27001:2024
Attribution data is only available from the day your pixel went live — with a 90-day GA4 fallback window before that
Why attribution is hard
User journeys are fragmented. A customer might discover your brand via an influencer story on their phone, browse on a laptop the next day, and convert via branded search a week later — three devices, potentially three different browsers and IP addresses. Most tools either rely on cookies (increasingly blocked) or over-simplify with hard IP mapping, which introduces errors at scale. Klar's approach is to collect as many signals as possible and weight them by reliability.
The first-party tracking script
Instead of firing under a shared third-party domain, Klar generates a script that runs under your own subdomain (e.g. 56789.yourshop.eu). Ad blockers and browser privacy tools block third-party tracking domains by default — a first-party script on your own subdomain doesn't get caught. It's also the durable long-term approach as third-party cookie restrictions continue to tighten.
What Klar collects
Once the script is live, it captures three types of signals:
Marketing signals — UTM parameters and ad click IDs (fbclid, gclid, ttclid)
Identity signals — user IDs, session IDs, IP address, device fingerprint, user agent, email address at checkout — 25 dimensions in total, used to stitch sessions together
Behavioural signals — time on site, add-to-cart, purchase events
How Klar stitches sessions together
With this data, Klar builds unified user profiles across devices and touch points. A simple example: a customer clicks an influencer post on their phone, then buys via branded search on their laptop at home. Klar links the first sessions via user ID and the laptop session via IP address — one unified journey: Influencer → Branded Search → Purchase.
Not all signals are equally reliable. Email address is a hard match. IP address isn't — it can be shared by many people on the same mobile network. So instead of mapping on IP alone, Klar analyses the ISP, country, and network type to build a probability curve that dynamically determines how long an IP can be trusted as an identifier. Multiple identifiers (e.g. IP + user agent) are combined to produce tighter curves and higher-confidence matches — well above 60%.
Historical data & the 90-day fallback window
Attribution requires a tracked customer journey for each order. Klar only starts building real journeys from the day your pixel goes live.
For orders placed before the pixel was installed, Klar uses a fallback: it pulls the order's source/UTM information from GA4 and treats that as the attribution signal. This is not a multi-touchpoint journey — it's a single last-click data point from GA4, used as a best-effort substitute.
This fallback only covers the 90 days immediately before your pixel's first tracked sale. Orders older than that have no journey data and carry no attributed revenue in Klar.
Example: If your pixel's first tracked order was on August 12, 2025, Klar will show GA4-based attribution for orders from May 14, 2025 onward. Everything before May 14 will have no attribution data.
What this means in practice: If your attribution report shows a gap or an unusually high "Direct / Unattributed" share in the early months, this is expected — those orders predate the pixel or fall outside the 90-day fallback window.
Zero-party data enrichment
Post-purchase surveys and discount codes layer on top of the first-party tracking to capture what click-tracking misses:
If a customer says "TikTok" and a TikTok touchpoint exists → that touchpoint gets boosted in the attribution model
If they mention a channel with no tracked touchpoint → Klar injects a touchpoint into the journey (e.g. Pinterest at the start, or an influencer at the end)
This means your attribution reflects the real customer journey, including dark channels that never generated a trackable click.
Data security & compliance
Tracking customer journeys means handling sensitive data. Here's how Klar treats it:
ISO/IEC 27001:2024 certified by GUTcert — the internationally recognised information security standard. Full certificate at trust.getklar.com
EU hosting — data processed and stored exclusively in EU data centres, never leaving the EU
Encryption — TLS 1.2/1.3 in transit, AES-256 at rest
Access control — RBAC, least-privilege principles, MFA on all internal systems
GDPR compliant — DPA available at trust.getklar.com/dpa
Want to understand how eCom measurement actually works?
This is Klar's free video course on Marketing Measurement - covering platform attribution, MTA, MMM, incrementality and unified measurement.