Know the source behind every lead and customer.
Not which platform claimed it — which channel actually earned it. One first-party dataset, nine ways to divide the credit, and no re-tagging when you change your mind.
Attribution models
One dataset. Nine ways to count the same conversion.
Switching model does not re-run your tracking or re-tag a campaign — the journeys are already stored, so a model is only a question you ask of them. All nine are on every plan.
Single touch
One channel takes the whole conversion. Fast to explain, and still the right answer for a short journey.
- First Touchcredits the channel that introduced them
- Last Touchcredits the channel that closed them
Single touch, direct excluded
The same two models with direct sessions skipped, so a bookmarked return visit can't quietly take credit for the work that earned it.
- First Touch (Non-Direct)first source that wasn't direct
- Last Touch (Non-Direct)last source that wasn't direct
Multi touch
Credit split across the journey. Four distributions, because a nurture sequence and an impulse purchase do not deserve the same shape.
- Lineareven split across every touchpoint
- Time Decay7-day half-life
- U-Shaped40 / 20 / 40
- W-Shaped30 / 30 / 30 / 10
Journey influence
A lens rather than a split: which conversions had an AI assistant anywhere in the path, whether or not it took the final click.
- AI journey influenceassistant present anywhere in the journey
Attribution problems
When every platform claims credit, you need a referee.
Add up what each ad platform reports and you will exceed the number of orders you actually shipped. Nobody is lying — each one is measuring inside its own walls. SourceTrack runs alongside them and keeps one count.
Meta says it won.
Its attribution window counts the conversion because someone saw an ad within it. That is a real signal, measured entirely inside Meta.
Google says it won.
Its own window counts the same conversion, for the same reason, from the other side. Both numbers can be defensible and still not add up.
Your data says what happened.
One first-party dataset, one journey per visitor, and nine ways to divide the credit — none of which have a budget to defend.
Attribution data captured
Designed for modern customer journeys.
Attribution is only as good as what got captured on the way in. These are the details that explain why a conversion happened rather than just that it did.
UTMs and source params
Campaign data preserved across landing pages and session boundaries, not dropped at the first redirect.
Referrers and AI traffic
Traditional referrers plus sixteen named assistants and answer engines, which arrive with no UTM at all.
Landing page paths
Which pages created intent and which merely assisted it, per journey rather than per session.
Conversion events
Forms, chats, meetings, trials, signups and payments — business outcomes, not pageviews.
Turn messy journeys into one clear count.
28 days free, no card. Install the script and watch your first attributed session land.
Product visuals shown use illustrative data.