Skip to content
Revenue and attribution

What is Multi-touch attribution?

Multi-touch attribution divides a conversion's credit across every source that contributed rather than giving it all to one. Common rules are linear (equal shares), time-decay (later touches weigh more) and position-based (usually 40% first, 40% last, 20% spread between). Every rule is a chosen assumption, not a discovered fact.

The common models

ModelHow credit splitsSuits
LinearEqual share to every touchLong journeys where you cannot justify weighting
Time-decayExponentially more to recent touchesShort cycles, promotional selling
Position-based40% first, 40% last, 20% divided among the middleJourneys where discovery and closing both matter
Data-drivenWeights derived statistically from your own conversion pathsLarge volumes, and only with real data-science support

The honest framing

None of these models measures causation. A visitor’s journey is a list of things that happened before a purchase; the model is a rule for dividing credit among them, chosen in advance by a person. Run the same sales through linear and time-decay and channels will change rank — not because anything about the business changed, but because you changed the assumption.

That is not a reason to avoid multi-touch. It is a reason to be explicit: state the model on the report, keep it fixed long enough for trends to mean something, and treat a change of model as a change of instrument rather than a discovery.

Practical guidance

  • Start with two single-touch models, not a complex one. First-click and last-click side by side reveal most of what multi-touch would tell you, and neither hides its assumption inside a weighting formula.
  • Linear is the least arguable multi-touch rule. If you cannot defend why one position should weigh more, equal shares is the honest default.
  • Data-driven models need volume. Below a few thousand conversions a month they are fitting noise, however sophisticated the description sounds.
  • Every model needs cross-visit memory. The longer the journey you want to model, the more visitor history the tool must retain — which is a real privacy cost, and worth weighing against the precision you actually gain.
Questions

Frequently asked.

Cookies, install and pricing, answered. Still stuck? Ask us anything .

01 Can sonex show revenue next to my traffic?

Yes. Connect Stripe or Polar with a read-only key and sonex reads revenue straight from your payment provider, per website. Revenue then appears as a focusable series on the Overview chart and as its own report, beside the traffic that earned it. No tracked event is needed for it to work.

02 Does sonex use cookies?

No. sonex sets no cookies and needs no consent banner. It counts visits without cookies, fingerprinting, or any personal data, so it is GDPR, PECR and CCPA-ready by default.

03 How do I install sonex?

Add one script tag to your site's <head> with your website id. It is a single lightweight tracker — no build step and no SDK required.

04 Is sonex a Google Analytics alternative?

Yes. sonex gives you the reports that matter — visitors, pages, referrers, funnels, revenue and a world map — without surveilling your audience or drowning you in configuration.

05 How is sonex priced?

By monthly tracked events. Free covers 2k events, Pro is $20/mo for 200k events, and Business is $200/mo for 2M events with team seats.

See what your traffic actually earns.

Revenue beside the visitors that produced it. No cookies, no credit card, no consent banner.

Get started