How Tailkey's attribution modeling works

An overview of how Tailkey's model turns raw visitor activity into content-level credit, and how that differs from conventional attribution models.

Many channels. One journey. Credit at content level.

The example journey referenced throughout this page
Organic search
Paid social
Retargeting email
Referral
Direct
The core journey
Category page: running shoes
Page
Customer reviews block
Content block
Product detail page
Page
Free shipping banner
Content block
Checkout
Conversion

A channel shows how someone arrived, and the same person may arrive through several over time. What matters more is the journey underneath: the pages they move through and, within those pages, the specific content that drives the outcome.

Attribution modeling

Buyers rarely convert on their first visit. They arrive through search, return from an ad, come back through an email, and read multiple pages along the way, sometimes within a single afternoon and sometimes over several months. Attribution modeling is the method used to determine how much credit each of those touchpoints deserves for the conversion that eventually follows.

Touchpoints

Every page, ad, email, or piece of content a visitor interacts with on the way to converting.

Paths

The full sequence of touchpoints a single visitor moves through, rather than any one moment in isolation.

Credit

The share of the outcome assigned to each touchpoint, reflecting how much it actually contributed.

Touchpoints and paths are straightforward to record. Credit is where most attribution tools fall short.

Most attribution models guess. This one doesn't.

Almost every model in use today shares the same flaw: credit is assigned by a rule fixed in advance, based only on where a touchpoint sits in the sequence rather than what actually happened there. Switch between the models below to see how differently the same journey gets interpreted each time.

Guide: “Attribution basics”
0%
Section: why last-click fails
0%
Comparison: multi-touch vs last-click
0%
Pricing page, FAQ block
0%
Sign-up
100%
Last-touch

All the credit goes to the final click before conversion, even if it just closed a deal someone else already won.

Position in the sequence, nothing else

None of the first five models actually measure anything: they are fixed rules, applied uniformly regardless of what visitors did. That is why the same page can appear essential under one model and invisible under the next, with nothing on your site having changed.

Tailkey starts from a different premise entirely, by redefining what a touchpoint actually is.

Source, page, content block

Most attribution tools stop at the source, something like "organic search" or "paid social." Better ones reach the page level. Neither reveals what actually persuaded the visitor. Tailkey resolves credit at all three levels, so a source becomes the pages that earned it, and a page becomes the specific blocks of content within it.

Where people arrived from.
Organic search
42%
Paid social
24%
Retargeting email
21%
Direct
13%

Organic search earned 42% of conversions. That is the whole answer a source-level tool can give you, and it does not tell you which page to write next.

Every touchpoint, scored on real behavior.

Tailkey doesn't apply a fixed rule. It captures every touchpoint, reconstructs the full path, and scores each one by how much it actually influenced the outcome.

We record every source, page, and block of content a visitor encounters, not just the first or last thing they saw. That includes the parts that never get a click, such as a pricing FAQ read halfway down a page.

Sources
Pages
Content blocks
Engagement signals

The model

An algorithmic ML based approach.

Most attribution tools pick a single rule, whether first touch, last touch, or an even split, and apply it everywhere. Tailkey blends multiple established attribution techniques, including path-based modelling that compares converting and non-converting journeys and fair-credit allocation drawn from game theory, then layers in further signals: measured engagement with each block, the shape of the journey around it, content type and position, and recurrence across other converting paths. Machine learning re-weights that blend continuously against your real outcomes, so the model you run next month is not the model you run today.

Path-based modelling
Shapley fair credit
Engagement signals
Content & journey factors
One conversion-strength score
Per source, page, and content block, re-weighted continuously by ML.
Multiple methods, one score

Established attribution techniques are combined rather than used alone, so no single method's blind spots dominate the result.

Real engagement, not just clicks

Dwell time, scroll depth, and on-page interaction feed the model, so a page that's actually read counts for more than one that's merely visited.

Re-weighted by ML, continuously

The blend is fit and re-fit on which journeys actually convert, so the weighting keeps adjusting as your content and audience change instead of staying fixed.

Better data. Better decisions. Better results.

Start turning visitor activity into insight, or keep reading to see how that credit becomes audiences you can act on.

Attribution based Segmentation

How Tailkey builds segments from your goal sets

Once every touchpoint is scored, those scores become segments you can export into a campaign.

Learn about segmentation →