Six models and a data-driven view
The models are lenses on the same touches. Demander shows them side by side so you can see where they agree.
| Model | How credit is split |
|---|---|
| First touch | All credit to the first touch in the lookback |
| Last touch | All credit to the last touch before the conversion |
| Linear | Equal credit to every touch |
| Time decay | Credit halves for every 7 days a touch lies before the conversion |
| Position-based | 40 percent to the first touch, 40 percent to the last, 20 percent shared by the touches between |
| W-shaped | 30 percent each to the first touch, the touch that created the lead and the one that created the opportunity; 10 percent shared by the rest |
| Markov | Each channel’s removal effect: the share of conversions that would disappear without it, computed over converting and non-converting journeys |
Touches after the conversion or older than the 90-day lookback get no credit, and the credits for each conversion add up to exactly 1. For causal answers, use a holdout test.
What you can attribute
-
01
Down to the message
Credit rolls up by channel, campaign, persona, creative or message. By message tells you which line of argument creates pipeline, whichever channel carried it.
-
02
Joined with spend
Each row shows spend, pipeline, revenue, ROAS, pipeline ROI and CAC, so a cheap channel that never opens opportunities stands out.
-
03
Any conversion
Attribute opportunities, any primary conversion or one conversion such as booked demos. Value then replaces pipeline, with cost per conversion per row.
-
04
Drill down
From any row, open the opportunities or conversions it contributed to, and from there the full path.
What counts as a touch
The tracker writes one touch per visit, keyed by the visit and its source, so a visit that sends ten batches still counts once. Paid clicks, organic post clicks, form fills, e-mail clicks, outreach e-mails and replies, direct messages and blog reads are touches, each with the pages seen and the strongest intent of that visit: pricing, comparison, case study or docs. Bounces and automatic replies are not touches.
E-mail clicks are traced through the sequence and step in the link, never through an open pixel or a recipient id. Ad impressions cannot be placed on a journey, because the platforms report them only in totals, so the path explorer shows exposure per channel next to the paths.
Account timelines with CRM milestones
-
01
One timeline per account
Touches from every contact at the company, lifecycle changes, CRM milestones and opportunities, in order. People are named only when they identified themselves; anonymous visits stay anonymous.
-
02
Meetings and stages from your CRM
The CRM sync reads meetings from HubSpot, Pipedrive and Salesforce, dated when they were booked and updated in place when rescheduled. Opportunity stage moves, wins and losses appear as milestones too.
-
03
Milestones never take credit
A meeting or a stage change shows on the timeline but gets no attribution credit, so it cannot crowd out the touches that led to it. A cancelled meeting stays visible with no weight.
-
04
Matched without guesswork
CRM records are matched to accounts by domain, else by the CRM id Demander stored when it pushed the account, else through a contact it pushed.
Paths: one opportunity, and all of them
-
The path to one conversion
For an opportunity, a form fill or a demo request, every touch of the account in the 90-day lookback, across all its contacts, in order, with the credit each one gets under the chosen model. It uses the same touches and engine as the report, so the numbers reconcile.
-
The path explorer
Each journey becomes a sequence of steps: the channel, a pricing visit, a form, a demo request. Repeats collapse, so a send followed by its click reads as the click, and long paths keep the first step and the last ones.
-
Converting and non-converting
Paths are counted for accounts that converted and for accounts touched in the same window that never did, with conversion rate, time to convert, touches and pipeline. A step-by-step flow shows where journeys go next.
-
By channel family
Group channels into families (paid search, paid social, display and video, organic social, e-mail, site visit, messaging, sales) to see the pattern behind many small paths.
Questions people ask
Which CRMs does Demander read?
HubSpot, Pipedrive and Salesforce, synced every hour. Opportunities, revenue, stage changes and meetings come from there. See CRM.
Why several models and a Markov view?
Each model answers a different question, and none of them is causal. Seeing first touch, last touch and position-based side by side shows which channels open accounts and which close them. The Markov view also counts journeys that never converted, which the rule-based models ignore.
Can I attribute booked demos without opportunities in a CRM?
Yes. Pick any conversion you defined, such as a booked demo or a sign-up, and attribution, cost per conversion and paths follow that conversion. Anonymous visitors’ conversions are credited over their own visits.
Do ad impressions count?
They are shown per channel next to the paths, and they get no credit. Platforms report impressions only in totals, so they cannot be tied to an account’s journey.