Which channels touched the conversion path?
№ 170 · Attribution waterfall
Definition
MetricDistribution of credit across channels on the conversion path, comparing first touch, last touch and multi-touch models
Unit% of conversions or revenue credited per channel
Bars showing % of conversions touched by each channel (multi-touch attribution). Build from MTA platform output. Identifies channels that get credit they don't deserve under last-click. Example: Search 32%, Meta 24%, Email 18%, Direct 14%, Other 12% = 100%.
- MTA
- Multi-touch attribution. Credit for a conversion split across every touchpoint rather than the last one.
Benchmarks
| Bottom 30% | Median | Top 30% |
|---|---|---|
| A model swap that moves a channel's credit by 40 points or more, which means budget decisions are being made by the model rather than by the data | B2B journeys average 76 touchpoints across 3.7 channels; full-journey tracked touchpoints run 222 to 266 and rose 19.8% between 2023 and 2024 | Not applicable as a performance metric; the healthy reading is a waterfall where no single model changes any channel's credit by more than roughly 20 percentage points, meaning the conclusion is robust to model choice |
B2B SaaS platform-tracked journeys, Dreamdata and HockeyStack 2024 to 2025 datasets; touchpoint counts are tracked digital interactions including ads, web visits, email opens and content views, not sales conversations · There is no benchmark for what share of credit a channel should receive, so this card benchmarks the shape and the fragility of the waterfall rather than its values. Touchpoint counts differ by an order of magnitude between sources because they count different things: Dreamdata's 76 touchpoints per deal and HockeyStack's 222 to 266 are both tracked-interaction counts on different definitions, while RAIN Group's 8 touches to a first meeting counts direct outreach only. Do not put them on one axis. Forrester's finding that a typical B2B decision now involves 13 internal stakeholders and 9 external influencers means single-touch models are structurally unable to describe the path.
| Category | Bottom 30% | Median | Top 30% |
|---|---|---|---|
| B2B SaaS, deals under $100K ACV | n/a | roughly 266 touchpoints and 2,879 impressions per closed deal | n/a |
| B2B SaaS, deals at or above $100K ACV | n/a | roughly 417 touchpoints and 5,500 impressions per closed deal | n/a |
| B2B, first-touch credit distribution | n/a | organic 30.84%, direct 30.44%, PPC 22.31% | n/a |
| B2B, last-touch credit distribution | n/a | PPC 32.90%, direct 26.56%, organic 23.39% | n/a |
When it looks bad
The first-touch and last-touch bars for the same channel sit at opposite ends of the chart, so branded search takes most of the credit under last touch and almost none under first touch, and the waterfall is measuring where the journey ended rather than what caused it.
Branded search holds 44% of last-touch credit and 6% of first-touch, paid social holds 5% last-touch and 34% first-touch, and the budget was set on last-touch.
What to do about it
- Show first touch, last touch and a linear or position-based model as three parallel bars per channel rather than choosing one; the decision-useful information is the spread, and a channel whose credit is stable across models can be budgeted with confidence while one that swings cannot.
- Route any channel with a large first-to-last spread into an incrementality test rather than arguing about the model; branded search measured at 19% incrementality against Google's attributed conversions is the canonical case where the waterfall and the truth diverge.
- Cap the touchpoint window at the observed journey length and state it on the chart; with an average 272 days from first touch to revenue in the Dreamdata panel, a 90-day lookback will systematically hand credit to late-funnel channels.
- Track the count of tracked touchpoints per won deal as a companion series; it rose 19.8% between 2023 and 2024 in the HockeyStack benchmark, and a falling count usually means tracking loss rather than a shortening journey.
Sources
- Dreamdata Average 272 days from first touch to revenue; a deal is influenced by an average 76 touchpoints across 3.7 channels dreamdata.io ↗
- Dreamdata Average B2B customer journey 192 days from anonymous first touch to closed won; large companies average 29 sessions in the buying cycle against 35 for mid-sized dreamdata.io ↗
- SyncGTM Full buyer journey 222 to 266 tracked touchpoints, up 19.8% from 2023 to 2024; SMB deals close in 5 to 12 direct touches, mid-market 15 to 30, enterprise 250+ syncgtm.com ↗
- 4M Digital Consulting Typical B2B buying decision now involves 13 internal stakeholders and 9 external influencers; 67% of B2B buyers prefer a rep-free experience and 45% used AI tools during a recent purchase 4mdigitalconsulting.com ↗
- Dad's Growth Lab Enterprise deals above $250K average 192 days from first touch to closed won per Ebsta and Pavilion 2024; buyers consume approx. 13 content pieces before first sales contact dadsgrowthlab.com ↗
- ZoomInfo Forrester: 70% of B2B marketing organizations tracked marketing-sourced pipeline in 2015, 47% by 2020, projected 14% by 2025; sourced attribution runs 5 to 20% in ABM and enterprise contexts; coverage heuristic is 2 to 3x revenue quota pipeline.zoominfo.com ↗
- Growthspree Marketing-sourced 25 to 45% of pipeline with median 35%; marketing-influenced 60 to 85% with median 72%; healthy sourced-to-influenced ratio near 1:2.0 growthspreeofficial.com ↗
- Marqeable Practitioner median 25 to 50%, PLG and inbound-led up to 60 to 80%, enterprise outbound roughly 15 to 35%; win rates run 19 to 21% per Ebsta x Pavilion marqeable.com ↗