The 7-Model Multi-Touch Attribution Audit: Reallocating B2B Marketing Spend

Marketing Academy · Field Work●Analytics & Attribution

CoreAudit· 50 minutes

The Trade Desk

Objective: Audit a multi-touch B2B marketing pipeline dataset across 7 attribution models (Last-Click, First-Click, Linear, Time-Decay, U-Shaped, W-Shaped, and Data-Driven) in Google Sheets and GA4 to uncover channel cannibalization and reallocate a quarterly media budget.

You are the growth analytics lead at The Trade Desk (TTD), auditing acquisition efficiency across five core channels: Paid Search (Branded & Non-Branded), LinkedIn Sponsored Content, Industry Podcasts/PR, SEO/Content Hub, and Lifecycle Email. Leadership currently allocates budget using last-click reporting, which suggests cutting podcast and top-of-funnel content spend. Your job is to run a full 7-model comparative audit to demonstrate the true pipeline contribution of every touchpoint.

Analyze a dataset of 120 multi-touch customer journeys leading to $1,800,000 in closed-won enterprise ACV. Run 4 diagnostic steps: (1) Last-Click vs. First-Click Channel Bias Diagnostic, (2) Multi-Touch Model Credit Distribution (Linear vs. Time-Decay vs. U-Shaped vs. W-Shaped), (3) W-Shaped B2B Milestone Attribution, and (4) Data-Driven Attribution & Incrementality Budget Reallocation Plan.

How do you mathematically evaluate channel performance across multi-touch B2B journeys to prevent last-click attribution from defunding your most effective top-of-funnel channels?

Multi-Touch Attribution/W-Shaped Modeling/Incrementality Testing/Media Budget Allocation

Before you start

What you'll need

—Knowledge of the 7 major attribution models and their mathematical weighting
—Experience building pivot tables and weighted formulas in Google Sheets
—Understanding of B2B CRM sales stages (Lead, MQL, SQL, Opportunity, Closed Won)

Demand Creation vs. Harvesting: Demand creation introduces new prospects to your brand (social, PR, content), while demand harvesting captures high-intent prospects who are already actively searching to buy (branded search, retargeting).
W-Shaped Attribution: A position-based multi-touch model that allocates 30% to First Touch, 30% to Lead Creation, 30% to Opportunity Creation, and 10% evenly across remaining touches.
Incrementality Testing: Controlled experimentation (such as geo-holdout tests) that measures the true net-new conversion lift caused by an advertising channel compared to a baseline where ads are paused.

Free path (everything below is enough to finish)

Google Sheets

FreeAttribution model formulas and budget reallocation modeling

Transparent multi-touch credit calculation and incrementality reconciliation

Google Analytics 4

FreeTouchpoint extraction and path exploration

Pull multi-channel funnels and user pathing data

Paid upgrades (optional, faster/deeper)

Segment(optional)

FreemiumCross-platform identity resolution and event stream routing

Unify anonymous web touchpoints with CRM sales pipeline milestones

The process

4 steps

Step 01 of 04

Last-Click vs. First-Click Channel Bias

Last-click gives 100% credit to the final touchpoint (heavily biasing toward branded search and email), while first-click gives 100% to discovery channels (YouTube, PR, organic search). Comparing both reveals which channels create demand vs. which simply harvest it.

What percentage of total pipeline revenue is captured by Branded Search under Last-Click compared to First-Click, and what does this reveal about its true demand-generation role?

Google Sheets— Open multi-touch-export.csv in Google Sheets, create pivot tables for 'First Touch Channel' vs 'Last Touch Channel' weighted by Deal Value.

Procedure

Import the 120-journey multi-touch dataset into Google Sheets.
Calculate total pipeline attributed to each channel under 100% First-Touch vs. 100% Last-Touch.
Compute the Ratio of Last-Touch to First-Touch attribution for Branded Search, Non-Branded Search, Paid Social, Content/SEO, and Email.
Isolate channels with a Ratio > 3.0 (Demand Harvesters) and Ratio < 0.5 (Demand Creators).

Sample output

ATTRIBUTION BIAS ANALYSIS ($1,800,000 Total ACV Pipeline):

Channel              First-Click ACV    Last-Click ACV    L/F Ratio    Classification
Branded Search       $90,000 (5%)       $720,000 (40%)    8.0x         Demand Harvester
Lifecycle Email      $54,000 (3%)       $450,000 (25%)    8.3x         Demand Harvester
LinkedIn Ads / PR    $630,000 (35%)     $180,000 (10%)    0.28x        Demand Creator
SEO / Content Hub    $756,000 (42%)     $270,000 (15%)    0.35x        Demand Creator
Non-Branded Search   $270,000 (15%)     $180,000 (10%)    0.66x        Hybrid Discovery

Healthy

First-click and last-click models are compared side-by-side to expose demand creation vs. demand harvesting roles across channels.

Unhealthy

Allocating top-of-funnel budgets exclusively based on last-click numbers, causing demand generation channels to be starved and shrinking overall pipeline.

What this means

Branded Search claims 40% of revenue under last-click but initiates only 5% of deals. Cutting LinkedIn Ads or Content because of low last-click ROI would dismantle the top-of-funnel engine that feeds branded searches 60 days later.

So what do I do about it?

Symptom	Action	Effort
Executive leadership proposes reducing content and paid social budgets based on low last-click conversion rates	Present the First-Click vs. Last-Click comparison matrix proving that 77% of all won revenue originated from content and social discovery	30 min
Branded search budget uncapped while top-of-funnel spend is constrained	Cap branded search spend to match incremental demand and reallocate surplus to high-first-touch channels	half day

YouYou can do this yourself, no engineering access required.

Step 02 of 04

Multi-Touch Model Credit Distribution

Multi-touch models distribute credit across the entire path. Linear splits credit equally, while Time-Decay weights recent interactions. Position-based models (U-shaped) give 40% to first and last touches with 20% in the middle.

How does shifting from Linear to Time-Decay attribution alter the valuation of mid-funnel content downloads and product webinars relative to late-stage sales demo requests?

Google Sheets— Apply weighted formula columns in Google Sheets for Linear (1/N), Time-Decay (half-life 7 days), and U-Shaped (40/20/40) credit rules.

Procedure

Assign fractional credit to each of the 450 total touchpoints across the 120 customer journeys.
Calculate total attributed revenue per channel under Linear, Time-Decay, and U-Shaped models.
Quantify the credit shift between 30+ day old awareness touches and 7-day conversion touches.

Sample output

MULTI-TOUCH ATTRIBUTION COMPARISON ($1,800,000 Pipeline):

Channel              Linear (Equal)    Time-Decay (7d)   U-Shaped (40/20/40)
LinkedIn Ads / PR    $432,000 (24%)    $252,000 (14%)    $468,000 (26%)
SEO / Content Hub    $504,000 (28%)    $324,000 (18%)    $486,000 (27%)
Non-Branded Search   $324,000 (18%)    $288,000 (16%)    $288,000 (16%)
Lifecycle Email      $270,000 (15%)    $414,000 (23%)    $270,000 (15%)
Branded Search       $270,000 (15%)    $522,000 (29%)    $288,000 (16%)

Healthy

Model choice reflects the specific sales cycle duration: U-shaped properly credits both discovery and final action in multi-stakeholder B2B deals.

Unhealthy

Using Time-Decay for a 6-month enterprise sales cycle, which artificially mimics last-click bias by devaluing long-term nurturing.

What this means

Linear and U-shaped models recognize that content and social drive over 50% of value, while Time-Decay heavily concentrates credit into the final 7 days.

So what do I do about it?

Symptom	Action	Effort
Time-decay model shows low ROI on quarterly industry research reports	Switch B2B executive reporting to U-shaped or W-shaped models that reward top-of-funnel lead creation	30 min

YouYou can do this yourself, no engineering access required.

Step 03 of 04

W-Shaped Attribution for B2B Milestones

W-shaped attribution allocates 30% to First Touch (awareness), 30% to Lead Creation (form fill), 30% to Opportunity Creation (sales stage validation), and 10% to intermediate nurturing touches.

Which channels emerge as the primary drivers of the Opportunity Creation milestone (the second 30% W-anchor) compared to First Touch and Opportunity Close?

Google Analytics 4— GA4 Explore > Path Exploration & Custom Funnel / Segment integration mapping CRM lifecycle milestones to web touchpoints.

Procedure

Map each touchpoint in the journey to its closest CRM milestone: First Touch, Lead Creation, Opportunity Creation, Opportunity Close.
Apply 30/30/30/10 weighting to calculate channel revenue shares across the 4 key milestones.
Identify the top channel for each of the three major 30% anchors.

Sample output

W-SHAPED MILESTONE ATTRIBUTION BREAKDOWN:

Milestone Anchor            Top Contributing Channel        Attributed Revenue
1. First Touch (30%)        SEO / Content Hub (52%)         $280,800
2. Lead Creation (30%)      LinkedIn Sponsored Content (44%)$237,600
3. Opp Creation (30%)       Product Webinars & Case Studies $259,200
4. Middle Nurturing (10%)  Email Sequences & Retargeting   $180,000

W-Shaped Channel Totals: Content/SEO: $486,000 (27%) | LinkedIn/PR: $450,000 (25%) | Non-Branded: $306,000 (17%) | Email: $270,000 (15%) | Branded: $288,000 (16%)

Healthy

W-shaped modeling reflects the distinct operational roles of discovery content (first touch), targeted lead-gen (lead creation), and deep product proof (opp creation).

Unhealthy

Treating all middle touches as homogeneous instead of isolating the specific asset that converted an MQL into a validated sales opportunity.

What this means

Content and SEO dominate top-of-funnel discovery, LinkedIn dominates lead capture, and webinar/case-study content drives pipeline qualification. Every channel plays a clear, complementary role.

So what do I do about it?

Symptom	Action	Effort
Sales team complains that MQLs from social ads fail to convert into qualified pipeline	Invest in mid-funnel case study content and product webinars to strengthen the Opportunity Creation milestone	half day

YouYou can do this yourself, no engineering access required.

Step 04 of 04

Data-Driven Attribution & Incrementality Reconciliation

Attribution models show observed paths, but incrementality testing and MMM prove causality. Combining attribution with holdout tests prevents over-investing in channels with zero incremental lift.

When reconciling W-shaped attribution against a 4-week geo-holdout test on Branded Search and LinkedIn Ads, what is the true incremental ROI of each channel?

Google Sheets— Google Sheets Incrementality Model tab, comparing test vs. control market conversion lift against model-attributed revenue.

Procedure

Review 4-week geo-holdout results (15% control market with ads paused vs. 85% test market).
Calculate Incremental Lift Factor for Branded Search (12% lift when ads ON) vs. LinkedIn Ads (84% lift when ads ON).
Adjust W-shaped attribution revenue by channel incrementality factors to compute True Incremental Revenue.
Formulate the finalized Q3 budget reallocation recommendation.

Sample output

INCREMENTALITY RECONCILIATION & BUDGET REALLOCATION:

Channel            W-Shaped ACV    Incrementality Factor   True Incremental ACV   Old Budget   Recommended Q3
Content / SEO      $486,000        95% (Organic Baseline)  $461,700               $40,000      $65,000 (+63%)
LinkedIn Ads / PR  $450,000        84% (High Causality)    $378,000               $50,000      $75,000 (+50%)
Non-Branded Search $306,000        72% (Moderate Lift)     $220,320               $45,000      $40,000 (-11%)
Lifecycle Email    $270,000        65% (Customer Nurture)  $175,500               $15,000      $20,000 (+33%)
Branded Search     $288,000        12% (High Cannibalism)  $34,560                $50,000      $20,000 (-60%)
TOTALS             $1,800,000      --                      $1,270,080             $200,000     $220,000

Healthy

Attribution models are audited against real incrementality testing before reallocating major budget lines.

Unhealthy

Treating 100% of attributed branded search revenue as net-new growth when 88% of users would have navigated organically anyway.

What this means

Branded search has high attributed revenue but low incremental lift (12%). Reallocating $30,000 from Branded Search to Content/SEO and LinkedIn Ads increases overall incremental pipeline by an estimated 28% without increasing total marketing budget.

So what do I do about it?

Symptom	Action	Effort
Branded search consuming 25% of total paid media budget with diminishing returns	Reduce branded search target impression share from 99% to 80% and divert freed budget to LinkedIn awareness and SEO	30 min
Lack of incrementality data causing ongoing debates between performance and brand teams	Establish a quarterly geo-holdout testing cadence on the largest paid channels	dev ticket

YouYou can do this yourself, no engineering access required.

Analyze your findings

What to look for

Last/First Ratio: Identify channels with L/F > 3.0 (harvesters) vs L/F < 0.5 (creators) to understand true channel roles.
Milestone Anchors: Verify which specific channels drive First Touch, Lead Creation, and Opportunity Qualification in W-shaped models.
Holdout Lift: Reconcile attributed revenue against experimental holdout data before shifting budget.

Make the call

Your CFO notices that Branded Search generated $720,000 in last-click revenue with a 14x ROAS, while Top-of-Funnel Content generated only $180,000 with a 2.5x ROAS. She proposes shifting 50% of the content budget into Branded Search. How do you respond?

Recommendation · Priority: High

“Adopt W-shaped attribution for B2B pipeline reporting and calibrate all channel allocations against quarterly incrementality holdout tests. Cap branded search spend to capture legitimate competitive conquesting, and reinvest the surplus into high-first-touch content and targeted social channels.”

Common mistakes

What trips people up

Treating branded search as a standalone growth driver rather than a navigational touchpoint. — Branded search has high last-click conversion rates because buyers already decided to visit; crediting it with 100% of deal value leads to over-bidding on existing brand equity.
Using Linear attribution without separating milestone touches from routine page views. — Equal weighting overvalues inconsequential middle page views while undervaluing the critical content assets that generated the lead or qualified the opportunity.

Final deliverable

A comprehensive 7-Model Attribution Audit & Budget Reallocation Plan containing channel bias ratios, multi-touch credit comparison tables, W-shaped milestone allocations, and an incrementality-adjusted quarterly budget model.

See a reference example

Sample output

KLAVIYO / SEGMENT ATTRIBUTION AUDIT REFERENCE REPORT

1. Touchpoint Bias Analysis ($2.4M E-Commerce Revenue):
   - Email Last-Click Revenue: $1,200,000 (50%) | First-Click Revenue: $120,000 (5%)
   - Meta Ads Last-Click: $360,000 (15%) | First-Click: $1,080,000 (45%)
   - Insight: Meta creates 45% of customer discovery; Email captures 50% of conversions.

2. W-Shaped Milestone Modeling:
   - Discovery Anchor (30%): Meta Ads & Influencer PR (62% share)
   - Lead Capture Anchor (30%): Welcome Quiz & Popups via Segment (48% share)
   - Purchase Anchor (30%): Abandoned Cart Flows & SMS (58% share)

3. Incrementality Reconciliation & Spend Adjustment:
   - Branded Search Holdout Test: 8% incremental lift -> Cut spend from $40k/mo to $15k/mo.
   - Meta Prospecting Holdout Test: 78% incremental lift -> Scale spend from $60k/mo to $85k/mo.
   - Net Outcome: +22% incremental revenue with flat total ad spend.

Success criteria

You're done when you can:

Quantifies demand creation vs. demand harvesting across channels using first vs. last-click ratios
Calculates multi-touch revenue attribution across Linear, Time-Decay, U-Shaped, and W-Shaped models
Maps channel contributions to specific B2B buying milestones using W-shaped attribution
Reconciles attribution models with incrementality holdout tests to formulate a waste-free quarterly budget reallocation plan

Key takeaway

Attribution models explain the journey, but incrementality proves causality. By pairing W-shaped milestone attribution with holdout testing, growth teams can protect vital awareness channels, eliminate wasted spend on non-incremental clicks, and maximize true pipeline growth.