Executive Decision Playbook: Reallocating a €500,000 TV Flight

Scenario: Omnichannel Post-Buy Campaign Audit & Budget Optimization
Audience: Chief Marketing Officer, VP of Growth, Media Agency Planning Committee
Question Answered: "How do we move beyond naive 'cheapest channel' sorting to cut wasteful spend, capture untracked Amazon retail demand, eliminate €18,500 in PPC cannibalization, and boost total omnichannel response by +34% in our next flight?"

Benchmark Scenario & Methodology

The figures in this playbook are generated from SpotLift's reproducible DACH & Omnichannel benchmark suite (spotlift demo --scenario dach). They demonstrate how data science and media teams transform raw broadcast logs (Schaltpläne), GA4 minute telemetry, and Amazon Marketing Cloud (AMC) clean-room data into defensible C-suite budget decisions.

Executive Memo: Summary of Campaign Findings

To: Media Planning & Budget Committee
From: Head of Performance Analytics
Date: Post-Buy Flight Audit Review
Subject: Q1 Broadcast Campaign Response, Amazon Retail Halo & Portfolio Optimization

Executive Summary

Our 14-day flight deployed €500,000 across 600 broadcast airings on 4 major networks (RTL, ProSieben, VOX, Joyn).

Traditional web analytics (GA4 alone) presented an incomplete picture that almost led to a catastrophic budget misallocation. By running SpotLift's high-frequency causal engine with Amazon Marketing Cloud (AMC) clean-room telemetry and Constrained Portfolio Optimization, we uncovered three critical realities:

The "Naive Average CPIV" Trap & Broadcaster Commitments:
- Pure web attribution showed Connected TV (Joyn) delivering direct response visits at €7.98 CPIV, compared to €14.20 CPIV on RTL.
- A junior analyst proposed shifting €150,000 into Joyn. The Media Committee rightly rejected this: Joyn's available ad inventory is a fraction of RTL's, meaning massive spend would cause rapid audience exhaustion. Furthermore, our agency deal with Ad Alliance requires a contractual minimum 45% spend share to preserve annual volume discounts.
- Solution: We deployed SpotLift's Constrained Marginal CPIV Optimizer (spotlift optimize), modeling diminishing returns ( $β = 0.70$ ) to find the mathematically optimal allocation that respects our agency contracts.
The Amazon Mobile Blind Spot (AMC Clean Room Discovery):
- Evaluating the campaign solely on GA4 website sessions penalized prime-time network broadcasts.
- Integrating Amazon Marketing Cloud (AMC) clean-room telemetry revealed that 43% of viewers opened the Amazon mobile app rather than visiting our direct website.
- RTL and ProSieben weekend airings drove +18,200 Amazon Product Detail Page Views (DPV), +2,400 Add-to-Carts, and a multi-day retail sales surge that legacy tools missed completely.
Amazon Sponsored Brands PPC Cannibalization:
- SpotLift's PPC Cannibalization Auditor discovered that our agency's auto-bidding Amazon Sponsored Brands campaigns were aggressively spending €2.60 per click on our own brand trademark during the 2-hour window following TV spots.
- These campaigns intercepted 64.8% of organic shoppers who were already searching for our product because of the TV ad, generating €18,500 in unnecessary PPC brand tax.

The Evidence: Omnichannel Scorecard

Station	Media Spend	Spots	Total GRP	GA4 Incremental Visits	AMC Amazon DPV Lift	Amazon Orders (36h Lag)	New-To-Brand (NTB) Share	Blended Omnichannel CPIV	Actionable Assessment
Joyn (CTV)	€65,000	150	115.0	+8,145	+3,420	+165	71.2%	€5.62	🟢 Top Direct Performer: Expand to saturation bound
VOX	€110,000	159	122.5	+10,018	+4,890	+285	65.8%	€7.38	🟢 High E-Commerce Affinity: Maintain allocation
ProSieben	€135,000	132	96.8	+10,007	+7,110	+390	61.5%	€7.89	🟢 Strong Retail Spillover: Prune low dayparts, expand daytime
RTL	€190,000	159	126.7	+13,430	+11,090	+570	58.9%	€7.75	🟡 High Daytime Halo, Late-Night Waste: Prune late fringe

The Omnichannel & NTB Turning Point

On direct web traffic alone, RTL appeared 78% more expensive than Joyn (€14.20 vs. €7.98). But when factoring in Amazon detail page glance views and 36-hour delayed retail purchases, RTL's blended omnichannel cost per response dropped to €7.75. Crucially, 62.4% of total Amazon purchases were New-To-Brand (NTB) customers—proving that broadcast TV expanded our customer acquisition reach into net-new households rather than simply subsidizing repeat buyers.

Daypart Efficiency & The Late Fringe Trap

Segmenting performance by daypart revealed an acute divergence between Daytime (10:00–16:59), Early Evening (17:00–19:59), and Late Fringe (23:00–01:00):

text

DAYPART EFFICIENCY MATRIX (Omnichannel Response by Hour of Day)
======================================================================================================
Daypart          Hours (CET)    Media Spend    GA4 Visits    Amazon DPV    Blended Cost/Response  Status
------------------------------------------------------------------------------------------------------
Daytime Fringe   10:00 - 16:59   € 115,000      + 10,840       + 7,210           € 6.37           🟢 High ROI
Early Evening    17:00 - 19:59   € 145,000      + 14,215       + 9,450           € 6.13           🟢 Sweet Spot
Prime-Time       20:00 - 22:59   € 175,000      + 14,105       + 8,620           € 7.70           ⚪ High Scale
Late Fringe      23:00 - 01:00   €  65,000      +  2,440       + 1,230           € 17.71          🔴 Wasteful
======================================================================================================

Critical Finding: Late-Night Fringe Waste

Broadcaster sales teams bundle post-23:00 airings as "discounted GRP packages." However, our attribution logs show that viewers after 23:30 watch passively and go to sleep. Paying €65,000 for late fringe yielded a dismal return (€17.71 per blended response)—nearly 3x more expensive than daytime and early evening airings.

Constrained Portfolio Optimization: Beyond Naive Sorting

Rather than relying on human guesswork or unconstrained sorting, we executed spotlift optimize with contractual boundaries:

Objective: Maximize total omnichannel response across Web and Amazon.
Elasticity Parameter: $β = 0.70$ (modeling diminishing returns).
Boundary Constraints:
- Maximum shift per channel: ±35% (Balanced Scenario).
- Contractual Ad Alliance (RTL + VOX) spend share ≥ 45%.
- Minimum brand reach preservation: RTL spend ≥ €135,000.

mermaid

flowchart LR
    subgraph Current["Current €500k Flight"]
        RTL1["RTL: €190k (38%)"]
        PRO1["ProSieben: €135k (27%)"]
        VOX1["VOX: €110k (22%)"]
        JOYN1["Joyn: €65k (13%)"]
    end
    subgraph Optimizer["SpotLift Constrained Optimizer (KKT)"]
        direction TB
        Diminishing["Diminishing Returns (β=0.70)"]
        AdAlliance["Ad Alliance Share >= 45%"]
        mCPIV["Equalize Marginal CPIV"]
    end
    subgraph Optimized["Optimized €500k Flight"]
        RTL2["RTL: €145k (29%) (-€45k)"]
        PRO2["ProSieben: €125k (25%) (-€10k)"]
        VOX2["VOX: €142.5k (28.5%) (+€32.5k)"]
        JOYN2["Joyn: €87.5k (17.5%) (+€22.5k)"]
    end
    Current --> Optimizer
    Optimizer --> Optimized

Specific Operational Changes

Prune Late Fringe on RTL and ProSieben (-€55,000): Strip out all spots airing after 23:00. Restrict linear broadcast inventory to daytime (10:00–16:59) and early evening (17:00–19:59).
Scale VOX Early Evening (+€32,500): VOX early fringe lifestyle and cooking programming proved to be our highest single converter for direct e-commerce and Amazon searches.
Expand Joyn CTV to Optimal Saturation (+€22,500): Expand Joyn from €65,000 to €87,500. This captures available digital inventory right up to its marginal cost inflection point, avoiding audience burnout.
Throttle Amazon Sponsored Brands PPC (-€18,500 Saved): Apply automated daypart bidding rules in Amazon Advertising to pause exact-match brand keyword PPC during broadcast windows, letting organic TV-driven searches convert without paying Amazon a €2.60 fee per click.

Modeled Impact of Reallocation

By shifting €55,000 into high-efficiency dayparts and optimizing portfolio allocations:

Metric	Current Flight	Modeled Reallocated Flight	Projected Delta
Total Media Spend	€500,000	€500,000	€0 (Budget Neutral)
Incremental GA4 Web Visits	41,600	54,912	+13,312 (+32.0%)
Amazon DPV Glance Views	26,510	36,053	+9,543 (+36.0%)
Total Omnichannel Responses	68,110	90,965	+22,855 (+33.6%)
Blended Cost Per Response	€7.34	€5.50	-€1.84 (-25.1%)
Amazon PPC Cannibalization Savings	€0	€18,500	+€18,500 Net Reclaimed

Reproducing this Analysis

Run this entire attribution, Amazon clean-room, and budget optimization pass in your own terminal using SpotLift:

bash

# 1. Generate realistic DACH & Amazon benchmark flight:
spotlift demo --scenario dach --out-dir ./campaign_audit

# 2. Run causal attribution pass across GA4 and Amazon telemetry:
spotlift run \
  --spots ./campaign_audit/tv_schaltplan.csv \
  --web ./campaign_audit/ga4_minutes.parquet \
  --amazon ./campaign_audit/amazon_hourly.csv \
  --format all \
  --out ./campaign_audit/attributed_spots.csv

# 3. Solve constrained portfolio budget optimization:
spotlift optimize \
  --spots ./campaign_audit/attributed_spots.csv \
  --budget 500000 \
  --max-shift 0.35 \
  --scenarios \
  --out ./campaign_audit/recommended_budget.csv

Executive Decision Playbook: Reallocating a €500,000 TV Flight ​

Executive Memo: Summary of Campaign Findings ​

Executive Summary ​

The Evidence: Omnichannel Scorecard ​

Daypart Efficiency & The Late Fringe Trap ​

Critical Finding: Late-Night Fringe Waste ​

Constrained Portfolio Optimization: Beyond Naive Sorting ​

Specific Operational Changes ​

Modeled Impact of Reallocation ​

Reproducing this Analysis ​