HBS Reading: 2.1.4 (Attribution, Online-Offline Interaction,Omni Channel Shopping)

Overview

The effectiveness of outbound marketing in the digital age is challenging because traditional metrics often fail to capture the complexity of the consumer's buying journey. Section 2.1.4 of the Harvard Business School reading explores five key issues in measurement: causality, attribution, dynamics, online-offline interaction, and Customer Lifetime Value (CLV) .

The Causality Problem: Correlation vs. Causation

The core debate centers on whether clicks prove that an ad caused a sale.

• The Challenge (eBay Study): A large-scale field experiment on eBay suggested that traditional click-and-sale measurements often mistake correlation for causation. The study found that branded keywords provided virtually no measurable short-term benefit, arguing that many consumers who clicked on a search ad would have clicked the organic link or bought the product anyway .

• The Defense (Google): Google countered this finding, claiming its internal studies showed that 89% of search ad clicks were incremental (truly caused by the ad), even when an organic link was present, encouraging advertisers to run their own experiments .

Attribution: The Path to Purchase

Attribution addresses how marketers assign credit to different media channels when a customer uses multiple touch points before buying.

• The Flaw of Last-Click: The most common approach, last-click attribution, gives 100% of the credit for a sale to the final ad clicked (e.g., a paid search ad), completely ignoring the display ads, social posts, or other media that may have helped move the consumer along the purchase funnel.

  • Ad Hoc Models: Models like First Interaction, Linear (equal credit to all), Time Decay (more credit to recent clicks), and Position-Based are common but lack scientific grounding and often overweight ads that appear frequently.

• Rigorous Models:

  • Regression- or Model-Based: Uses existing data to scientifically model the true effect of each interaction.

  • Experiment-Based: Randomizes ad exposure to treatment and control groups to determine efficacy, providing the most accurate way to measure effectiveness, although it is complex to manage across multiple ad networks.

• Synergy: Search and display ads do not work in isolation. Combining them provided a 22% lift in conversion compared to search alone, suggesting that ignoring their interaction underestimates display ad effectiveness .

Dynamics: The Delayed Impact of Advertising

This issue accounts for the time delay between seeing an ad and making a purchase.

• The ZMOT: For expensive or complicated products (like cars), the consumer engages in an intense, iterative information search process over weeks or months, a phenomenon known as the Zero Moment of Truth (ZMOT) .

• The Problem: Conventional measurement focuses on short-term effects and ignores this potential long-term impact.

• The Finding: Studies using dynamic models found that ignoring the long-term effect of ads underestimated ROI by up to 40%.

Online-Offline Interaction & Omnichannel Shopping

This addresses how digital ads influence physical shopping and sales, which is critical for retailers like Coca-Cola and Unilever whose sales primarily occur offline.

• Synergy: There is strong synergy between channels. In-store experiences are as influential as online ones, especially for products requiring personal fit (e.g., clothing, footwear) . Companies that plan budgets separately may be underperforming.

• Omnichannel Challenge: Consumers research online but buy in the physical store (e.g., checking product details on a website, then going to the store to try it on and buy it). This means simple metrics like CTR and conversion rates significantly underestimate the value of online ads.

• Measuring Offline Conversions: Advertising platforms now help measure the true impact: IKEA found an 11% increase in store visits among customers who saw Facebook ads, and Michael Kors measured a 31% increase in in-store transactions .

• Moment-Based Marketing: Mobile technology enables marketers to target consumers at micro moments ("I want to know," "I want to go," "I want to buy"). For example, sending ads for nearby restaurants to an Uber user on a Friday evening.

Customer Lifetime Value (CLV)

CLV is emerging as the ultimate metric for measuring ad effectiveness, moving beyond short-sighted metrics.

• Beyond CTR: CTR doesn't always correlate with profitability.

• The Gaming Example: The mobile gaming industry shifted from focusing on Cost Per Install (CPI) to CLV because the majority of users abandon games quickly (94% after a month), and few spend money . CLV ensures acquisition efforts target customers who will provide long-term revenue.

• The Risk of Online Customers: Customers acquired through online channels may be more price-sensitive, potentially leading to lower retention rates and consequently lower CLV compared to those acquired through traditional means.