Étude de cas
Le retour sur investissement de l'IA
Nielsen + Google :
Quantifier le pouvoir
de l'IA dans la publicité
Introduction
Le nouveau rôle de l'IA dans les résultats marketing
Artificial Intelligence (AI) is often celebrated as the ultimate solution to advertising efficiency, helping brands plan better, work smarter and save time. It’s the headliner at every industry event and a hot topic in boardrooms. And so far, AI has proven a valuable asset in streamlining the planning process. But the next big question is: Can AI drive measurable impact after the planning phase has passed? And how could you possibly have enough data to measure every touchpoint and control for enough external factors to truly find out?
Objectif
Quantifier l'impact de l'IA sur les résultats
Google sought to uncover the bottom line impact of their latest AI-powered solutions across YouTube and Search by putting them to the ultimate test: measuring over 50,000 brand campaigns and over 1 million performance campaigns in the U.S. across their AI tools. These campaigns spanned a wide range of categories like food, home and personal care, retail, apparel, telecom and automotive over a two-year period ending June 2024. Their AI-powered solutions in question included:
1. Video Reach Campaigns (VRC): Maximize reach with AI, combining multiple formats and optimizing towards a singular goal of efficient reach, non-skippable reach, or reach at a target frequency.
2. Video View Campaigns (VVC): Drive increased consideration by leveraging AI to optimize ad placement, delivering more views at a lower cost.
3. Demand Gen: Generate demand, drive performance, and deliver strong ROI with AI-driven, multi-format ads across YouTube and other visual surfaces, ideal for social advertisers.
4. Performance Max: Achieve more conversions and value across all Google channels, with AI optimizing performance in real-time.
5. Broad Match: Expand reach beyond exact and phrase match, using AI to target additional, relevant queries and attract valuable customers, increasing conversions and improving ROI.
Défi
L'IA n'a pas fait ses preuves
Measuring the performance of AI-powered solutions is a new challenge in the marketing ecosystem. Google needed a measurement solution that could isolate the distinct impact of AI from manual campaigns and also access rich category-level data to effectively capture their wide campaign footprint. In order to prove whether AI-powered solutions moved the needle on return on ad spend (ROAS) and sales effectiveness, they needed a marketing mix modeling (MMM) solution with the ability to answer their research questions robustly.
Solution
Relier les points de vue sur la performance
Google turned to Nielsen Marketing Mix Modeling’s new AI/ML-powered modeling platform to uncover ROAS and sales effectiveness. Nielsen took Google’s over 50,000 brand campaigns and over 1 million performance campaigns, controlled for external factors, and used advanced modeling techniques to aggregate digestible and accurate insights. Due to Nielsen’s extensive data categorization and data partnerships, Nielsen was able to provide the deep cuts of data needed for granular and speedy analysis.
Principales conclusions
17%
L'IA génère plus de ROAS que le manuel
Les campagnes vidéo alimentées par l'IA de Google sur YouTube offrent un ROAS 17 % plus élevé que les campagnes manuelles.
23%
Les synergies entre les campagnes d'IA renforcent l'efficacité
Le VRC pour Efficient Reach + VVC
10%
La génération de demande stimule le ROAS et l'efficacité des ventes
L'ajout de Demand Gen
8%
Performance Max permet d'augmenter les performances de la recherche autonome.
Performance Max
15%
Le "Broad Match" génère un ROAS et une augmentation des ventes significatifs.
Les campagnes de recherche avec correspondance large (Broad Match) alimentées par l'IA de Google offrent un ROAS 15 % plus élevé et une efficacité commerciale 10 % plus élevée que les autres stratégies de correspondance de mots clés.
Résultats
La recette du succès
The results showed that Google AI-powered advertising solutions consistently outperformed manual campaigns in both ROAS and sales effectiveness. But the real magic happened when brands combined multiple AI ad solutions strategically. This AI mix unlocked the best balance of reach, efficiency and returns.
Comment cela fonctionne-t-il ?
Nielsen Marketing Mix Modeling is a statistical analysis technique used to measure the impact of various marketing activities on business outcomes, such as sales and ROAS. By analyzing historical data from multiple channels, Nielsen isolates the effect of each marketing element while controlling for external factors like seasonality and economic shifts. With this approach, Nielsen can identify the most effective strategies, optimize budget allocation and better understand how different marketing efforts work together to drive success.
Conclusion
L'IA stimule les résultats publicitaires
AI isn’t just a fad or a futuristic planning tool – it’s a performance driver. While there are many use cases for single format campaigns, it’s undeniable that AI-powered campaigns can drive a bottom line impact. However, the real expert should find and harness that perfect blend of AI tools to maximize impact and always keep measuring to uncover the truth in today’s ever-changing AI landscape.
“The Nielsen research rigorously validated the significant impact of Google AI-powered solutions across both brand and performance campaigns. The data demonstrated substantial ROAS improvements over manual methods, along with valuable synergies between AI formats. These insights derived from the Nielsen study reinforce advertiser confidence in the tangible results they can achieve with Google AI.”
Shannon Trainor Stark – Managing Director, Solutions and Thought Leadership, Google
Vous souhaitez vous entretenir avec notre équipe d'experts ?
