AVIVA boosts digital campaign efficiency via machine learning automation

Full digital campaign automation will transform how the industry optimizes digital marketing. Media agency Zenith has achieved a breakthrough in digital campaign automation that significantly boosts advertiser effectiveness and will fundamentally change how agencies and clients optimize digital media efforts.

Digital campaign automation

Over the past six months, a team of data science experts and strategists from Zenith worked on an advanced digital campaign automation project using proprietary machine learning technology and custom-built algorithms. The goal was to explore how machine learning could efficiently process large volumes of data and automate the most complex and time-consuming aspects of digital campaigns. The global Zenith team led the execution of this pilot project.

Cookie and sales data

By collecting data from Aviva’s ongoing campaigns, the team combined cookie-level campaign data from a leading DSP (demand-side platform) with the client’s corresponding sales data. Applying Zenith’s custom machine learning algorithm, the team precisely attributed conversions to specific user interactions.

Increased campaign efficiency

Closed-loop automation

Next, Zenith — for the first time in the industry — automatically optimized Aviva’s digital campaigns by feeding the algorithm’s results back into the DSP system. This closed the automation loop: data collection, attribution, and media plan adjustments across all digital touchpoints now occur automatically.

Further, Zenith integrated its own data on demand-influencing factors into the algorithm to enhance the effectiveness of automatic media plan changes. This allows the automated optimization to consider factors such as price impact on sales or creative performance used in campaigns.

Full digital campaign planning automation is powered by cloud computing data processing technology. Throughout the process, the client retains full ownership of their data.

The application of machine learning and automation delivered improvements in Aviva’s CPA* metrics — 6% for search and 10% for display campaigns.

Zenith leads in business digitalization. This groundbreaking project is part of our strategy to leverage data and technology to increase our clients’ business profitability — said Vittorio Bonori, Global Brand President Zenith.

The project will be rolled out across all markets where Zenith operates.

Article glossary

Attribution — assigning conversions to channels. A rule allocating credit for a conversion among touchpoints: last click, first click, linear, time-weighted, or data-driven model.

Why it matters: The choice of model shifts budget between channels — a business decision disguised as a tool setting.

When to apply: Configuring analytics, debating channel effectiveness, budget planning.

Risk of omission: Without conscious selection, default “last click” systematically overvalues channels capturing ready demand and undervalues those generating it.

Summary

Zenith conducted a pilot automation project for Aviva, combining DSP cookie-level data with client sales data. Zenith’s proprietary machine learning algorithm precisely attributed conversions to interactions, and the results were fed back into the DSP system — enabling automatic data collection, attribution, and media plan adjustments. Zenith is the first in the industry to close this loop.

Results include a 6% improvement in CPA for search and a 10% improvement for display. The algorithm also incorporates demand factors, including price impact on sales. The project will be deployed across all Zenith markets.