03/31/2018

Case study: Optimizing ad pricing at Dictionary.com

Brock Ferguson
Co-Founder, Principal Data Scientist

Dictionary.com is the first and largest digitally-native English language dictionary, available anywhere at any time. Tens of millions of global monthly users visit Dictionary.com and Thesaurus.com for reliable access to millions of definitions, synonyms, audio pronunciations, translations and spelling help to alleviate their anxiety with the English language, understand meaning, and choose the perfect word, anywhere or anytime the need arises.

In a recent project, Strong Analytics partnered with Dictionary.com to design and build a solution that could leverage machine learning to optimize ad pricing using impression- and bid-level data.

Operating at a truly global scale

As one of the most popular websites and mobile applications in the world, Dictionary.com generates data on a massive scale each day. Moreover, with a truly global reach, successfully optimizing Dictionary.com’s ad-ops required a rich, segmentation-based approach. As online advertising is a dynamic, fast-paced market, the right solution couldn’t just work today, but must adapt to continuous market changes in the future.

Our solution

Strong Analytics built an ad floor optimization platform powered by Apache Spark.

Along with powering day-to-day production algorithms, this platform empowered Dictionary.com’s growing, in-house data science team with the ability to seamlessly prototype new algorithms and continuously upgrade optimization strategies.

We also built a new dashboarding tool that visualized trends and key metrics from each of Dictionary.com’s segments. This enabled ad-yield managers to make decisions with the help of accurate market snapshots and optimal pricing recommendations.

In an experiment, Dictionary.com saw 31% revenue gains in one segment for the new optimization strategy over the prior strategies.

Share: