A leading independent ad platform company needed an automated, highly accurate and extremely fast screening solution that provided real-time updates and was capable of scaling up to meet spikes in traffic volume.
Client Profile:
LKQD™ is a leading independent video advertising platform that enables publishers to serve video ads across multiple devices and formats.
The Business Need:
The LKQD platform works with thousands of publishers who utilize their self-serve platform to serve video ads globally. The platform gives publishers a self-service tool to manage and grow their video monetization business. LKQD was using self-declared categorization across their publisher inventory to ensure that brand safe sites were appropriate for inclusion in their platform.
With a commitment to ensuring brand safety, LKQD pre-screens all publishers prior to on-boarding; however the sheer volume and manual effort created a resource-intensive, unwieldy and potentially error-prone process. An automated, highly accurate and extremely fast screening solution was required. Additionally, since many of publishers represented new sites, the LKQD Platform needed a solution that provided real-time updates and was capable of scaling up to meet spikes in traffic volume.
LKQD identified a number of requirements necessary for this categorization and on-boarding process including:
| The evaluation criteria included: | |||
|---|---|---|---|
| Granularity: The solution needed to provide granular categorization of inventory in order to provide the strongest brand safety | |||
| Accuracy: an extremely accurate database of categorized URLs | |||
| Coverage: breadth of web coverage for the broad range of publishers with which they worked | |||
| Speed: Rapid categorization of new web pages | |||
| Scalability: a solution that could accommodate high query volumes with large spikes in volume due to traffic. | |||
| Responsiveness: zero latency in categorization | |||
| Flexibility: Custom URL category mapping with the ability to map the categories of the URL database to the Platform’s custom categories as well as those established by the IAB | |||
| Multi-language support: Working with global ad networks required a solution that offered broad multi-lingual support | |||
| Exceptional Service: Focus on client success from initial evaluation through deployment | |||
The zvelo Solution:
LKQD evaluated zveloLOGIC with a specific emphasis on speed and identification and categorization of brand-safe and inappropriate content across a wide range of websites.
| The evaluation criteria included: | |||
|---|---|---|---|
| zvelo categorizes the web in nearly 500 categories with automated support of two dozen languages, including multiple objectionable categories for brand safety as well as IAB Tier 1 and Tier 2 categories. | |||
| zvelo uses a combination of artificial intelligence, machine learning and human quality assurance to generate high accuracy | |||
| zvelo actively categories 99.9% of the Active Web | |||
| zvelo provides real-time updates for new and dynamic websites | |||
| zvelo delivers zero latency via microsecond lookup times | |||
| zvelo supports customers doing billions of queries per day | |||
| zvelo allows custom URL category mapping to the DSP’s category sets, as well as those established by the Internet Advertising Bureau | |||
| zvelo provides full white-glove support through evaluation and deployment | |||
| zvelo offers SDK and API implementation options | |||
The Results:
After integrating zvelo, LKQD tested the performance throughout high traffic periods, running nearly 300 billion queries in a few days. Based upon zvelo processing speed, accuracy, coverage and multi-lingual support, LKQD was able to achieve significant optimization of their on-boarding processes which reinforced the LKQD commitment to quality and brand safety.
