Theta Collaborates with Chainlink to Fight Online Video Ad Fraud
Theta has been collaborating with Chainlink in the past several months building an innovative blockchain-based technology platform to solve the multi-billion dollar Ad fraud problem, through increased levels of transparency and reliability about the performance data of online video data. Our solution, which is currently live on TestNet, uses Google’s Big Data solution, BigQuery, to process video performance data generated on the Theta Network and create a reputation score for each streamer based on their viewership. Chainlink oracles then broadcast the reputation scores to the Ethereum blockchain where they are stored as immutable records for online advertisers to reference when deciding where to allocate capital.
Our design combines the inherent capabilities of the Theta blockchain to redundantly validate video content data and Chainlink’s ability to secure off-chain data delivery, with the benefits of Big Data processing to provide advertisers a golden source of truth regarding online content viewership, filtering out ad fraud and false metrics. Doing so increases the effectiveness of online advertising by empowering more informed decision making about which content creators and platforms reach the largest number of unique human users.
Additionally, we plan to use Chainlink to expand upon this initial integration by leveraging its secure oracles to obtain additional off-chain data about viewership. This data will be used in the future to develop advanced multivariable reputation scores about online content, further increasing the reliability and transparency of online advertising campaigns. We are confident that our unique approach can unlock billions of dollars in savings for online advertisers both through our immediate and advanced solutions.
Video Ad Fraud: A Systemic Issue in Online Advertising
Video content platforms are now a definitive foundational pillar of the core Internet experience. Trend research shows that 6 out of 10 people now prefer online video platforms compared to live TV, and by 2025 over 50% of viewers under 32 will not subscribe to a pay-TV service. As such, the advertising industry is increasingly shifting more of its resources towards promotion through online video content sharing platforms. However, one of the common problems with determining the value of sponsoring video content is calculating the true value derived from it.
The difficulty derives from malicious content creators who are able to fake viewership data through the usage of click-farms and non-human bots, ultimately inflating their click-through rate and tricking advertisers into thinking an advertisement was viewed by real users. In fact, it’s estimated that in 2019, the worldwide cost of digital ad fraud was $19B USD, and by 2022 it’s estimated to grow to $44B USD. Not only does this result in reduced ROIs for advertisers, but it skews incentives as it redirects large revenue streams away from valuable publishers (content creators) and rewarding publishers with falsified viewership metrics.
The solution we have engineered with Chainlink will be the first blockchain-based approach to reduce digital ad fraud, and increase the effectiveness of each ad dollar spent.
Initial Implementation: Calculating Content Viewership by Tracking User Bandwidth
Our initial implementation uses Theta’s native blockchain ledger to track user bandwidth supplied to video data, BigQuery to process this performance data and develop a reputation score, and Chainlink to make the data available on-chain.
First, Theta’s decentralized network generates highly reliable performance data by assigning a unique video ID to each video streaming on the platform. Next, our network of distributed nodes provides bandwidth to video streams; the amount of bandwidth that each node provides is recorded on the blockchain and identifiable by each stream’s video ID. The Theta delivery network incorporates intelligent trackers which optimize and efficiently group viewers together for stream data sharing based on their proximity, connectivity quality, behavior and other variables.
In particular, to combat click-farms and bots, artificial intelligence and machine learning driven filtering algorithms can effectively detect bad actors or bots based on their interaction and sharing patterns, and rule them out from the sharing groups. In addition, randomness is injected into the peer grouping process so that any undetected bad actors would have minimal chance to be grouped together. Instead, they are more likely to be grouped with legitimate users, and thus provide real bandwidth and augmented utility to the network. This process guarantees the data generated and recorded by the Theta blockchain are robust to manipulation, and can provide an accurate picture of how each video stream is being consumed and shared among real users. Thus, this on-chain record of bandwidth consumption provides tamper-proof evidence that nodes are legitimately expending computing resources to provide bandwidth to online streams.
Theta’s validated viewership data is then transferred into BigQuery through Pub/Sub, which is an open-source Big Data ingestion service. Since the Theta blockchain is natively compatible with the Ethereum Virtual Machine (EVM), for which Pub/Sub already supports, data from the Theta network is easily integratable. From here, Blockchain ETL and other statistical analysis functions can be performed on the viewership data to generate deeper and refined insights into each content creator’s true video content metrics and how they stack up against others. Once analysis is complete, the end result is a reputation score for each content creator based on a more trusted set of viewership data.
Chainlink’s decentralized oracle network then utilizes a PubSub external adapter to redundantly fetch and securely broadcast this reputation score onto the public Ethereum blockchain where it is stored as a tamper-proof on-chain record of node operator and video content performance. Advertisers can use this highly accurate and reliable set of data to decide where they should place ads and to confirm their true performance versus what they paid for ad placement. Advertisers can now see the true ROI of their ad campaigns and more efficiently deploy their capital.
The data delivered on-chain by Chainlink can also be used within smart contracts on the Ethereum network to automate contractual agreements between advertisers and content creators including payments and reward programs. Lastly, a vast array of decentralized applications, such as prediction markets, can be created that execute based on the data delivered on-chain regarding content creator performance.
Possible Future Development: Create Multi-Variable Reputation Scores
While this initial implementation is a major first step towards mitigating the effects of video ad fraud, we hope to employ a defense in depth approach by bringing a diverse set of data on-chain to further harden the reputation score against any form of ad fraud. Also, we can make the reputation score available on the Theta blockchain, enabling Theta nodes to store both the original unprocessed and the final refined data.
Given the flexibility of Chainlink’s external adapters, we can make any piece of off-chain data available on-chain and use it as a variable in the calculation of real content viewership. This allows us to mitigate ad fraud to the fullest extent possible while being able to adapt alongside the ever evolving fraud market.
Conclusion
Building on the existing data validation guarantees provided by the Theta blockchain, Chainlink provides the oracle solution needed to securely bridge our data to and from various on-chain and off-chain environments, as well as access key metadata needed to further filter bots and click farms. This could fundamentally alter how advertisers analyze the effectiveness of content and how they design contracts with their clients.
Making decentralized, well-validated data available to all stakeholders in a business process brings new levels of reliability and transparency to the whole advertising, media, and entertainment industry. Advertisers can now ensure that content viewership data reflects reality and is resistant to the current forms of ad fraud, resulting in more effective deployment of advertising capital and more fair compensation to honest content creators. This opens up a wide range of possibilities previously impossible by lowering the trust barrier required to obtain trusted data and ensuring each party in an advertising deal is held to the highest standards and integrity.
