“Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Eric Berry, CEO at TripleLift.
Such is the state of the tech industry today that companies are trying to solve nearly every problem with the blockchain.
Advertising technology is not immune from this trend. We have seen blockchain applied as a solution to nearly every facet of the industry, yet never with any material effect. Should blockchain really be applied to ad tech problems?
The decentralized blockchain technology uses mathematical complexity to make certain assurances about the integrity of the contents encoded in the blockchain, while also making the contents available to all. It was popularized through its role with Bitcoin, but its roots can be traced back over a decade before the cryptocurrency.
The blockchain was selected for Bitcoin because of its particular needs. The cryptocurrency required a completely decentralized arbiter for transactions that could store transaction history and currency ownership without a centralized authority, because any single entity, whether a corporation or a government, could eventually come undone or be corrupted.
Cryptocurrencies also operate in an environment where there are clear incentives for malicious actors at every step. Thus, elaborate systems were put in place to ensure the integrity of the ledger, including computational complexity for adding items to the blockchain. A decentralized system where each actor cannot be trusted creates a material impact for how it runs and adds substantial complexity that should be added only when necessary.
Advertising also faces a number of challenges, including fraudulent actors, deceptive or secret fees and nonviewable ads. Yet for blockchain to be chosen for a solution, blockchain’s characteristics should be required – and there should be no other simpler solutions available that work better.
A blockchain could be used to record the viewability of an ad – in theory. Instead, trusted parties such as Moat and Integral Ad Science emerged. They are corporate entities with significant monetary incentives to honestly and accurately measure and report on viewability. The data structure behind the scenes didn't matter – using blockchain would have undermined the outcome – because industry participants could simply trust the overall incentive structure that compels these companies to report accurately.
Logs can be made available that store the underlying data for the right participants, which results in the same level of transparency as a blockchain. But the companies that provide viewability services were able to innovate faster and add more functionality because they were not beholden to the complexity of using a blockchain.
The viewability example can be extended to the industry at large. Ad tech is not an industry where centralization is necessarily bad, nor is it an industry where there can be no trusted parties. There are certainly many examples of bad actors – fraudulent traffic, undisclosed fees, auction shenanigans and more.
But for the blockchain to be the best solution, given the complexities it introduces, the same predicates that made the blockchain ideal for Bitcoin would have to be relevant: the need for a decentralized system and untrustworthy individuals at potentially every position in the ecosystem.
In ad tech, these traits are never really part of the challenges being solved. Private enterprises can instead stake their success on integrity, develop data structures and computation systems that are much more capable and can develop reporting frameworks that use the elements of transparency that are part of the blockchain.
Fees and fraud are often cited as use cases for the blockchain. Some think fees would be appropriate because they require multiple, often competitive entities to record the price paid for ad inventory. The blockchain could record what each individual says the fees are and thus provide a public accounting for reconciliation without needing coordination – but a third-party verification company that provides this service could do the same by staking its reputation on accountability and even making monetary guarantees to the same.
So, while the blockchain could, in theory, handle this challenge, it is not, in practice, a good solution given that it struggles to handle transaction speeds that are orders of magnitude below what RTB requires, and it also introduces unnecessary complexity. Corporations that use arbitrary data structures and make logs publicly available would simply do this better.
Similarly, for fraud, given that the blockchain represents a form of decentralized trust through mathematical complexity, the theory sounds appealing to combat ad fraud. Yet, in practical terms, the pain point addressed by the blockchain relates to trusting data, which is an entirely different problem than identifying fraudulent traffic or domains.
The blockchain uses the mathematical difficulty of identifying content that hashes to particular outcomes to verify integrity. But maintaining a list of fraudulent domains or fraudulent IPs is a solved problem – especially through whitelisting. Identifying fraudulent traffic involves analyzing behavior through deterministic and stochastic signals and relies on a centralized party, such as Google or White Ops, building large data sets and analyzing the signals accordingly. This is simply a different challenge, and not one best solved through distributed, nontrusted algorithms.
There are very appealing attributes of the blockchain, especially its ensured transparency. That said, the ad tech ecosystem would be wise to resist the temptation to embrace this particular shiny thing because more often than not, it is simply not the right solution for the problems of our industry.
That said, this is not exclusively the case, and there may well be some appropriate situations. They simply must be considered soberly in determining whether this data structure, and the constraints that it introduces, produce the best outcome.