A phenomenon we tend to see in ad tech is that, because of similarities in terminology, people from finance often try to apply concepts from finance to ad tech, or people in ad tech try to borrow concepts from finance. This includes trying to implement financial instruments (e.g. futures, options - and derivatives generally), calling programmatic buying desks “trading desks,” and - most recently - using the blockchain for stuff in ad tech. In this Lift Letter, we discuss the blockchain, what exactly it is, and a few potential applications to ad tech.
The blockchain was popularized as a technological piece of Bitcoin. One challenge that Bitcoin solved is that, as a digital currency, how do you make sure that a given piece of digital currency is not “double spent” - meaning if I own one bitcoin, that I don’t just spend that same bitcoin with two different vendors before each of the vendors find out that I did that - and also without using some centralized store of truth (if it’s centralized, either though a government or some other entity, this becomes the point of failure for the entire currency). This requires some sort of representation of the state of all bitcoins - who owns which ones and where they’ve been. This, in turn, requires that no one can control or inaccurately change this representation, and that the representation must be trustworthy.
Through the use sophisticated cryptography and distributed algorithms, the various computers that participate in the blockchain record encrypted versions of all transactions onto a publicly accessible “ledger.” In bitcoin, the ledger is based on the concept of a blockchain. The blockchain itself is a series of “blocks” that use a reliable method of encryption to point to the previous “block” for a set of transactions, the pointer to that previous block being saved in the next block. This goes back for the history of a given chain, meaning you can trust the history of a block chain’s transactions as long as the cryptography is valid. The blocks themselves are then redundantly distributed through the set of computers that use bitcoin (or whatever system is being based on the blockchain) and cannot be undone, only modified to reflect further transactions. This creates a universal truth for who owns which bitcoins (or whatever) and where they’ve been.
One of the challenges with blockchains generally is that it’s expensive and relatively slow to save something onto the blockchain. A transaction has to be distributed throughout the blockchain, and ultimately saved by a significant number of computers that are randomly distributed through the system. This simply takes time and costs money in compute time. For this reason, among many others, Bitcoin has struggled to gain adoption. That said, people have seen the blockchain and found a number of great uses for it, and it serves as the basis for all sorts of other digital currencies, including Ethereum (not exactly a currency) and Zcoin. Blockchains have also been extended beyond the realm of digital currencies, including serving to store who owns various stocks and bonds, representing the state of contracts, and similar concepts.
Blockchain is increasingly being mentioned in conjunction with ad tech. One example is NYIAX - the New York Interactive Advertising Exchange. NYIAX aims to allow publishers to sell their inventory through their exchange. The ownership of the inventory would be stored on the blockchain. Thus the buyer could resell the inventory to another buyer, and the ownership of and right to sell this inventory could be reliably known through the blockchain. Somehow, though it’s unclear to me exactly how, the blockchain representation of the inventory would eventually be turned into actual ad inventory. From the perspective of a blockchain allowing a distributed set of buyers and sellers to trust that the right to sell a piece of inventory is trustworthy, and to understand who actually owns it, NYIAX seems to be using blockchain in a well-considered fashion. That said, the ability to let arbitrary parties pass around ownership of a publisher’s inventory through various resellers seems like an objective that runs contrary to the direction of the industry. Specifically, through initiatives like ads.txt, ad tech generally is trying to bring the final buyer and seller of ad inventory closer together.
Another example of leveraging blockchain for ad tech is a company called MetaX. MetaX has created the “adChain” - a blockchain for RTB to detect fraud. The adChain registry uses Ethereum (a digital currency / contracting medium) to identify and store “fraud-free” domains. AdChain has created a fairly elaborate mechanism for the identification of these domains and maintaining the list in a decentralized fashion. There are many ways to fight fraud, however, it is unclear that the best way is through a decentralized list of good domains. Indeed, the main challenge with adChain is that creating / curating this list is not really the problem - buyers in the industry generally know “good” and “bad” domains and have expertise creating whitelists, but that traffic may not always be real on those “good" domains. The proposed resolution to this problem - a more salient problem in ad tech - as defined in the adChain white paper is entirely unrelated to the blockchain and is inadequate in the RTB context, meaning adChain doesn’t really do anything itself to address the most common forms of fraud.
It is quite possible that blockchain will have genuine utility as it relates to ad tech. That said, it must come in the form of being the best solution to an actual problem, and done in a way that solves that problem within the realities of the blockchain. Storing individual RTB impressions on a blockchain, for example, will likely never be a reality - it is too expensive and slow. Currently, we are in the phase of the hype cycle where people are trying to apply blockchain solutions to non-blockchain problems, or to use blockchain when other solutions are simply better and the benefits of blockchain are irrelevant. That said, there are likely challenges where having a distributed, immutable source of “truth” that can support transactions (slow) and supporting anonymity will likely have some benefit - but it’s not immediately clear to me that anything in the industry today satisfies all the criteria above.