Blockchain Business Use Case (1) — Digital Advertising settlement

Now that speculation is not as trendy…

The sentiment about crytocurrencies has turned bearish recently. Bitcoin has once fallen below US$4,000 on the 25th of November. Initial Coin Offering (ICO), a method for startups to raise capital, has dropped by over 80% in total capital raised. As the speculation appeal of crytocurrencies is fading, the interest in blockchain technologies is also dwindling somewhat for the general public.

Regardless, even blockchain optimists find it hard to deny that so far no business use case has deployed blockchain technologies for production purposes at scale. Private blockchains have some good pioneers in interbank settlement, trade finance, tokenized asset trading, many of which are either in proof-of-concept stage or still in their infancy. The most famous production use case for a public blockchain, on the other hand, is still CryptoKitties. Articles or podcasts about blockchain technologies focus a lot more on the technology instead of any practical use case that the technology can bring about.

Still the king for public blockchain use cases

I intend to put together a series of business use cases that could make use of the key components of blockchain technologies. While I may get the use case ideas from a particular blockchain protocol (as in this piece where I learnt of the use case from Zilliqa), I will focus more on how general blockchain technologies can derive business value. I will stick to a largely fixed use case assessment framework (heavily borrowed from Gideon Greenspan of Multichain and Antony Lewis of R3), which includes the below questions:

Do you really need blockchain?

1. Who are the participants, their jobs-to-be-done, and the current workflow?
   Blockchain is effectively a type of database that multiple participants need to use and share. Similar to most databases, business participants need to record their business tasks on the database for the database to have any value. As such, it is important to understand what business tasks the participants need to accomplish and how they are accomplishing them now.
2. What are the challenges the participants are facing?
   Most use cases blockchain aims to tackle is already being solved in some other forms. To convince participants to switch to a blockchain-based solution — a non-trivial effort, participants need to observe some major challenges in the current solution.
3. What is the blockchain based alternative?
   After assessing the current situation, a blockchain based solution needs to be designed to solve the challenges.
4. How will the blockchain based solution address the challenges? Can the benefits justify a transition?
   This addresses how the blockchain based solution will solve the challenges in practice and if it brings sufficiently large net benefit to the users to justify a transition.

Let’s not make the new workflow more complicated…

5. Can a trusted intermediary solve the challenge(s)?
Note that a trusted intermediary can often times host this shared database for the use of the participants. For instance, despite the recent poor run of events, Facebook, a trusted intermediary for social networking database, is doing a pretty decent job to maintain the social media account of over 2 billion monthly users. Majority of these users may not be overly bothered by (or even aware of) all the recent controversies, i.e. most users do not observe major challenges using Facebook to fulfill their social networking needs. If that is the case, the benefit of a blockchain based social network may not be able to lure enough users off Facebook to kickstart its own network effect. Afterall, who wants to hang out on a social network which gives back data ownership to users (a major benefit of the blockchain alternative) BUT with only a few of their friends on it?

Use Case Assessment: Digital Advertising reconciliation and settlement

(As said, this use case originates from Zilliqa, a blockchain protocol based in Singapore. They have been working with a few partners in this use case, which they call Project Proton)

1. Who are the participants, their jobs-to-be-done, and the current workflow?

Digital advertising is a gigantic and complicated value chain with a market size of US$88 billion as of 2017. To find out what happens within a split second between when the user starts browsing a page and when the content is shown to him/her, here is a good short intro video. For the purpose of this article, let’s focus on the below 4 types of value chain participants.

a. Advertisers (e.g. Nike or P&G) who want to maximize the number of views of their digital advertisements from the target audience

b. Advertising agencies (e.g. Mindshare, Carat) | Service providers (e.g. Mediamath, Beeswax) who want to help their advertiser clients maximize the number of views while minimizing their costs.

c. Ad exchanges (e.g. Rubicon, DoubleClick) who want to attract a large pool of buyers and sellers to facilitate transactions and fee income.

d. Publishers (websites with traffic) who want to maximize advertising income while making sure the advertisements fit its target audience.

The current workflow has many variants and much more details but broadly speaking:

1. It starts with advertisers working with agencies to set an overall campaign focus, strategy, and budget.
2. If the focus includes programmatic buying, they may work with specialized service providers and bid targeted “views” (i.e. targeted audience viewing the advertisement) from ad exchanges.
3. The advertisement will then be placed in the respective publishers’ sites real-time when the targeted audience browses the publishers’ sites. Views will be logged.
4. Agencies advance the payment to service providers and ad exchanges, who in turn pay the publishers.
5. Advertisers will pay agencies by the number of views typically up to 30 days after the campaign. The actual delay is case by case and some agencies charge regularly instead (e.g. quarterly)

2. What are the challenges the value chain participants are facing?

a. Advertisers: First, transparency — advertisers currently have limited information from agencies about the purchased publisher mix , the margin mix and pricing structure of campaigns, etc. Data typically comes from the agency / service provider platform, which makes them subject to manipulation and difficult to audit. Marc Pritchard, P&G Chief Brand Officer, has publicly urged the industry to tackle the transparency issue publicly in early 2017. Second, fraudulent views — payment from advertisers is determined by the number of views. Many of these views could be generated by a robot (i.e. computer program). According to Juniper Research, advertisers could lose $51 million a day on ad fraud.

Just kidding… the real robot is more like a few lines of code

b. Advertising agencies | Service providers: advertisers typically do not settle right after the campaign and the payment terms can be up to 30 days. Since agencies need to pay downstream even before the campaign ends, they are effectively funding the advertisement costs for the advertisers during the campaign plus any payment delay period. Rubicon CEO Michael Barrett said that “long payment terms are a tax on the industry for technology companies that are responsible for programmatic transactions”. Late payment means agencies will need working capital financing to match the revenue and expense, raising the cost of operation.

c. Ad exchanges: while agencies usually take the hit of late payment so that service providers and Ad exchanges receive payment on time, it also means part of the spending from advertisers is lost to the cost of working capital financing. Ad exchanges will make less as a result. Ad exchanges also face challenges in terms of filtering out low quality / fraudulent publishers that may impact advertisers’ interest in the exchange.

d. Publishers: outside of the known reputable publishers like Google and Facebook, it is difficult for honest publishers to compete with fraudulent ones. Real audience (whose views advertisers want to pay for) take good content to attract and content costs money to create. Robotic audience costs practically nothing to create. If advertisers cannot easily distinguish the honest from the fraudulent, digital advertising market effectively becomes the “Market for Lemons”, where the honest sellers would be slowly driven away.

3. What is the blockchain based alternative?

Credits: Zilliqa Blog

Leveraging a blockchain smart contract layer, the new workflow (illustrated by Zilliqa’s diagram above) goes like below:

Before the campaign starts…

1. Advertisers work with an agency (Mindshare) to identify in advance all the participants involved and the contract terms, including the targeted audience, the price per view, the revenue split, etc.
2. For this to work, advertisers need to deposit the daily cap amount into the smart contract. This could be a stablecoin (a kind of crytocurrency that is pegged to a fiat currency like the US dollar). A less elegant alternative is to have the advertisers deposit to payment providers obligated to execute payment order from the smart contract.
3. An ad verification company (IAS) is brought in to make sure the generated views are not from a robot. The advertiser needs to agree that views verified by the ad verification company is already a payable view. Note that, with or without blockchain, ad verification companies can be brought in (though reconciliation typically gets more complicated with more players…)
4. Once all the parties agree, the terms are written into the smart contract layer and validated by a blockchain consensus protocol (e.g. Proof of Work of Ethereum, practical Byzantine Fault Tolerant of Zilliqa), which makes them immutable and ready to execute.

With the contract in place, the campaign can then be launched…

1. The agency (Mindshare) then works with a service provider (Mediamath) to buy from an ad exchange (Rubicon) as usual. The target audience requirement is sent to the publishers. Target audience view data is then verified by the ad verification company (IAS) — this part is the same without blockchain & smart contract.
2. The ad verification company then sends the verified viewability data (in encrypted format to ensure security and privacy) to the smart contract. The data is processed and the smart contract logic triggers the payment to the ad verification company, the publishers, the ad exchange, service providers, and the agency named in the smart contract.

4. How will the blockchain based solution address the challenges? Can the benefits justify a transition?

From my high level outside-in assessment, the benefits of the new solution are as below:

1. Terms transparency: The advertisers now have transparency about the publisher mix and revenue split, alongside an auditable trail of verified viewability data — something they have already been demanding openly.
2. Auditable fraud detection: With an auditable trail, advertisers should be more comfortable using the an ad verification company because an internal analytics team of the advertiser can then analyze the result and hold the ad verification company more accountable. More auditable verified viewability data should over time incentivize the improvement of the ad verification solution.
3. Instant settlement: Agencies, service providers, ad exchanges, and publishers can now receive instant payment for their services without losing revenue to working capital financing (due to late payment from the advertisers).
4. Administrative cost saving: Without the blockchain solution, it is likely to have a lot of back and forth negotiation and reconciliation between the advertisers and the agency before the payment is settled. With the blockchain solution (combined with the verified data provided by ad verification company), this effectively becomes a straight through processing (STP) workflow that can shave off much of the reporting, negotiation, and administrative overheads.

While difficult to fully quantify in dollar terms in an outside-in assessment, the blockchain solution addresses many of the burning questions of advertisers and agencies. As such, the business proposition does appear to have a solid foundation.

5. Can a trusted intermediary solve the challenge(s)?

Currently there is no intermediary playing this role in reconciliation and settlement. The most neutral participant in this value chain is the ad verification company. However, it is unlikely that advertisers want to be so dependent on any one ad verification company to play this role. Otherwise, the ad verification company that can sign up a large enough number of publishers will kickstart the generation of network effect, creating much more market power (much like a credit card network or social network). The remaining participants in the value chain all have conflicting interest and cannot play the role of the trusted intermediary in this workflow.

Challenges ahead but looks good to start

All in all, my view is that digital advertisement payment settlement is a solid use case for a blockchain based solution. Having said that, this solution is certainly not without its challenges. For instance,

1. Viewability data volume: Is there a blockchain-smart contract protocol that can handle the volume of viewability data? Number of views can easily go to billions if the solution is to scale (3 billion internet users X easily hundreds of ad views per day per user). This is a lot of data to be processed by a public blockchain network. Do we need to process every view on chain? The typical objective to process the data on chain is to ensure its immutability, which makes economic sense for high value data like a bitcoin or other digital asset. Given the low value at stake of each view (e.g. The average cost per click of an online Facebook ad is $1.72, cost per view is even lower), does it make more sense to write the summary data on chain while storing the raw view data at the InterPlanetary Filing System (IPFS) instead?
2. Payment solution: What is the payment solution? Stablecoin is quite new so there is no time-tested stablecoin in the market at the moment. The most famous stablecoin, Tether, is far from stable. Historically, pegged exchange rate usually becomes the target for speculation attacks. If the stablecoin needs to be backed by a fiat currency 1-to-1, then how can the validators of the stablecoin be incentivized? Fiat payment triggered by a smart contract order could work but the solution provider charges additional fees and also has the potential to kickstart a network effect since it effectively becomes the trusted intermediary (in payment but not in reconciliation).
3. Commercial contract in the open: Given that the smart contract is in the open, leakage of ownership information of the participants’ public addresses would mean a commercial contract becomes fully in the open. Is there a leakage-proof way to keep the public address ownership information secret forever? Unlike bitcoin where the number of unique users using the network in a day can reach a million, it may be not overly difficult to figure out the owner of a public address in this use case if the protocol used does not have many other unique users with a similar transaction pattern. Transparency for the value chain participants is one thing, transparency for competitors is quite another.

Given that this use case is relatively new, it is more than possible that solutions will be developed for all of the above challenges at some points. Zilliqa, for one, is designed to use network sharding to solve the scalability issue. Would that represent some other trade-offs? It is hard to say at this point. Nevertheless, given the on-paper business robustness of this blockchain solution, I am excited to see how this use case will evolve in the years ahead.