Introducing LinkExchange powered by Userfeeds

’It’s like Google AdSense But With Tokens’.

Maciej Olpinski
userfeeds
7 min readOct 2, 2017

--

TL;DR: If you don’t want to read the article and just play with the demo widget just go to http://rinkeby.etherscan.io and click on the Ether icon in the sponsored link.Thanks to Etherscan for supporting our alpha! PS. Make sure to have Metamask installed with some Rinkeby test Ether to be able to bid on links.

At Userfeeds our mission is to help users allocate their attention more consciously and effectively in the increasingly noisy and competitive environments of the modern Web.

The core belief that underlies everything we do, is that the global economy is in the middle of a paradigm shift toward a new model where the most valuable scarcity is human attention. This shift has much more profound implications than most people recognise today and will transform the future of business models, governance, politics and even individual lifestyle choices.

Individuals and organisations that will most successfully navigate this transition are the ones who can understand this new economic reality earlier than others. And quickly adopt new mental models and technological tools to take advantage of new opportunities. The challenge that most will face is that our external economic environments and internal, “natural” biological impulses evolved to constantly satisfy a scarcity of information that no longer exists.

At Userfeeds, our goal is to:

  1. explore and evaluate mental models and heuristics designed for the new economy of attention scarcity
  2. build blockchain-based software platforms that help users allocate their attention using these heuristics

We look at blockchains as a communications technology that allows signaling of incentives between economic actors that have scarce attention to allocate. We think that in the nearest future, all attention allocation decisions (‘what I want to see and what I want to ignore’) will be based on the algorithms that use incentive alignment between sender and receiver as one of the key ranking signals.

Apps, Developer Platform and Research on Attention Economics

Since our funding announcement a couple months ago, we’ve been thinking a lot how to prioritize efforts given our resources, maturity of the blockchain technology and goals we want to achieve.

We made a decision to focus on 3 areas:

  1. end user applicationsLinkExchange which can be described as ‘Google AdSense For Tokens’ aimed at users, advertisers & publishers in the crypto space that we built on top of the Userfeeds Platform.
  2. developer platformUserfeeds Platform which is the platform for app developers that enables them to run signaling metaprotocols and ranking algorithms based on tokens. LinkExchange is the first product built on top of that platform.
  3. research & conceptual worktheoretical work around attention economics that is foundational to both applications and the platform.

All three areas are on different levels of abstraction — with LinkExchange being the most concrete and practical and the research work being the most abstract and related to fundamentals of attention economics.
Compared to our initial plans, we decided to build the initial applications ourselves (like the LinkExchange) rather than target 3rd party developers from day one. It allows us to test the underlying platform and showcase the examples of what can be built.

We’re super excited about LinkExchange and think it might be a new way of monetisation for the online publishing industry and the method for reinventing incentives between users, publishers and advertisers.

In the meantime, we’ll be releasing more information on the technical and architectural choices behind Userfeeds Platform and papers on the underlying attention economics.

What is LinkExchange? — the elevator pitch

LinkExchange is like ‘Google AdSense for token-based communities with incentive-based targeting mechanism’.

LinkExchange offers publishers in the blockchain space a native method of monetisation by allowing them to display sponsored links on their websites in exchange for Ether or any ERC20 token.

For advertisers, the token they use to transact with the publisher has a dual use case:

  • it’s a medium of exchange and unit of account — just like USD or EUR they’d use in traditional ad networks.
  • but more importantly — it’s the incentive-based targeting mechanism.

In other words, advertisers use the token owned by the community they want to reach with their message. Token is a method of signaling their alignment with token holders and being more attention-worthy than the ‘outsiders’. Transparency and open nature of the blockchain allows users to audit every single ad that is competing for their attention.

Token-holders can always see who is bidding for their attention, what token they’ve used to buy the link and all other details they might be interested in. All transactions are recorded on the Ethereum blockchain and are fully transparent and auditable.

You can play with the demo widget today at: https://rinkeby.etherscan.io

It’s all about incentive (mis)alignment

Traditionally, advertisers used keyword, behavioural or demographic ad targeting methods.

While extremely effective, these methods require publishers or platforms to collect data about the users, in order to allow for more precise targeting. The more data about users they collect, the better targeting they can offer.

The challenge with this model is that interests of users, advertisers and publishers diverge quite quickly. Everyone is pulling in their own direction. Platforms have to silently collect more data into their private silos because that’s the only way to remain competitive. Users respond with ad-block limiting the amount of attention publishers can sell.
With incentive-based or token-based targeting, all three groups are incentivized to support each other by using the same token to exchange attention to capital.

For obvious reasons, this method can only work for a subset of advertisers who are interested in reaching holders of specific tokens. Also, the number of token-based communities is still small relatively to the number of communities on i.e. Facebook. But the hypothesis we want to verify soon is that for in the niche of token-based communities, LinkExchange can provide more value to all parties than Google AdSense or other 3rd party ad networks.

How It Works?

LinkExchange allows every Ethereum address to act as a decentralised advertising network. Just like with Google AdSense, the publisher configures the widget and embeds it on a website they control. Potential advertisers can pay to display ads in this widget.

The difference lies in how everything works under the hood.

In LinkExchange, the publisher specifies the Ethereum address that identifies the source data feed for the ads. We call that Ethereum address — a Userfeed address.

All links sent by advertisers to that Userfeed address are added to the data feed that will be ranked by the algorithm in the next steps.

Technical note: ‘Sending links to the Userfeeds address is done by our proxy contracts that forwards the money to the Ethereum address specified by the publisher and generates an event log with a link metadata (we call it a ‘claim’). The Userfeeds Platform then aggregates and parses the event logs using the algorithm selected by the publisher.

The algorithm can contain arbitrary ranking rules and its output is the ranking of the links.

To prevent spam and abuse the publisher can turn on the whitelisting option that allows them to moderate the ranking and only display links that were explicitly reviewed and accepted by them.

The publisher can additionally set how many top links from the ranking will be displayed in the widget. We call this number — a number of ‘slots’.

For example: there might be hundreds of whitelisted links but if the number of ‘slots’ is 10 then only 10 top links will have a chance of being displayed.

The links that are high enough in the ranking and make it to the slot are assigned a probability score which determines their probability of being displayed on the next impression of the widget.

Pay-per-Probability model

In the Pay-per-Probability model that is central to LinkExchange, the advertiser uses the token transfer to the address specified by the publisher to increase probability of their link being displayed in the widget on the next page view of publisher’s site.

The probability score is calculated using the algorithm specified by the publisher and the amount of tokens sent to an address is just one of the variables. The publisher might choose to use other variables such us as time, to modify the probability calculation to influence incentive structure in their link widget.

So depending on the widget’s expiration date (the publisher can declare how long they intend to maintain the widget) and expected traffic, the advertiser can place their bids.

But… does this mean more advertising that everyone hates?

On the surface LinkExchange widgets look just like the regular advertising widget so where’s the hype and do we really need more ads?. In reality, it’s like looking at the electric and internal combustion engine car.

They might look the same externally but their engineering, design, economics and business models are worlds apart.

The idea behind LinkExchange is to reimagine incentive structures behind ad placements that we’ve learned to hate so much. We use the same familiar UX of ad placements but feed them with data coming from the open data layer (the blockchain) with token-based economics.

The distinction between content and advertising becomes blurry and is dependent on the incentives of the reader.

Are you a crypto-publisher? Would you like to join the pilot?

We’re still polishing up the product but if you’re a publisher that serves token-hodling audiences and would like to pilot LinkExchange — please let us know (via the contact form at userfeeds.io or in the comments)

--

--

Maciej Olpinski
userfeeds

Co-founder of Userfeeds.io — the information relevance platform for the Web 3.0 stack enabling token-curated feeds and ranking algorithms.