Sthan - Virtualizing Physical Addresses

Sthan replaces Postal Addresses with a privacy-protecting reimagination of what it means to physically locate a person or a place

By some estimates, as many as 3.2 billion people do not have an address. A lack of an address can lead to exclusion from essential services like banking or health care. Even when addresses are available, they can be woefully inaccurate or inadequate to help find a location. This results in costly delays in the first/last mile, decreasing affordability of all solutions that depend on moving things from A to B. Moreover, these costs are regressive, they affect the poor more than the rich. Bad Addresses cost India $10-14B annually, ~0.5% of the GDP.

Humans know how to communicate where they stay in real-life. The problem is that the way a user communicates it, may not be adequate for certain organizations. The problem isn’t people, it is the address text itself. It is defined by organizations to find places in a manner convenient to them. Think about how unnatural a pincode is.

Fragmented geocoding solutions are appearing, which only granularize the pincode, not replace it. Google provides a 11 digit plus code. What3words, uses 3 random words to describe every location in the world. Companies like Zippr enroll door-to-door by actually assigning every door a 9 digit number. These solutions all have specific advantages, and fatal flaws discussed here. Ultimately, very few are finding user traction.

Sthan is a technology-neutral protocol. Sthan offers the user consented sharing and privacy-by-design. It is designed for hard addresses. By being interoperable, it offers great flexibility, and relies on a market ecosystem for adopting geocoding scheme(s). Basically, Sthan is a new way of answering “Where are you?”

Sthan is human-centered. We understand that addresses are meant to locate people or for people to locate a place. Sthan is designed for real-time/mobile/dynamic physical address of the user - “Where are you living?” as well as “Where are you now?”. A user’s identity doesn’t have to change when they move, their Sthan can stay the same.

Sthan protects privacy by only resolving those attributes that are necessary in the context to answer “Where are you?”. E.g. you may share your exact location with a rideshare, whereas you would only share the current city with your travel agent.

The architecture of Sthan is briefly described below :

First, a Sthan is

1. an Address of the format user-name@entity e.g. tanuj@fedex, 555-1234@ups, tajmahal@india

2. A series of address data attributes such as Geocode, unstructured address text, structured address attributes (city, state, country, pincode), etc.

Second, each Sthan has API endpoints for :

1. Authorized entities to encode and resolve addresses at an attribute level

2. Creation of one-time tokenized addresses for sharing with logistics providers

3. For users to give consent, retrieve user logs of resolved addresses

If the protocol is implemented, and a central switch is made as a public good (i.e. non-excludable), We expect market players to start building innovative applications on the Sthan protocol.

Many attempts have been made at address standardization - what3words, google +codes, robocodes, etc. They all suffer from known problems. All these problems stem from the fact that these are private solutions forced to find a one-size-fits-all approach globally. -Sthan is a layer above these solutions, presenting an interoperable framework, which incentivises inclusion. -it provides for market-driven, context-specific solutions for adoption & innovation.

-it allows for mobility and migration, a need of developing nations.

-it is privacy protecting. Imagine tokenized addresses and access logs, to help users understand who is looking at their address.

Sthan is inspired by UPI’s VPAs & Aadhaar’s tokenization.

Currently a user's address passes in plaintext in any system. People can get doxxed or swatted. Sthan virtualizes the address i.e. instead of sending 123, Maple Drive, Bangalore, my address will either be a user generated (e.g. tanuj@indiapost) or system generated token (eg. 1fksd57asd@zippr).

Querying of addresses will be attribute-based, limiting information shared.
e.g. APIs for “is tanuj@indiapost in Bangalore?”, etc.
Users will have a log of who has resolved the address.

Address Tokens may only be resolved by authorized entities with consent. Even if company A shares token with company B, they can’t resolve it.

We think Sthan is a great fit, and potentially even more powerful when implemented within digital identification systems. Specially for those who have never had an id.

First, there is a high overlap between those who do not have an id, and have no/difficult addresses.

Second, Digital Identification Systems bestow people with a digital existence. If we issue a virtualized Sthan address at the same time as we issue a digital id, then we have a potent combination. One can allow users to either choose, or procedurally generate a Sthan address that is part of the digital id (and on the printed id card, if there is one). People will have an id number to say “I am me”, and a Sthan address to say “I live here” that has a high degree of trust.

Third, Tokenization of address can be implemented in eKYC-like systems.

Fourth, Users can update their own addresses and geolocation by authenticating themselves via digital id.

For a digital identification system - Sthan replaces a unstructured address text with a simple virtual id (e.g. jack@post), which can be chosen by the user so that they can easily remember it. Every Sthan address is linked to a user, and digital identity solutions can extend this linking seamlessly.

For end-users, the “user friendliness” will be provided by market or government players, who implement the standard APIs on top of the digital identification system.

Let’s see 2 apps built on payments protocol UPI as example.

 BHIM emulates digital wallets. 

Whatsapp, you send money like you send memes.

Our solution is a technology-neutral protocol. It is a set of APIs that any entity can communicate with. The central switch required to operate can be built as a public good, or we can create a market of switches that all implement the protocol. Addresses can be created by anyone, and resolved by authenticated entities.

The whole system is built on the idea that instead of giving everyone the same full-address can we instead give them a virtual address. This virtual address will only resolve for authorized entities. Moreover, it will only resolve the minimal information necessary for the task.

Connectivity is not required by the user except at the time of onboarding. Moreover, Sthan is completely retrofittable with existing address systems. Where an explicit virtual address isn’t available, Sthan can work with unstructured address text, and providers to geocode. The interoperable network allows Sthan to fetch the best geocode for that text from multiple providers, incentivizing clean databases.

Low literacy and numeracy can be solved by UX layers. An example is given above on how Whatsapp vs BHIM use Virtual Payment Addresses. Moreover, unlike any other solution, we allow users to choose their own address with no constraints.

Over the next few years, we want to ensure an interoperable, privacy-protecting, geocoding system in India. Simply put, there should be no address that is hard to find, without making everyone’s address easy to find. The second order effects of this through more efficient logistics, e-commerce etc. will begin to kick in.

IndiaPost, which delivers most mail in the country would be able to give a kickstart to Sthan. In return, IndiaPost will be able to unlock the value of the local knowledge of its large network of post offices, letter boxes and postmen through payments for address resolution.

Tanuj Bhojwani - Fellow at iSPIRT. I consult on many population scale platforms for societal problems - India Stack in the past, and now the upcoming Logistics Stack. We've been designing a Sthan like system for address standardisation as part of the Logistics Stack.

Sai Sri Sathya – Founder of S20.ai, a startup democratizing AI in a privacy-preserving manner. I’m also a researcher collaborating with the Camera Culture Group at MIT Media Lab, mentor at REDX, a global AI for Impact initiative through innovation labs and student clubs and volunteer at iSPIRT, helping create large scale digital platforms for public good.

We have the right mix of skills and experience to solve such problems.

Tanuj has led the effort to go from no drone regulations to creating a drone policy framework, creating a digital public good called Digital Sky, and enlisting market support on the idea.

Sathya has done extensive research on addresses and its economic impact in India and was a part of a social impact initiative by Facebook’s Connectivity Lab in developing open source generative street addressing scheme (Robocodes).

He has invented advanced proximity and mesh technology based protocols that can solve communication and several other emerging market problems.

In India, we have the nod of the Indian Government to build a Sthan-like solution.

Both Sathya & I are motivated by solving this deep problem for the economic value it will unlock in India (Sathya has a written a paper explaining how much), and hopefully the rest of the world.

Since we do this as a pro-bono consultants to the Indian government, we need very little money to solve this, mostly for occasional travel. In the past, we've been able to find sponsorship for such travel through philanthropic foundations and/or not-for-profit organisations.

The expenses to implement this in India will be borne by the government, as it is designed to be a digital public good.

We're creating Sthan in India because of our unique advantages, and networks in India. But the addressing problem is global. There are many startups and tech giants trying to solve this problem, but with a keen eye on accruing the platform power. The inherent network effects favours monopolies.

We believe that address privacy and address standardisation can only be solved together if we use the public goods philosophy, i.e. non-rivalrous and non-excludable. Hence, building them into platforms like MOSIP makes most sense. Moreover, in terms of adoption, issuing standardised Addresses while issuing id can lead to leapfrog adoption.

There are some key incumbents who have built a business on maps, address scraping, geocoding and reverse geocoding. These businesses tend to enjoy the classic network effects which have concentrated the industry. Sthan creates an underlying framework for interoperability and opens up the system to multiple actors who can onboard new users and drive adoption. The main challenges we foresee :

1. Key incumbents will want to resist Sthan.
2. A well-designed incentive structure to keep running the data marketplace smoothly