Breaking down the Dwolla network to its core.
Dwolla, as a network… has no money. It’s not a bank. Dwolla does not hold money. The Dwolla network talks to banks on behalf of members of the network.
Your money, my money, is already virtualized. The first time I was explaining that concept to a bank they looked at me sideways and the room fell silent. That happened again and again and again and still happens today.
It’s happened in local banks, regional banks, and global banks.
The cash in your pocket is a UI. It’s a trigger to post to a balance sheet in a computer. That check you write, it’s a trigger to get a balance from point A to B. The card you carry — just calls a server and tells the money to go from you to the person or company they represent who swiped it. Each one of these sentences could be a blog post on its own.
Dwolla is an efficient trigger. It is real-time by default, and there is nothing physical in the whole equation except for the computers that access the network and the people writing code, answering e-mails, and talking about it, to make the network better.
Dwolla, is built on the internet. Dwolla, is code.
There are no cards, checks, or paper. If plastic cards didn’t exist… Dwolla would work… If checks didn’t exist… Dwolla would work… If there was no cash…. Dwolla would work.
Dwolla doesn’t use hardware. It doesn’t have borders and it could care less what currency you transact in, from a technical perspective. From a business perspective we’ve limited it currently to the US and USD to focus on growing in one market. The network can learn currencies like your computer learns new characters and part of the thesis is that the network should always be currency agnostic.
Let me explain all that…
Someone explained Dwolla back to me in this way one time and I’m going to just steal it from them. It’s also the way that I’ve seen multiple fed faster payments initiatives described in the theoretical form.
There are 3 core pieces to the Dwolla network.
Put these 3 things together and you have a completely virtual way to find who or what you want to pay/buy, record the transaction, and report to all relevant parties (legally), and actually get the money from point A to B.
Simple in theory. Hard in practice.
I’ll explain each generically. The ledger concept is something popularized largely by Bitcoin and has helped us evolve our thinking from a ledger being something you have because you must record that a transaction happened, and turned it into a feature. You could think of this generically as “accounting” in one sense, but it’s much bigger than that.
I’ll touch more on the ledger later.
You need to be able to pay who you want with the network. This is one of the basic functions of any network but it’s something that most people think about what they think of Dwolla. It’s Dwolla.com, the mobile apps, and using the API to tell it to pay someone. It’s the layer that connects the two parties.
Getting permission to include people in that directory takes time. Today we’ve successfully deployed:
- Private search for individuals with user level control.
- Public indexing and private search for businesses with user level control.
- Alias accounts for those who need multiple entities indexed (in private beta).
- APIs for allowing buyers and sellers to close transactions without coming to Dwolla.com.
The directory is simply how people who want to engage in a transaction find one another or connect to close the transaction loop.
This is one key difference that some people love and some people hate about Dwolla. Dwolla, is driven on a redundant, secure, permission based ledger system.
The ledger is essentially the heart. It is where all the transactions are recorded. The Ledger is also the thing that everyone who participates in the network utilizes to prove that a transaction occurred.
This is an idea, which has recently been popularized by the Bitcoin’s blockchain, has been at the core of Dwolla since its inception and was made accessible over three years ago when we opened our network.
Where Dwolla’s ledger is very different from Bitcoin’s blockchain is that it has:
- Revokable access at the user level
- Tiered privileges at the ledger level
- Driven by authorizations at all levels
You can add multiple parties holding copies of the ledger as a feature and that day is likely going to come for Dwolla’s ledger.
Dwolla’s ledger doesn’t speak BTC though. It speaks USD. $, for example. But, you can teach it new tricks and any other currency. The currency is an archetype… a symbol. Adding a ¥ or € in front of something isn’t complex. Being a trusted and functional part of the ¥, €, or $ ecosystem is incredibly difficult from a social, regulatory, and compliance perspective.
$ — -> record (on ledger) — -> $
To have a functional ledger like this with the currencies that drive the world’s economies you have to have access or permission to electronically settle in that currency. That’s another piece, but it’s important to note.
A distributed permission based ledger system provides:
- Necessary leverage for financial institutions and enterprise companies
- Visibility for government, regulators, and policymakers
- Privacy and transparency for consumers
When it comes to privacy… part of this is driven at this point by my own thoughts on privacy. I admit fully, when Dwolla was started I wasn’t thinking about privacy as a feature largely because I didn’t know how borked the current system actually is.
The owner of the transactions (you, me) retains control of whom has access to our transaction in the Dwolla ledger. Unlike something like the blockchain where all your transactions are public, your Dwolla transactions are private until you select to share them. You share them by agreeing to share your transaction history over the API via the Transactions — Listing method.
You’re in control.
Dwolla doesn’t create currency or forms of value, it just talks to all the necessary parties involved to exchange it. Whatever you do in the network is your decision to share or not share. It’s private by default.
The ledger doesn’t make anything… it does however, record everything.
There is a glorious future application with the ledger as we add additional currencies. Using the ledger to create new efficiencies in currency conversion working with existing financial institutions is a logical application.
The money at some point has to get from point A to point B. One thing Dwolla did right from the get-go was look at money as data.
That is to say that when we talk to a bank through any messaging system (ACH, FiSync, credit interface, wire, and, and, and) there is never any physical asset transfer and the money is treated as bits. This was a good idea.
The settlement is the back end black box full of spaghetti that actually figures out the most efficient path to get money from point A to point B. In the US, building a settlement system requires a few core competencies that make it a bit harder to do well.
- Consistent access
- FI participation
- Regulatory approval & oversight
Getting access to the systems to settle on, or building your own like we did, is an exercise in insanity. It takes a long time and it’s difficult. But, it’s up and running and it’s virtually free to use.
What we have also found is that while a lot of companies sometime look at the simple cost of a payment, the actual benefits are in things like compliance, service, and security.
Doing these three things well means doing something in the US that hasn’t been done before and laying groundwork for lots of other things. For example:
- Recurring payments are just a /send transaction with a timer.
- Escrow payments have a trusted third party involved to act as a custodian.
- Credit is an agreement between 3 connected parties. Buyer, seller, bank.
All three wouldn’t work seamlessly without a functional directory, ledger, and settlement layer. All the while, physical money never even shows up… It’s all just digits in a computer.
One of the points of confusion is driven by one of the network’s biggest advantages… The applications are virtually limitless.
The money, man… It’s just not real.
Originally published at benmilne.com on September 8, 2014.