Presented at Xignite API World 2016 - San Jose
Look in to the history and role of APIs in Financial services and the Fintech Revolution, assess the current situation in fintech and finserv markets and discuss what the future of APIs in financial services might look like.
8. xignite@sdubois xignite@sdubois
1910’s-60’s
TICKER TAPE
1990’s
CLIENT SERVER
2000’s
WEB
2010’s
APIs
1970’s
MAINFRAME
quotron
Vehicles
Terminals
Displays
Mobile Wearables
IoT
Robo-Advisors
Trading
1,000s OF BROKERS 10,000s OF TRADERS 100,000s OF ANALYSTS MILLIONS OF USERS BILLIONS OF DEVICES xignite
EXPLOSION OF END POINTS
13. xignite@sdubois
API HISTORY
HTML
W3C
Standard
IMG Tag
First pointer to
resources on
external servers
JavaScript
First smart
client
capabilities
IFRAME Tag
Complex resource
embedding
1996 Flash
Complex resource
embedding
XMLHttpRequest
First Web API
Implementation
AJAX
First smart
client logic
First API
(Maps)
JSONP
Call a server
without a server
REST
Roy Fielding
publishes his
dissertation
on REST
CORBA
First
standard
RPC
First API
First API
First API
JSON
First Spec
XML
W3C
Standard
xignite
First Commercial API
First API First API
First API
S3 EC2
HTML5
First Spec
SOAP
W3C
Standard
xignite
First Cloud API
First API
1990 1991
1993
1995 1996
1999 Feb-00
2000 2001
Jul-02Nov-00 Jun-05
Feb-05 Dec-05
1996 Aug-03 Jul-04 Aug-06 Sep-06Mar-06
Sep-121998
Oct-10Oct-08 Nov-13
FINTECH
WAVE
First API
Oct-08
Jun-07 Jun-08
App Store
iPhone 1
First API
Oct-02
Jeff Bezos
API Memo
24. xignite@sdubois
0
500
1000
1500
2000
2500
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
MARKETS AT AN ALL-TIME HIGH
DOW JONES INDUSTRIAL AND S&P500 2000-2015
$5,000,000
$500,000
$50,000 $5,000
2000 2005 2009 2011
FINTECH REVOLUTION DRIVERS
ECONOMICAL
FACTORS
SOCIAL
FACTORS
TECHNOLOGICAL
FACTORS
NEW MILLENIAL EXPERIENCE
COST TO BUILD APPS DROPPED
Maturing
Of App
Development
Scaling
of Public
Cloud
Social
Customer
Acquisition
Dawn
of APIs
29. xignite@sdubois
FINSERV BIG SQUEEZE
Fintech
User Experience
Pressures
Business Model
Pressures
Cost Structure
Pressures
Regulator
Balance Sheet
Pressures
Business Model
Pressures
Operational Cost
Pressures
Since 2013 Since 2008INNOVATION REGULATION
FinServ
INNOVATE
TRANSFORM
REDUCE COSTS
COMPETE COMPLY
31. xignite@sdubois
AWARENESS AT THE TOP
xignite@sdubois
"Silicon Valley is coming.“
Jamie Dimon
Chairman & CEO – JPMorgan Chase
Letter to Shareholders - 2015
32. xignite@sdubois
THE MARKET DATA CHALLENGE
Market data is one of the largest
expenses for many firms.
Previously neglected or hidden
is the cost of market data
infrastructure.
Legacy market data
infrastructure stifles innovation,
prevents transformation and
drives up costs.
xignite
MARKET DATA
@sdubois
33. xignite@sdubois
ColocationColocation
TYPICAL LANDSCAPE
Lots of legacy in-house infrastructure
Vendor lock-in and point to point interfaces
Significant compliance risks and costs
Many data silos across vendor technologies
Reliability and maintenance issues
Redundant or unnecessary expenses
Rigid, expensive and innovation killer
New app development nearly impossible
Unable to disconnect legacy systems
Costs continue to increase with scale
EXCHANGES VENDORS
Databases
Handlers
Databases
Analytics Engines
Databases
Tick Databases
App Servers
Cache Servers
Fanout Servers
Messaging Servers
Processing Engines
Search Engines
Trading & Algos
Handlers
Feeds FilesFeeds
Proprietary
Processing Engines
Vendor Proprietary
Servers
Redundancy
Compliance
Silos
Personnel
Upgrades
Lock-in
Point to Point
Legacy
Front-Office Mobile & Web
35. xignite@sdubois
INDUSTRY-LEVEL ISSUE
Much replicated data and infrastructure
• Same feeds from same exchanges
• Same files from same vendors
• Same databases to store and manage the
same data.
• Massive replicated data centers, storage an
processing
• Same hardware and system administration
burden
• Same staff to monitor and administer
No significant competitive benefit
• 95% of infrastructure is commoditized
• Billions of dollars wasted on replication
• Enormous innovation inertia
41. xignite@sdubois
WHY APIS MATTER
Reducing the Cost Structure
• Build a new tech stack from scratch
• Leverage the public Cloud
• Leverage third-parties
Improving the Experience
• Improve User interface & workflow
• Reduce friction and automate
• Increase access to information
42. xignite@sdubois
NEW THINKING
Trading & Algos Front-Office Mobile & Web
ColocationColocation
Databases
Handlers
Databases
Analytics Engines
Databases
Tick Databases
App Servers
Cache Servers
Fanout Servers
Messaging Servers
Processing Engines
Search Engines
Trading & Algos
Handlers
Feeds FilesFeeds
Proprietary
Processing Engines
Vendor Proprietary
Servers
Redundancy
Compliance
Silos
Personnel
Upgrades
Lock-in
Point to Point
Legacy
42
EXCHANGES VENDORS
Cloud?
43. xignite@sdubois
ADOPTION ROADMAP
First Use Case
• Customer Facing/Backup
• Mobile/Web
• Critical Innovation
• Critical Time to Market
• No New Infrastructure
• Get Comfortable
• Align with API Strategy
Adopt Broadly
• Deploy for all new apps
• Reduce dependency on legacy
• Eliminate fringe infrastructure
• Replace some data sources
Remove Legacy
• Onboard data sets
• Add contributed data sets
• Migrate legacy back ends
• Eliminate core infrastructure
6-18 Months
1-3 Years
2-5 Years
43
Hello Everyone
My name is Stephane Dubois. I am the CEO and founder of Xignite.
I am excited to be talking about fintech and APIs today.
I’d like to thank API world for giving me the opportunity to share this with you.
The name of this session is APIs are powering Fintech Innovation. What is next?
So this is not a technical session, it’s more of a strategy and a business one, looking at a specific vertical: financial services.
We are going to talk about APIs in fintech.
Specifically about APIs and market data.
First I am going to discuss the world of financial market data APIs.
Then I will talk about what happened with the fintech revolution and what role market data APIs played.
Then I will talk about what is happening in financial services and I will wrap up with talking about what is next.
As you know there is this fintech revolution going on.
And it seems that it will have a significant impact on the financial services industry.
What I’d like to do is first look at the role API played in this revolution….
This is an API conference so it should be no surprise we deal with API.
What we do is one type of API: financial market data API.
We bring data from the financial markets, and deliver it to applications that consume this data.
Our APIs are pure-play commercial APIs.
We are one of the few pure-play API providers out there where our product is actually the API itself.
People pay to use them.
They pay for the data and they pay for the convenience of using an API.
Some of you may not be familiar with Market Data.
Without market data, there is no finance. There is no financing. There are no markets. There are no investments. There are no startups.
Financial services is probably the industry that is the most dependent on external information.
So market data is the lifeblood of finance.
This industry was not born yesterday.
The concept of market data started when Paul Julius Reuter, who founded the news agency that bore his name started to use carrier pigeons between Berlin and Paris.
Those pigeons were speedier than the train, and gave Reuter faster access to financial news from the Paris stock exchange.
One might have said he was gaming the market.
It’s a large market by all standards: $26.6B in revenue.
But it’s stagnating. Even after the post crisis recovery, the industry is not recovering.
The reason is that the way people have been getting most of their data is via terminals like this one.
This is the latest and slickest version of the Bloomberg Professional.
This baby has a crapload of data and cool functionality and having one on your desk is the same status has having a Maserati in your garage.
On the downside, it’s not easy to use and it will set you back a cool $24,000—which is about ½ the median salary in America.
Terminals are amazing products, but this year, Bloomberg will be lucky if they sell more terminals than they lose.
The reason is that places like this that consume a lot of terminals are becoming a thing of the past.
The problem is that following the mortgage crisis, the financial service industry has experienced its biggest slump in employment in history.
We still have not recovered from the high of 2006. And it’s been 10 years.
There are less and less people using terminals. And this is unlikely to change.
And even as the industry re-hires, they are not re-hiring people that use terminals.
Because in these 10 years, the world has changed dramatically.
It started with the iPhone of course. And the Ipad. And the Apple Watch. Devices are proliferating. We have cars connected all the time. IoT devices.
Now you can check the value of your portfolio in the morning as you open your fridge to grab eggs and bacon to fix breakfast.
So if you look back at the industry, you can see that every wave of technology allowed financial information to reach more and more end points.
An order of magnitude more.
The industry started with pigeons in the 1800’s that would only help serve a few hundred individuals.
It then went to ticker tape with Edison at the beginning of the last century, a technology that reached a few thousand brokers downtown New York.
Then mainframes in the 1970 brought enough firepower to target 10,000s of traders around the world.
Client-server technologies in the 80’s made it possible for Bloomberg and others to target 100,000 of thousands of users.
And finally it was the web that allowed millions of individuals to get easy access to financial data. Yahoo Finance for instance, the granddaddy of all web sites has around 50M monthly visitors.
And now we are talking billions. Billions of devices of all kinds.
And many many of those consume financial data.
Planes trains and automobiles do.
Tablets, phones and computers do.
Maybe your nest device won’t but your Samsung fridge or TV does.
Gartner says that there will be 21 Billion connected devices.
And I estimate that a majority of those devices will consume financial data.
Then the question is how do you serve data to those 21 B devices?
The only answer o course is Cloud APIs.
What we have now, is the concept of Cloud APIs.
They are not just an API.
Like Amazon (the underlying infrastructure), they are designed to be a utility: Simple, Standard Based, and with unlimited power.
Cloud APIs are not toys, they are not novelties, they are not nice-to-haves.
They are APIs you can run your business on.
They are APIs you cannot afford not to run your business on.
The history of APIs of course was laid out by the invention of new technologies and standards such as HTML, XML, SOAP, JSON and others as you can see on the bottom of this timeline.
The first concept of pointing to resources on an external server can actually be attributed to Marc Andreessen and Mosaic (back in 1993—before it was Netscape), with the invention of the IMG tag.
Unlike a simple hyperlink, the IMG tag was pointing to and integrating a resource sitting on external servers.
Then Microsoft and FutureWave extended the concept with the invention of Flash and the IFRAME tag.
Microsoft again laid the framework for what an API would be by implementing the XMLHttpRequest API in JavaScript.
Great work done by Seattle again here.
The first commercial API can probably traced be back to SalesForce which launched theirs in February 2000, followed by eBay and Amazon.
We actually launched our first commercial financial data API in 2003, way before Flickr, Google, Facebook or Twitter.
All I can tell you, is that we were way way way before our time.
The world would never be the same after Amazon launched S3 in March 2006.
Two years later Apple launched the App Store and the next generation of API-based mobile applications—changing forever our expectations regarding software and user experience, but also sealing the destiny of service oriented architectures.
We launched our first AWS-based API–a cloud API—in October 2008—just a couple month after that.
And you can see, that as it relates to this timeline, the Fintech Wave—which started in November 2013 (according to Google Trend) is quite recent evolution.
http://apievangelist.com/2012/12/20/history-of-apis/
http://pamelafox-samplecode.googlecode.com/svn/trunk/tokyotalk/slideshow.html#slide37
Just for kicks lets look at Roy Fielding’s REST dissertation from 2000.
And for more kicks. Let’s look at the infamous Jeff Bezos memo.
For the story, we indeed did launch our first commercial API more than 12 years ago.
Way before it’s time I said.
But as you can see, persistence and survival can pay, as this little API, that has evolved but not changed dramatically since it’s inception, is one of the most popular APIs in the world today.
Today we have more than 1000 customers in 56 countries.
We have raised more the $37M including our $20M round last month.
We are one of the largest API providers out there and certainly the biggest fintech API provider I know of.
We have more the 650 API end points serving data from more than 150 data sources and covering more than 35 million instruments.
And all of those with a meager 35 employees based in San Mateo, CA.
The causes are a combination of economical factors, social factors, and technological factors.
The impact is that by combining the benefits of those technologies, new companies—or large ones who step up to them—can achieve 100 times the velocity in 100 times less cost.
The combination of all those factors caused the #fintechrevolution.
The problem we solve is that consuming and integrating market data with applications today involve massive amounts of legacy technology and heavy infrastructure.
These hidden technologies prevent change, stifle innovation and drive up costs inside firms.
And this is not just a firm problem. This is an industry level problem.
95% of the data a bank consumes is the same as other banks.
Everyone gets the same feeds from the same exchanges, the same files from the same vendors.
Everyone builds and replicates the same commoditized infrastructure.
This costs the industry billions of dollars and prevents it from innovating.
It’s like a frozen field of icebergs.
This means that the next generation software, is going to be based on Cloud APIs.
And just like what we have done before, we will rewrite most of our applications to take advantage of that new technology stack.
We did it from the Mainframe to Client Server
We did it from Client Server to the Web
And we are doing to do it from the Web to the Cloud.
And our kids will do it for the next big thing that comes around.
A great one is the problem of delivering real-time data to mobile users.
This has historically always been done in house.
Firms don’t even think “can I outsource this?”.
Fanning out real time data, like stock quotes, to tens or hundreds of thousands of users is a bitch—pardon my French.
You need tons of infrastructure, you need to scale for peak activity—total waste.
It’s hard to find programmers to code it. Hard to maintain and operate.
Firms figure it out but it costs them a fortune to build.
How do you do it with CloudAPIs?
Simple. A few lines of code in your mobile or web app using our new CloudStreaming APIs and your are done.
You can rollout to hundreds of thousands of users in no time at all.
If you have more users, no problem, our friend Scott is gonna let us crank up the volume and serve your needs at a fraction of the cost.
It’s perfect for new apps but you can even replace existing technology with it.
Let’s take another example: Engaging users with alerts about market conditions and investments.
Great way to get people to trade and talk to your advisors.
In financial services, if customers are not engaged, they are not making you money.
How do people do it today? Same way. In house!
Same problems. Tons of infrastructure and blah blah blah.
What’s the alternative: CloudAlerts
Same thing here. A few lines of code in your app lets your user define rules for what they want to be notified about.
Then we do all the heavy lifting in the cloud of crossing those rules with everything that is happening in the market
And when an event occurs, we call you back—via an API of course—so that you can in turn notify you users.
You can deploy this in no time at all for many different types of market alerts and hundreds of thousands or even millions of users.
And we are going to do it by taking advantage of a whole ecosystem stack of APIs starting from the infrastructure, the data like ours, to the Analytics such as those of our partners Estimize, to the transactional API of a Tradier, and the workflow APIs of an Advisor Software.
And of top of that, the business APIs created by the banks themselves.
These are hopefully some great examples of where a firm can start deploying the market data cloud.
The way you want to think about the market data cloud is a specialized technology layer on top of the infrastructure APIs provided by Amazon.
Market Data Cloud APIs provide easy, scalable and cost-effective access to the financial market data you need to run your business.
This includes market data, reference data, but also data like contributed data or data you get from other vendors.
So just like you are starting to move some of your operations to Amazon, you can start moving your market data infrastructure there.
And as you move more and more of your infrastructure to the cloud, the cloud will provide you with more and more capabilities there like those we showed.
And you are not the only firm benefiting from this, the whole industry benefits from the mutualization of market data.
Just to wrap up, let me just remind you what the scope of the Xignite Global Market Data Cloud is.
We curate and integrate data from more than 150 sources including exchanges, public sources, and data vendors—with whom we partner.
We have more than 50 API products available via unlimited subscriptions covering every asset class such as Equities, Funds and Fixed Income and every data type such as real-time quotes, fundamental data or valuation data.
We have the easiest, largest and most popular suite of Market Data APIs in the world.
And if you are ready to move to the cloud, our APIs are ready for you.
So now let’s go ahead and start the Q&A section.
Hello Everyone
My name is Stephane Dubois. I am the CEO and founder of Xignite.
I am excited to be talking about fintech and APIs today.
I’d like to thank API world for giving me the opportunity to share this with you.
The name of this session is APIs are powering Fintech Innovation. What is next?
So this is not a technical session, it’s more of a strategy and a business one, looking at a specific vertical: financial services.
As you know there is this fintech revolution going on.
And it seems that it will have a significant impact on the financial services industry.
What I’d like to do is first look at the role API played in this revolution….