Matthew Cheung: Hello, everybody. Today we're going to talk about data as a service. And according to one of our customers, it's about getting the right data to the right client at the right time and in the right application. But how easy is it to do this? And what is the right application? So we've seen the rise of big data and data lakes and data assets, but delivery of data into end user apps, or what we're calling the last mile of delivery is often neglected.

Matthew Cheung: Even though this is often the first touch point that your clients may have with your data, and it shouldn't be that way. Delivery into Excel, into chat, into APIs is more important than ever. And Excel, you know, is ubiquitous. The Microsoft CEO says it's their most important consumer product. While chat has seen a meteoric rise in the last year, Microsoft Teams usage is now 150 million users, which has doubled in the last 12 months.

Matthew Cheung: Symphony and Capital Markets users have grown to over half a million and grown by a third. And then looking at APIs, usage of APIs, recent surveys have shown that more than two thirds more APIs were used in 2020 than the year before. Now, financial market data is unique in the fact that it's often live data That's powering and driving real time decision making.

Matthew Cheung: And just to be clear today, we're not talking about market data. But we're talking about the delivery of time, critical, actionable data that your firm might be producing or aggregating data like streaming OTC prices that you might be quoting on bond axes, curve pricing, index or ETF data, or even reference data.

Matthew Cheung: So in this session, we're going to talk about what is data as a service. Why is there a gap between end user apps and some of the more sophisticated technologies out there? Why it's beneficial for data producers to focus on their core services rather than that last mile. And we'll also talk about new ways to deliver and monetize data.

Matthew Cheung: So I'm Matthew Chung. I'm CEO and co founder of iPushPort. We're a real time data sharing and workflow platform, and we've got customers across the financial market space who are using us for delivering data and also building data driven workflows using our no code platform. So I'm super excited today to have a first class panel join us.

Matthew Cheung: To discuss this topic, so just going around just on the order on the slide, guys, do you want to give a quick 20 seconds, 20 second introduction about yourself or your role is and what your company does? Do you want to start with James?

James Tromans: Thanks, Matt. Yeah, my name is James Tremans. It's a pleasure to be here.

James Tromans: I'm part of the Google Cloud team. I'm in the office of the CTO. So the Google Cloud does not have a single CTO, but there is a team of us that do that job. And we're geographically distributed and I'm based in London. Thanks, Matthew.

Ovie Koloko: Thanks as well, James. Good afternoon, everyone. My name is Ovie Koloko. I'm global head of product management and product development for Parameter Solutions. Parameter Solutions is the data and analytics as well as post trade risk management services division of the TPI cap group. And we're a leading provider of OTC market content as well as execution services to both at sell side.

Ovie Koloko: By side and non banking institutions. Thank you.

Bernie Thurston: Thanks. Oh, sorry. It's you, Luke. It's not me. I

Luke Ryan: said Luke Ryan, senior product manager for Morningstar market data within the market data business. We've been Morningstar. We collect, normalize and dispute exchange content. And OTC content from various providers on a direct from exchange to consumer point of view.

Bernie Thurston: Now I'll jump in. Thanks for that, Luke. I'm Bernie Thurston. I'm the CEO of Ultimus. We're indexing the ETF aggregator, and we're the largest ETF calculation agent on a global basis. And again, we distribute data to a number of the financial institutions, both buy and sell side. Matthew, I'll

Matthew Cheung: pass back to you.

Matthew Cheung: Cool, thanks everybody. So just a quick bit of housekeeping. So, we're using a platform called BigMarker. I actually chose that because of the Zoom fatigue that you may have seen. Great for any feedback when you're using it. Some things that's on this platform is a little bit more interactive. You can see on the right hand side, you've got the ability for a chat, so you can ask questions amongst yourself and to us.

Matthew Cheung: There's a Q and a part on the right hand side. If there's specific questions you want to ask us, you can put them in there and other people can actually upvote them. So the ones that are the most upvoted questions at the end will cover in the Q and a section and there's also some handouts on there.

Matthew Cheung: Some case studies and also a report as a white paper report where we've interviewed all of the panelists and put together. PDF that you can download that. So we're first of all going to kick off with our first poll. Steve, if you're there, can you just shoot that up? So what's the biggest challenge you face with delivering data to your clients?

Matthew Cheung: Is it delivery to user desktops? Finding developer resource, cost of ongoing maintenance, or audit and control.

Matthew Cheung: And then Steve, as they're coming in, do you want to kind of just run through the results?

Matthew Cheung: Sure. Okay. So we've so far got majority of people saying deliver to user desktops is the most popular, quickly followed by finding developer resource. Those are the top two.

Matthew Cheung: Great, fantastic. Right, so we're going to kick off first, going to talk to James Tromans who in his role in kind of office of the CTO in Google Cloud, which has quite a broad view of the ecosystem and infrastructure data and lots of things that are going on. So James, the cloud has made all technology more accessible.

Matthew Cheung: And we've seen things like big data and data lakes continue to grow yet accessibility or accessibility rather to data in end user applications has lagged behind somewhat. Why do you think that is?

James Tromans: Yeah, so I, so I definitely agree with the overall statement. I think there are some industries that are seeing a commensurate growth with end user data in front of the business users, but it's not a wide thing, and especially not so true in financial markets.

James Tromans: But I would say that, you know, data itself, just because we have the data lake or data warehouse still needs to be findable. It still needs to be usable with other data sets. So interoperable, it still needs to be accessible and it still needs to be. Usable in a way that, you know, once you've done it once, you don't reinvent the wheel every time.

James Tromans: And those are generally developer challenges. And that means that we're sort of constrained by, you know, how engineers that are already working with data are able to sort of. Find the data they need, do the things they want with it and get it in front of the business. So when I answered your poll just now, you know, from my perspective, I still think there is that contention on engineers.

James Tromans: So technology that can help automate or semi automate that process obviously is is desirable. And I think that right now, most companies are still in the position of seeing that final mile of data delivery into an application is very much their problem, their challenge, and one that they're solving with their own resources.

James Tromans: And so I think, you know, despite having this massive growth in data, which cloud can obviously help service from a fully managed sort of data warehouse solution, there is still the need to rely on. You know, either partners or incumbent in house engineers to make that data usable in a way that brings business value.

James Tromans: And so, you know, short of more companies recognizing that that's potentially a problem that can also be solved through software. I think that we're going to still see that gap. Between, you know, all of this promise of data actually finding its way into the hands of the business owners that can most or best make use of it.

Matthew Cheung: And on that point where your developers are still playing a key role in this. What do you see with the industry where, because definitely we've seen as a FinTech where business users are much more tech savvy than they were, which that's been a long term trend, I think, but more interestingly, is the technology side actually beginning to understand the business more and there's the convergence a lot more now than there was in the last few years, at least.

James Tromans: Yeah, plus 1 on that. I think, you know, even before I joined Google when I was at Citi, I was observing that trend and that trend was accelerating significantly. You know, when I was working on the trading floor, you know, with both technology and traders, you know, the more junior traders that were being hired were, you know, even.

James Tromans: You know, folks that would have historically never have really been interested in technology or even beginning to use Python, for example, right? And I know that for my work at Google, where I see across a number of everything from tier 1 banks through to much smaller companies that, you know, it's not massively you know, proliferating yet, but it's definitely happening.

James Tromans: And then similarly, similarly, the fact that technologists are getting a greater understanding of the business reasons as to why they're building things is bridging that gap. And the companies are getting the most out of their technologists are the ones that have that kind of most porous membrane between kind of the business and tech and actually start seeing the two as more partners.

James Tromans: And this links back to, you know, moving away from just the pure data. To lake data warehouse mindset to 1 to borrow a phrase, a sort of data matches becoming popular is a phrase that sort of best captures this idea of a business owner working with technologists right from the point of data creation, all the way through to data consumption.

James Tromans: So the idea of, you know, a financial services provider. You know, a product is an option or a swap, whatever it might be, you know, actually thinking that data itself can be a product for use internally to do more. And obviously cloud is in a position to make the burden on technology a lot less by taking away and managing the databases, scaling the databases, making the databases highly available, all of which are traditionally things that are kind of a cost to it.

James Tromans: And that frees up resources to work on these more important business problems. And so I have more time. To understand the business

Matthew Cheung: and you used to work on a on a trading desk back back in the day. Do you see a kind of a gap between some of those end user requirements and the people on the desk and some of the existing technology and knowledge that's out there?

Matthew Cheung: So. I remember kind of, you know, 10, 15 years ago when I was trading, everyone was an Excel whiz. And that person generally was anything technology related, went to that person on the desk. And just through that, that person just got better and better and everyone else probably remained the same. And now you've got on the other side where um, software is kind of eating financial markets, you know, kind of hear that term a lot.

Matthew Cheung: And in that sense, There's not many people actually can just hook together a couple of API as far from the court tab on the desk. Do you see that gap kind of changing as well? Like we've seen with. Tech techs and business people you're seeing the gap between kind of knowledge changing at all.

James Tromans: It's a great question I definitely still think that's a problem.

James Tromans: I think that You're absolutely right, you know certain individuals on a trading desk get relied upon to bridge those gaps But technology is also playing a role, to help make and build apis you know apigee would be an example of one of those that we have in google cloud But there are others too and I think that the the reality Is the idea of Excel has a perfect fit for certain use cases, and it should and will continue to be used.

James Tromans: But there is there are situations where, you know, you want to sort of bring things out of that environment to make them more shareable to make that work, you know, usable by someone else and that you're not just sending an Excel file around. And I think that's where the opportunity lies and where these people that can compose together different data sets in a way that's sort of auditable and reusable are going to really benefit from it.

James Tromans: Get the advantage in

Bernie Thurston: the future. So I think just to follow on from James's point though, in terms of obviously we're talking about the last mile delivery. My question is in terms of what is the actual destination I'm seeing there is a difference between what the, and to take James's point, the financial and technical experts that are now sat on the desks and I've got to say, I'm seeing the uptick in terms of the.

Bernie Thurston: Technical capability of the people sat on the desks. And as James said, people are taking using Python are they're taking our APIs directly? I would say in terms of there isn't just that one single person who used to do VBA. The speed with which people can now consume APIs is quite phenomenal, as it is no longer the barrier to the desks themselves.

Bernie Thurston: But I think what I am seeing is that there is a difference of use case associated with the. Should we say the corporate, the risk department, the back office side, they want to receive their data in a certain way, be able to do it, but then the desk won't be able to see it. And I think that's really where this last mile, which last mile are we talking about in terms of, are we trying to get to the front door?

Bernie Thurston: Are we trying to get the back door with this data?

Matthew Cheung: Thanks, Bernie. Let's just move on to another quick poll. I'm Steve. Do you want to just open that one? So are you using public cloud providers to deliver any live and stream streaming data? So yes, you're planning to, or no, but we're using some market data infrastructure that you might've been using for the last five, 10 years.

Matthew Cheung: Or no, we don't currently share data at all.

Matthew Cheung: So just seeing those responses coming in now, most people are saying yes, that they are using public cloud providers. Hold on, James.

Matthew Cheung: Yeah, that's kind of a 54 percent saying that they do use public cloud providers. It's a good news for you know, probably I push Paul and Bernie as well, because we rely on cloud to enable the services that we provide. So, so moving on next to of yay. Can you tell us a little bit how your clients are typically consuming data and then how data as a service providers can help provide that last mile of delivery?

Ovie Koloko: Sure. From a parameter perspective. General mantra is that we want to be able to meet clients wherever they are. So if a client wants to consume data utilizing public cloud capability, or they want to use some of the, I want to call it legacy, but some of the existing technology which has been used over the last, as you described it, 5 to 10 years Our role is to make that integration as seamless as possible and provide the mechanisms which allow them to make the choice.

Ovie Koloko: Ultimately, though, the how is often dictated by the why, and it's the use case for what the end customer wants to do, which is going to, which is often going to drive you know, the decision as to what technology they use to receive the kind of content that we're providing today. And if we think about the investment life cycle and we.

Ovie Koloko: Break it down into those three really basic chunks. So pre trade activity, point of trade activity, and post trade activity. We can provide, we provide data across the life cycle to different types of market participants, but to do all three of those things. So if we think from a pre trade, screening or research perspective, much of what James and also Bernie were describing in terms of the transformation that we've seen in the market and the ability for people to utilize new tools and get access to that data more quickly is happening. In this area historically, if we were providing data for research or for analysis, that would be about us extracting data from either hours or one of our channel partners, historical databases and then probably delivering that in.

Ovie Koloko: Batch format using, say, an SFTP account to the client to enable them to pick the data up and take it into their own environment and then conduct whatever research or analysis they want to do. What we're increasingly seeing the trend towards though, and both James and Bernie touched on it in the last question, is people needing and wanting, especially in for that type of use case to significantly reduce their time to value.

Ovie Koloko: And where where. Data as a service providers and people who are able to facilitate and enhance that last mile delivery have a providing an advantage to the end client is by creating environments which enable them to do that. So, if I, you know, if I quickly want to look at a new asset class and understand whether or not as a new market entrant, there is an opportunity for me to generate alpha or to reoptimize an existing.

Ovie Koloko: Execution model. I don't necessarily want to wait for someone to run an overnight batch job, deliver that data into somewhere before I make that call. If I can utilize certain tools to be able to do that at inception, using sandboxes or the like, I can massively reduce my TCO and significantly increase my time to market and increasingly we're engaging people who want to conduct that.

Ovie Koloko: That type of activity, and that's both on the cell and on the buy side, if we look on the other side of the spectrum, and this is to Bernie's backdoor point. In the post trade on the post trade side of things, if we start looking at middle office functions, whereas, as James did say, traditionally, people are utilizing enterprise risk and enterprise application solutions and then enterprise delivery mechanisms in order to provide market data, transactional data and reference data into those services, the.

Ovie Koloko: We're not seeing anywhere near the same type of interoperable gains, which I feel exist. And when you start thinking about some of the regulations, which are coming down coming down the path. So things such as FRTV, if you start thinking about the middle office functions at banks, where the boundary and to that porous membrane that James just spoke about, the boundary between what the front office traders are going to need to do and middle office and back office valuation and mismanagement functions require.

Ovie Koloko: Again, having that last mile delivery capability, which enables people to have those interoperable games and understand how different applications can speak to one another and then shorten and reduce the time in which decisions can be made, I feel offers a huge, huge potential operational gain to the types of clients that we speak to today.

Ovie Koloko: So. You know, going back to the question, we're able to deliver data to our customers in multiple ways, which we can see ourselves where there is opportunity for for further gains to be made. And we're starting now to see how people want to exploit that utilizing data as a service technologies in order to do so.

Matthew Cheung: So, so you're essentially improving the client experience by giving them exactly what they're after. Does that also mean that? Because there are some of the technologies that are now available that you can essentially buy over build, it means that you as a company can actually focus on your core services and and innovating products that you're kind of taking to market.

Ovie Koloko: To an extent and I think we, you know, we, we, we, we, we make those choices by over build versus partner often in all of the, whenever we're making ideation or, or we're discussing new products initiatives um, delivery data delivery wasn't necessarily an afterthought, but it was, it.

Ovie Koloko: Historically have been far more simple, right? So we could think about building a product and focus all of our attention on how we build that product. And then knowing and being quite comfortable in the fact that there was some standardized delivery mechanisms, which we could use for the customer.

Ovie Koloko: Once we completed that work, however, now was part of the ideation process for my team and the proposition owners and the product managers who are responsible for Susan. building the new data led or data first services and solutions that we deliver to our customers. We have to think about how different user personas and how different user experiences can be augmented by you by exploiting the types of delivery technology that there are there.

Ovie Koloko: So whilst it definitely allows us to focus on our core skills, the skill itself has changed, right? As a data product manager, you have to understand and know What technology is available to you and you have to know how you can integrate that into your product initiative in order to fully satisfy the needs of what your customers are going to be today.

Ovie Koloko: And most definitely in the future. Brilliant.

Matthew Cheung: Thank

Bernie Thurston: you. I

Luke Ryan: would echo that from a market data point of view. Morningstar is it is less about going back 5, 10 years where you have 1 delivery mechanism and everyone has to take that. It's now got to be about how do you Have a product and deliver data that meets your client needs and delivers it in the way your client wants it, not how you're telling your client

Matthew Cheung: to take it.

Matthew Cheung: Thanks, Luke. I think that segues really nicely into a question for you, Bernie. So your call service is obviously index and ETF benchmarking and pricing, and you got kind of always set up with this delivery via API. And then your customer demand said, Oh, can we have an Excel add in as well? And you didn't really want to build it, but your customers wanted it, so you had to.

Matthew Cheung: Thank you. But something like that would make sense for you to be looking to outsource. So what do you think the, kind of the, the limitations are of going down this kind of one API route versus being able to give it out your data out into lots of different formats. I think,

Bernie Thurston: I think as in, as we were just talking about in terms of, we're currently calling it data as a service.

Bernie Thurston: What I think is we're actually just a lot of us. Data providers. We're now moving into the service section, which is how do we take the format that we've got and deliver it in different mechanisms? I think that's really now the overhead that we're now beginning to see. This is how people want to consume it.

Bernie Thurston: We don't want to constrain it. We can fit into any of these. mechanisms. We no longer have the CSV flat file that comes in a defined format, but it's now the fact that we've got these flexible structures and everything else whereby we can keep on increasing the amount of data available. Great. We've now got away from the legacy delivery mechanisms where we were constrained.

Bernie Thurston: We had a version upgrades and the like. We've now got the opposite problem. We've actually potentially got too much flexibility, and this is really now where the service part comes from. We now need to be able to deliver into the database. We need to be able to deliver into the Excel spreadsheet. We need to be able to deliver into different chat APIs as well.

Bernie Thurston: And it's that whole idea in terms of these loosely decoupled systems. It. It's a challenge, but also an opportunity associated with it. And I think that's the part that is very exciting these days is the fact that if we can provide data in a format that allows people to do all these varied applications that they want to utilize, that's really where the opportunity comes from.

Bernie Thurston: And that's really what's going to differentiate the people who've embraced the cloud type of technologies. I'm going to define them as service to be able to. And I'm going to use my bad analogy again, being able to have the ability to service both the front door and the back door distinct services with from the same set of data and going back to the.

Bernie Thurston: point that was raised about the FRTB is the fact that being able to provide that same set of data and making sure that people have got the capability of being the traders utilizing it, the middle office utilizing it, and the back office utilizing the same data set, but for different purposes and at different points in time, that's really whereby we can now actually add value to the consumers of that

Matthew Cheung: data.

Matthew Cheung: And I suppose not only that, it's the, it's the permissioning, the control. the tracking, the audit of all that data. Now, now you can make it freely available through all these different applications. You need to be able to control and audit

Bernie Thurston: that too. No, indeed. And that's one of the major parts that clients are now coming back to us.

Bernie Thurston: And the pushback is like, we're supplying all this data down. How do we then control who it's going to in these various aspects? And as James alluded to earlier, in terms of obviously the clients used to think of. Well, anything they used to build themselves, they were then responsible for. But now suddenly we've got effectively a disparate set of applications, which again, they're trying to control throughout their organization and be able to license correctly on that side.

Bernie Thurston: And that's really whereby the, again, we're now creating separate APIs for permissioning control as well. And permissioning reporting specifically so that again, that whole. Permissioning layer is actually becoming a new data set in itself. And actually a very, I'm not going to say profitable one, but a good way of controlling costs as well.

Bernie Thurston: People have got more visibility on that side and we're creating more and more capability for people to be able to view that data report on it and see who's been using the data, both from application perspective and from an Excel slide so people can decide whether or not they need that data going forward.

Bernie Thurston: Great. Thanks,

Matthew Cheung: Polly. Luke, over to you. The challenge, actually, that you're facing, or one of your challenges is around probably the other side. It would be the first mile, where it's getting data from different sources, ingesting it into your platforms and services. Can you kind of talk about Some of the, some of the challenges that you have around managing that data and, and, and essentially building servicing around it and transmitting it on elsewhere.

Matthew Cheung: Yeah. So as a,

Luke Ryan: As a data vendor and aggregator, not only do we have to try and get our data customers, we're also collecting a huge amount of data as well. And to do the aggregation. So at the moment, we're collecting data from over 260 different sources. We've got 25 million plus instruments on the, on the data feed.

Luke Ryan: So we have both the outgoing delivery issues, but we also have the issues of managing vast numbers of inbound applications, um, managing those ensuring that the data is normalized the same way that it's loaded the same way and transmitted across the data feed in the same. The same way so our customers can make use of it.

Luke Ryan: So as one of your polls went out earlier, our biggest issue is finding development resource maintaining the code afterwards. That all comes with a huge overhead and some of the core data sets. You can't get away from that from, but there are lots and lots of up and coming data sets that we are onboarding all the time onto the market data feed from all staff.

Luke Ryan: And this is where companies like yourself come in to help us normalize these datasets, which combined have a huge value to our customers. But individually don't have enough value for us to do find the development resource, maintain ongoing. So how can we outsource some of that work to companies like I fish ball to ensure that we can ensure that we can continue to give the value to our customers.

Luke Ryan: sO. As I said, a lot of this is down to how do we find the right partners as yourselves and, and normalize that data to do that ETL layer before it's actually loaded into our. Into our databases and now our data

Matthew Cheung: feeds and in the long run, doing in the long run, doing kind of ad hoc development is not very efficient or scalable.

Matthew Cheung: How do you. Decide what I'm going to take on on our top development versus looking at third party providers for services. So

Luke Ryan: you obviously got a very core set of data points and data providers that you need to go to, which is going to be where your core resource are going to be allocated to, but it's the value add information that you can collect from other sources that really becomes.

Luke Ryan: The key driver for outsourcing work. So where do you, where could you add value with other data sets that enable you to grow from a marketer point of view, an aggregator point of view to grow your business and your coverage to ensure that you can continue to provide value to your end users. aNd then that's where outsourcing the information, outsourcing that ET of the ETLF to, to third parties makes the most amount of sense because you're essentially creating a almost like an aggregation within the aggregation before it reaches to you to ensure that you're not overburdening your, your development resource to do too much,

Ovie Koloko: oNboarding

Luke Ryan: and therefore too much maintenance going forward as well.

Luke Ryan: And using tools like yourselves enables you to push some of that. That's analysis and loading away from the engineering teams and more to technical operational teams, which then lines are low, which then enables you to do more work with your core data sets to add more money later

Ovie Koloko: on.

Matthew Cheung: And how important is it to move around kind of that development lifecycle and the efficiency of building and maintaining these end user apps yourself?

Matthew Cheung: Yeah, we went, I think when we spoken, particularly Bernie, like you're. Your core offering is the aggregation of data, and it's not delivering it out into 20 different end user applications. How important is that speed to market for you on one hand, and also The stress of not having to deliver data in this, in this kind of last mile delivery.

Bernie Thurston: No, I'm going to say in terms of, and to speak to Luke's point as well, in terms of also the maintenance of all the feed handlers from us in terms of the number of providers that effectively are still providing data in flat files. We still have to, they change their FTP location. Suddenly they've.

Bernie Thurston: changed. If hopefully they've added two more commas in, we spend an awfully long time handling that side. But as you just alluded to as well, in terms of there's so many different use cases associated with this data, then especially with the expansion of ETFs in the global entity, obviously that we've just had a massive uptick in terms of the number of people who are interested in crypto ETFs.

Bernie Thurston: Again, we're having to adapt our interfaces and everything else on such a regular basis as people find new use cases associated with this. We've been asked to add in and flagging our feeds, various ESG products as well. And these are all new data points that people want to have visibility of. It's very easy for us to add these data points in it's then, and this is where.

Bernie Thurston: Again, Matthew, your organization comes into its own is the fact that giving them people visibility associated with that data is actually the most troublesome part from our side. We can easily build it into the feeds, but it's then the fact that we then have to go through release cycles with our clients testing and everything else.

Bernie Thurston: And that's really where the whole process slows down. Obviously, people have got. Requirements where they need to fulfill them in a timely manner, and then we run into the, I'm not going to say bureaucracy, but process associated with our end user clients, and we have to go through a step, and that's really what delays the whole receipt of the data, and which is why we tend to fall back to Excel again at that point in time, where in terms of how do people receive it, suddenly we end up almost One moment here.

Bernie Thurston: Going back and going, Oh, here's a CSV file with all the data pre formatted so they can consume it. So we almost sometimes slowly lose the capability that we've gained over the last couple of years just to be able to get past all the process associated with a large organization.

Matthew Cheung: Have you got any comment on that, Javier? You're on mute.

Ovie Koloko: I promised myself I wouldn't be the person who was who was on mute. I, I think from, from, from my side of things and starting thinking specifically around The burden in and around data, you know, the last mile delivery. I'm quite fortunate that I'm, I'm, I run products and I have a CIO who who looks after the technology side of things.

Ovie Koloko: So I can look to the left when, when needs be and be rest assured that that's something that's keeping him up at night. BuT I'm also very fortunate in that I work with a CIO who understands that the delivery landscape is massively evolving. And he's been a massive, a huge, both internal and external advocate of assessing and looking at the new technologies, which are coming to market and understanding exactly how we as an organization can leverage those for our own, for our own benefits for.

Ovie Koloko: Myself for my team and for our proposition owners, what we have to obviously do then is think about how we can leverage those and to the comments I was making earlier in and around on starting to understand and look at how a particular data set or a particular solution that we're building can be delivered using these tools and can accelerate the client experience is something that we are constantly having to worry about.

Ovie Koloko: And we can't wholesale just allow our colleagues in technology to pick that up. You know, depending on the asset class, there's a huge demand for speed in the areas in which we work, we work in. So if you look at things such as the LIBOR or IBOR transition at the moment, our ability to quickly ensure that we have, the right data for the right curves, which can enable and support customers when they're doing their transition and using the various applications that they have to do so that they need, obviously, to do so is key for us and understanding from that last mile perspective again, how we can leverage some of the technology that exists today in order to facilitate and speed that process up has been something that's been very interesting, but also very important.

Matthew Cheung: So one thing that we've noticed with our customers is it's actually for one reference data customer. They've started to looking at different new commercial models that's actually based on data usage, because, you know, your market data licensing is always a very contentious issue for everyone in financial markets, because it's very expensive.

Matthew Cheung: But when you're getting into the realms of OTC data or. Or data which you get only you have and you want to be sharing out the ability then to to distribute it in a audited and controlled manner actually gives you the ability to start tracking usage. Is that something that any of you all the way from James Google through to the data service providers here as well?

Matthew Cheung: Have you started to look at more kind of like a pay as you go commercial model to pricing of data?

James Tromans: I'll, I'll, I'll jump in real quick. I think you know, obviously in our position, we want to facilitate our customers to share. And and safely monetize their data as they would like. And so we have the tools on top of GCP, you know, as part of the platform that can be composed together very, very seamlessly that would for sure enable that type of model.

James Tromans: Should that be. The way a customer wanted to do so, and we're sort of going to various lengths to work at, you know, what our customers want us to build a generically for them, so they have to compose even less together. Right. And it's sort of more, you know, how far how far should we go? So we think quite a lot about that.

James Tromans: And, you know, at the same time, we are not. You know, in this case, you know, a market data company, it's not what we are. So I think it's, it's, you know, we're trying to find a place where we're really serving our customers to what they want and what they need from a cloud provider. So, so yeah, absolutely.

James Tromans: It's on our minds. And we're working with customers and currently to think about these types of challenges.

Bernie Thurston: And just to follow up on that and sorry, James, I'm here going to blame you as well in terms of I can see people wanting it. I think one of the big issues is obviously until everything's going through a single platform and people can control it in one place.

Bernie Thurston: We've got so many different reporting avenues and my big issue with cloud providers at the moment is They seem to keep releasing a new product, a new way of delivery, a new way of consuming this data on a weekly basis. Trust me, James, I'm not actually complaining, but it is that whole idea in terms of trying to cope with that way of delivery, consuming, processing all of that data, plus a usage base pricing model, I don't think any of our.

Bernie Thurston: Clients are ready for that type of approach as of yet. I'm sure we will get there, but it's just the fact that one, we're increasing the amount of data. We're increasing the amount of ways it can be consumed. I can't see any client currently willingly signing up to a um, if every consumption based model, because there'd be scared in terms of how much there'd be hit with at the end of the day.

Bernie Thurston: And trust me, and again, James, I apologize. I think this is the one side with. Cloud based providers. It's very easy to accidentally consume a lot of resources associated with any of these things, especially when consuming data. And it's having the controls in place to be able to ensure that, okay, what is the usage?

Bernie Thurston: Once we've managed to confirm the usage rights and credit to the cloud based providers, they have got an awful lot better and they have got better than on premise solutions to do this. But Until that's then the de facto standard, we can't go down the line of effectively consumption based pricing.

Ovie Koloko: Just that adding to that, I think from, from what I can see where there is definitely scope to, to further evolve is supporting, it's supporting flexibility.

Ovie Koloko: So I, I, I don't necessarily, I agree a lot with what Bernie's saying there in terms of the, the The need or even really the desire on the end client side to have a consumption based or pay as you go type model. I think people recognize and understand that they need different data sets for different things.

Ovie Koloko: And the use case ultimately will determine what what that means. However, once you've, once you've determined what the use cases, having that flexibility to look at other data sets or, or, or in other areas, I think it's something that, clients are expecting us to be able to facilitate because the technology is enabling them to do that significantly more seamlessly than they have done historically.

Ovie Koloko: So, thinking about how we evolve and adapt some of our licensing policies in order to encourage that, that type of participation, I think, will be very, very interesting over the next, over the next few years. And just to add to that final point, I.

Luke Ryan: I agree with what Bernie said about ongoing data. Pay as you go is very difficult to manage ongoing, so you don't know how much data is going to be published at the end of the day, and it does scare customers, that sort of model.

Luke Ryan: Morningstar have started doing the pay as you go model for historical data, which can be managed. You know how much data was produced on a particular day. And therefore, customers can manage that more successfully, but I think ongoing, we're still some way away from being able to offer that as a solution on a pay as you go model, because it's just so much of an unknown about what's going to happen in the markets from a day to day point of view, and being able to assess how much you're going to pay each day is kind of a black box at the moment you can go from.

Luke Ryan: 20 billion messages a day to 35 billion messages a day because Elon tweet. So it can be vastly different from

Bernie Thurston: day to day.

Matthew Cheung: Just for everyone listening as well, for the, for the audience, you know, if you do want to ask any questions, please post them in the side and we can, we can cover those in about 5 10 minutes. I've got a question for everyone on here around Excel. You know, Excel is used by... 750 million people around the world is ubiquitous in the markets and data providers are often kind of forced to a certain extent to provide Excel add ins to meet this demand.

Matthew Cheung: However, they are difficult to maintain, and often you might need outsourced resource to build an Excel add in. What do you think the future is? for Excel and what's the world beyond spreadsheets and how can kind of data data as a service providers bridge that gap.

Bernie Thurston: I'll leap in here in terms of Excel is here to stay. It's the easiest way for us to be able to consume data. I think all clients, all organizations hate it with a passion because effectively it's so difficult to maintain. There's so much flexibility associated with it. We are beginning to see more and more.

Bernie Thurston: Capability for people to be able to generate their own data sets out of APIs and everything else, but fundamentally it still gets downloaded into Excel so people can go and manipulate it themselves. They can go and make any changes. They can go and attach it to a Bloomberg terminal. That's effectively where I think everything.

Bernie Thurston: Falls down at the end of the day in terms of traders want to be able to manipulate the data they've got and make their own assumptions associated with it. So I think there obviously have been things that have been coming out. The Office 365 on the online portal gives you the capability of doing it.

Bernie Thurston: Google Sheets allows you to import APIs natively. So they're all obviously. Waste get through it. But again, the problem is obviously with our financial institutions and everything else. They likes to be controls processes around this type of thing, which then yes, we're moving the capabilities. We're beginning to see progression.

Bernie Thurston: But again, the financial institutions are probably 345 years behind in terms of deciding when to. Take these things in and unfortunately, the traders still want to be able to consume the data and play about with it. So we're stuck with Excel for foreseeable future.

Ovie Koloko: I think just deciding to burn these burning made a comment around flexibility for for an individual quickly looking to manipulate data.

Ovie Koloko: Who is semi. Literate or numerate, it still remains the easiest tool to do so, and the enhanced integrations that it has with third party apps make it the best place to do so as well. Control is definitely an issue, but I feel as some of the last mile providers, enhance some of the permissioning and auditability capabilities that they have.

Ovie Koloko: I feel for enterprise data providers such as ourselves, Luke or Bernie on their side, we've become far more comfortable and willing to accept the fact that the data is being consumed in that way. And we have, as IP owners and content owners, some way of being able to control how the data is used and then, and then, and then.

Ovie Koloko: Potentially that manipulated and distributed downstream. But, you know, I agree. I think it's a tool that helps facilitate and bridge lots and lots of gaps on DSO long as the data is provided in a controlled and auditable way, it's definitely something that we want to support the ability for people to be able to access and use our data.

James Tromans: I'll, I'll offer maybe the engineer's view or an engineering manager's view, perhaps. yoU know, you write code in Excel, you write, you write cell logic in Excel. It's not version controlled. It's not shareable. None of your other colleagues can reuse that code easily. Any data that you produce in Excel that's derived data from the data that was already there.

James Tromans: No one else can access that data. It's, it's your data. It's in there, right? You have to save your Excel file and send it over an email or put it on a SharePoint or put it in Google Drive or whatever, right? And then it's one of the few tools that in order to view the data, you need the authoring tool to view it.

James Tromans: Which again is kind of strange, like, you know, HTML5 web dashboards don't work that way, right? You can just use the browser. You don't need to have the end, you don't need to have Visual Studio code or something else, you know, just a viewer. So I think there's a lot of downsides to Excel from a software engineering standpoint.

James Tromans: And I think if you think about versioning your code so that you actually can fix bugs and other people can see what you've done, if you think about having access to data that's, you know, reusable or going back to those fine principles, it's a With that said... It's been around a long time. It solves a very good set of use cases where it's absolutely the right tool and the problem is it gets abused by people because they're comfortable with it because it's what they were brought up on and they use it in a way that it's sometimes not designed for and you end up having quite big, monolithic applications that do a lot in Excel and they should be refactored out and written as proper applications and have proper tools.

James Tromans: Bye. Bye. Software life cycle development controls around them, and I'm not saying you can't bring some of those software development controls to Excel. You can, but it's like shoehorning something in that's not natural. So I think, you know, to borrow a phrase that we, we, I've used in other talks and a colleague of mine you know, what's beyond spreadsheets.

James Tromans: I think that we're seeing, you know, an interesting trend to do certain analytics in, in Jupyter notebooks. And we're seeing, you know, from the, even the literate traders on the trading floor, for example. They'll quite happily boot up a Jupyter notebook and have a play. Now, that's not everyone, and I don't think Jupyter is going to fix, solve the problem for everyone either, right?

James Tromans: But I'm just seeing these little glimmers of where other tools are sometimes being used for problems that once, once would have only been considered to be Excel.

Matthew Cheung: So let's, let's have a look at, chat, you know, the last 12 months, we've seen chat used more than, than all, you know, all of history before it, where, what do you as a group kind of see the role of chat and chatbots playing in, in, in data kind of consumption, and I'm talking about.

Matthew Cheung: You know, there's Bloomberg, IB Chat, there's Symfony, there's Teams, there's Slack, there's Icon Messenger. There's a lot of different platforms that are used widely in capital markets. But where do you see chat evolving and how are you as firms looking to utilize that?

Bernie Thurston: I will take that one. In terms of the chat from our point of view has been absolutely phenomenal in terms of obviously not requiring the Bloomberg Instant Messenger. We use Symphony quite extensively for doing a lot of our support. The big issue that we've currently got is just in terms of, we unfortunately then fall back to email as soon as we have to then send data or the like, because obviously just the fact that we can't.

Bernie Thurston: Post CS fees or things within some of these chats, it's again, the restrictions of the large institutions, which then causes us the problems. So in terms of a lot of the support, a lot of the conversations are all done through chat, and then we have to revert back to email. So then it's a bit going backwards in some of these cases, especially if you're in a flow of a conversation, you want to be able to send data.

Bernie Thurston: That's the part that we would now love to be able to address. And obviously we're having the conversations. Currently about in terms of how can I then deploy that are some of our data sets into the chat associated with it. And that's the part that we're trying to address. And yes, if we managed to do that, we can see that as a huge leap forward in terms of the support of the clients, being able to ensure that the seeing what they should be seeing at the end of the day.

Bernie Thurston: And that's really where we want to be able to get to at the moment. There's a bit of a disconnect, but the chat is. Expand expanding that out quite considerably

Ovie Koloko: for us. I think where I see it going is around shortening the time to make decisions. If you have if you have bots, which make currently used applications and some of the various chat facilities speak to one another in a more seamless way it enables teams to collaborate in a far more and therefore speeds up the the decision making process.

Ovie Koloko: So from a pre trade from a and from a post trade perspective, I feel it can significantly again, as I described it at the beginning reduce the time to value. I feel that there's lots and lots of good and well known applications of chats and execution Chat rooms and execution platforms interacting with one another's.

Ovie Koloko: And Matthew, obviously you, I push, I push will help facilitate lots of of that interoperable interoperability today. But I feel that there are other areas within organizations that could also benefit from this. And as we see those applications and chat rooms integrate more and more closely with one another, you'll see decisions speed up and that, and people gain more value and benefit from it.

Matthew Cheung: So on that note, let's do another, or in our last poll so for everyone listening, are you interested in delivering data to your clients in any of the following things that we've been talking about? So Excel, chat, API, or anything else? Well, I think it's actually multiple choice, so if you want to do everything, you can tick those two.

Matthew Cheung: So just looking at those results and seeing chat and API, neck and neck.

Bernie Thurston: Definitely think we should have had all of the above in that poll.

Ovie Koloko: I think, I think that should have been,

Matthew Cheung: Great. So just taking Steve, if you're on, are there any questions from the audience that we should answer? Yeah, there's a couple of questions.

Matthew Cheung: Toby's posed a question. Microsoft are looking to better integrate Python into Excel. I don't think it's a case of Excel or Python scripting. They both have a place and can work well together, horses for courses. So it's probably just a more general comment. And I guess,

Ovie Koloko: do people agree or disagree?

James Tromans: Yeah, I I agree. Like I said, you know, Excel is a wonderful tool. And I will say, like, we can step out of just Excel, right? Like, Sheets in general is a wonderful tool to, to solving certain problems, but it's when it gets abused. And actually, you know, there are tools out there like Pixel, PYXLL. And others that are actually making it so that you can do stuff in Jupyter and it shows up in Excel and vice versa, like, you know, actual using the two tools together.

James Tromans: So you can go and if you're interested in that and go look it up. So I think the future, it was so early with with this. I think there's a lot yet still to be discovered, but there's a space for both for sure.

Bernie Thurston: I think it does come back down to the whole control mechanism, people and being able to pull it all into Excel.

Bernie Thurston: Excel is unfortunately the worst database known to man, and unfortunately everyone thinks they can utilize it. And that's the biggest issue I have with it in terms of controlling it and people saying, Oh, but I've done this and I don't get any data back. That's my major issue associated with that side.

Matthew Cheung: Any other questions, Steve? Cool. Yeah, Toby's just come back and just said that you don't want to be processing vast amounts of data in Excel. I think probably would

Ovie Koloko: agree with that.

Matthew Cheung: Another question from Neil. How can fintechs help with increasing regulation around benchmark vision collection and distribution?

Ovie Koloko: That's probably a whole nother webinar, isn't it?

Bernie Thurston: I was going to say, I'm, I've got to say, unfortunately, I'm loving it right now in terms of, obviously for the benchmark providers, our biggest issue is the lack of identifiers. So obviously ESMA with the BMR and everything else on that side, unfortunately, they assume that every index, every ETF has got an ISIN.

Bernie Thurston: Unfortunately, about 30 seconds later, they realized that wasn't the case. So we're spending a lot of time effectively going around all of those providers, web scraping and everything else. Unfortunately, that still does come back. Down to a lot of manual side, but there's also if every things that we're doing on the 871 M, the narrow board regulations were effectively fortunately due to the fact that the, I think this comes back down to one of the earlier points.

Bernie Thurston: We can do all the calculations in the cloud, get all the data out there, but then unfortunately we then run into that stumbling block in terms of the client's internal applications. Great. We've got all these flags about ESG, 871 M, BMR, everything else. Then the client's applications then have to be able to figure out how to consume this data, and that's the major stumbling block currently in terms of, can we, we can build all the data.

Bernie Thurston: Unfortunately, can the clients consume it for the regulations change?

Matthew Cheung: So, I suppose to wrap up, you know, in the, in the data driven financial markets of today. We've seen that availability of and access to data on its own is only part of the challenge. The last mile of delivery. So the integration of data into the workflows of the clients who consume it is often the last thing on the minds of data producers.

Matthew Cheung: But talking to some people today, you can see that is now becoming a little bit more. Important and gets that way day by day, but it is the first in the mind of data consumers. And from a client's perspective, it can be the key differentiator, the feature that drives them to choose one data source over another.

Matthew Cheung: So it's no longer sufficient just to provide data to your clients on another screen on their desktop or a custom fix API that they need to develop a resource to connect to. Clients want data to be delivered to them in real time in the format and the mechanism. By which they want to consume it, and we're seeing a new breed of fintechs a service providers like high push pull.

Matthew Cheung: Yeah, we're at the forefront of this new way of delivering and integrating your data seamlessly into your clients existing workflow tools into applications. They want and also. Providing that interoperability and extensibility for any future applications I might want to connect into. And importantly, for a data producer, you need to be thinking about the barriers to client onboarding and integration because the lower they are, the quicker the ideas and data as a product can be brought to market.

Matthew Cheung: So essentially we're, we're enabling and data service providers are enabling data producers and aggregators to simplify. And automate the data driven workflows and that frees up some of that last mile delivery. So data service providers can focus on core value propositions. We've also heard today about those last mile applications of Excel.

Matthew Cheung: Chat and API. There's obviously lots more, but us as a business is the common ones we've seen, which is why we've been talking about it today. So, James, Olivier, Luke, Bernie, thank you very much for your time today, and thank you everyone in the audience for listening.

Ovie Koloko: Thank you. Thank you. Thank you. Thank you very much.