Host: afternoon, everybody. My name is Clive Posselt and a very warm welcome, both to our audience and to our panellists for our discussion this afternoon. The plan is 35 minutes or thereabouts of discussion and debate, followed by Q& A. If I could ask the audience to please use the Q& A functionality in the webinar GUI to send in your questions and I can then put them forward to the panel.

Host: We'll also be having a number of polls during the session. Hopefully that will give us a better idea of the, the backgrounds of the audience members and also their relative understanding or experience with the topic so we can. You know address those our discussion topics and points accordingly.

Host: Hopefully as you know this afternoon, we're going to be discussing data as a service We have a fabulous panel with us this afternoon all of who contributed to the Recently published financial markets insights report entitled remote working very, very salient at the moment reveals critical importance of data as a service.

Host: So without further ado, I'd love to introduce the panel. And first off, can I introduce Julian Dugar. Julian gives a wave. So we know everybody knows who you are.

Matthew Cheung: Thank you very much. Julian is

Host: fixed income client execution platforms and digital sales with Matt West market. Next up, we have Patrick Flannery.

Host: Can you say hi, Patrick? Thank you very much indeed. Patrick's co founder and chief executive of Maestrate. We then have Mark Wolferdon. Thanks for giving us a wave, Mark. How you doing? Mark is managing director at Euromoney Trade Data. We've got John McPherson. Welcome, John. John is the deputy chair of the advisory panel at Enjin.

Host: The Investment Association's Accelerator. And last but not least, we have Matthew Chung. Hi Matt, how's

Matthew Cheung: it going? Hello.

Host: And Matt's Chief Executive Officer of iPushPool. So welcome all, and to kick off

Mark Woolfenden: what I'd like to do is start up with

Matthew Cheung: our first

Host: poll of the day and hopefully this will give us a better idea of hopefully whether the audience have a decent idea of the topic.

Host: So first question going up now and it's have you heard of data as a service? Yes, no or not sure. Answers coming in rapidly.

Host: Need some music here, a bit like who wants to be a millionaire probably. But yeah, we're getting close now. So there we go. 85 percent of the audience have heard of Data for Service. So that is pretty good news for us and the panelists. Mo was 12 percent and 3 percent weren't sure. So I'll just end that poll for now.

Host: And so given people have heard of what this is, I'd love Matt, if you could kick us off just from your experience and your view, what is data as a service and what are the salient points we should be should be

Matthew Cheung: thinking about. Yeah, sure. Thanks, Clive. Thanks everybody for joining us today. So data as a service from our perspective, you know, we see it is the ability to seamlessly connect your data to anyone else who wants it in whatever application they want.

Matthew Cheung: So you can connect your data, which could be sitting in a database and a platform, even in a spreadsheet. You can then share that with your clients, your customers or their colleagues into applications they're already using, like an Excel spreadsheet. You know, in Excel spreadsheet, a chat platform, a chat box an internal platform via an API as well.

Matthew Cheung: And that data can be live. It can be streaming or it can be on demand. And with that, cause we're working in capital markets, we need all of the enterprise security control and audit on it. Historically data sharing in capital markets has been pretty complex. It's been very manual. You see a lot of emails, file sharing, copy, paste and so on, and that's where you then have to streamline and automate an audit, or you use expensive developers to connect all this data together, and none of it being out of the box.

Matthew Cheung: So why is it pertinent now? Well, the financial markets obviously have seen a big change in working conditions, you know, people working from home because of COVID and so on. And that's really accelerated cloud adoption and digital transformation projects across the markets. There's really an interesting piece in the FT this week.

Matthew Cheung: I don't know if anyone read it about how big banks are now adopting the cloud. Yeah. We saw Microsoft actually in the last quarter saying that their CEO said they've seen two years worth of digital transformation in the last. Microsoft are actually reporting tonight. So keep an eye out after the bell to see what their comments are this time around.

Matthew Cheung: That would be really interesting.

Host: Yeah, definitely. I think two years is probably an understatement for some companies. Yeah, we'll

Matthew Cheung: see how those metrics get spread out when they comment on it later. But what's happening now. You know, it was inconceivable at the start of the year, you know, the people moving to the cloud adopting all these technologies.

Matthew Cheung: So what does this mean for data as a service? Well, it means embracing data as a service means you're no longer confined to internal workflows. You can utilize the cloud. So then you use the cloud as an enabling technology. So then share data to external clients and customers and teams to trigger workflows and so on.

Matthew Cheung: And using data as a service means you can do it in any application and it's all plug and play again using the cloud means it's incredibly scalable. It's significantly cheaper than you trying to build it yourself and there's a fast time to market. So I think with the with the panel we've got here today, they're going to really talk about, Some of the scaling and how you can use that with the cloud decisions, how kind of buying things in instead of building it and how those decisions were made.

Matthew Cheung: And it really why data as a service needs to be part of a digital innovation strategy. If that's what your organization is now looking at. Brilliant. Thanks, Matt.

Host: That's a really good overview, and hopefully that sets the scene well to the topic we're going to be discussing this afternoon. So moving on, I've got a couple more polling questions before we get out to the rest of the panelists.

Host: So first off, I'll launch this one. And what has driven your firm's digital transformation, given what Matt was saying about the change two years and two months? Has it been the CEO, the CTO, the CIO, COVID 19

Matthew Cheung: or other?

Matthew Cheung: My money's on COVID.

Host: Interesting, there's three a neck and neck at the moment.

Host: And yeah, COVID's one of them.

Host: Any more able to just click their polling numbers? I think we're going to call it there. So, top of the pops actual fact COVID and other. CEO CIO 11%, and in the 30s each, COVID and other. So, as you said, you know, COVID obviously has had a had a strong effect on people's digital transformation.

Host: So, the next question I was going to quickly ask, and this will help the panel to help sort of direct some of their responses, I think. And I'm just going to launch this now. So, is your firm buy side, sell side, technology vendor, or other, please?

Host: Funny enough, this question is actually being answered much more rapidly than the previous one.

Host: And if we just end the end the poll there. So we have 39 percent technology vendors, 17 percent sell side, 15 percent buy side, and about 35 percent other. So that hopefully gives the panel a bit more of an idea as to the sort of audience we're We're talking to you this afternoon. So Matt, thank you very much.

Host: And if you're fantastic overview, what I'd like to do is move on here and bring gin John in here, John in here, both from your experience of BML, but also working with your by side members at the Investment Association. Can you give us an idea where data service is being used

Mark Woolfenden: within the

Host: markets today?

Host: And some of the advantages that it's offering to people. Yes, Clive. Thank

John Macpherson: you. So I'll just put it a little bit of context. My background actually goes before the investment association engine program, which I'm now involved with and have been for the last two years. Before that I was CEO of BML Technologies, which is a fintech startup that has been scaling up in the UK and globally.

John Macpherson: And before that I had a career on the sell side in the investment banks primarily with golden sacks and jp morgan so i've been through the evolution if you like of Firms very much building first and buying last up until my departure from Goldman Sachs back in 2013. It was unthinkable that they would buy in technology.

John Macpherson: They would build everything themselves, you know, bespokely. Having caught up with some of the senior members of those teams in the last year or two now that's very much reversed. Now they're very much a a buy first and build last company, which I think is prevalent amongst all the major financial institutions that I've come across So that that that that ship has sailed if you'd like

Host: as it relates and that was very similar to yours there.

Host: John. Definitely. Definitely been a transformation

John Macpherson: Yeah, no, and it's very noticeable, you know people want to work with fintech firms once you have a problem to solve and you've identified it you're trying to identify a solution Now with data as a service it's interesting. My experience from the last couple of years has primarily been through the buy side of the business as opposed to the sell side historically and getting used to and understanding the type of problems that those institutions have.

John Macpherson: Data as a service for them they look at very much as a to be a cost efficient responsive service lets them focus on selling their products rather than sourcing, managing and activating the data. I think specifically amongst the asset management community, they, the primary benefit they look for is around the automation and streamlining of pre trade negotiation workflows, which is where I push pull very much come into the picture.

John Macpherson: But it can help them across lots of different things if they'll allow it. That can range from relationship management through the market data piece risk management, post trade processing uh, trading surveillance, even research. So there's, there's a whole bundle of things. But one primary entry

Host: point, I was going to say, John, you say there, if they allow it, do you feel there's an appetite out with the trading desk to embrace this sort of technology yet?

John Macpherson: Very much so. I mean, specifically again from the buy side, which is where I'm coming from today. The firms that have that appetite are very keen to adopt and move forward quite quickly. But we're seeing an interest from All the firms in learning more before taking maybe that next step.

Host: No, that's, that's very interesting.

Host: And can you sort of, sort of give us an idea of maybe some of the advantages, specific advantages they're saying, other than just at the trading desk, you gave examples, but there are any other sort of clear examples in? Sort of post trade or reference data that are particularly jumping out to your your members at the moment.

John Macpherson: Regulatory risk management and basically challenging firms to capture, store and analyze the data over many years, many departments and many regions at far greater granularity than ever before. And I think that's probably the one that I've seen the most interest in recently, but there are multiple examples.

Host: That's brilliant. Well, thank you very much indeed for that. That's a good good bit of background on the buy side Patrick, can I, can I come to you next? You know, obviously there's an appetite for this both on the buy side, sell side and also for data vendors. But as a technologist, you know, from your perspective, what are the key challenges that data rich firms really experience

Mark Woolfenden: when looking into offering data as

Host: a service to their customers and how can they overcome these?

Patrick Flannery: We think, you know, we break the problem down into five different areas. So collect, store, transform, maintain, and deliver. So, for whatever sort of data that you're working with, if you think about, if you think about digital transformation, a lot of, a lot of automating processes is, is getting access to the data and then providing computer code software with a place to run it.

Patrick Flannery: Right. So getting the actual inputs to the to automate the business process is really a foundational challenge, you know, actually actually getting the data. So when you think about collect the data, you know, in our thinking, there's both a business challenge, which can involve licensing rights, it can involve physical access security, but also a technology problem.

Patrick Flannery: How do you actually Hoover up All of the different bites. How do you ingest them? So once you've collected the data you have to store the data and increasingly the more data that's produced by markets But also internal systems that are automated There's a reflexivity of sorts the more data you produce the more data you have to work with the more things that become automated And the more data that's there.

Patrick Flannery: Well for us as a company, we have about 12 petabytes of data It's growing quite quickly linearly across all all sorts of asset classes. If you were to take that you go to the EWS calculator, you can google

Host: Yeah, it's going to come out pretty expensive.

Patrick Flannery: It's definitely hundreds of thousands of dollars per month per region.

Patrick Flannery: Now, most firms have a hard time storing all the data because they're not really quite sure what to do with it or they don't have the tools. So this whole process of once you have the data, actually having it in a place and oftentimes, you know, finance firms are global. So having different copies of the data in different geographies that may have different governance is certainly an issue so that those storage costs can multiply and they're only getting larger.

Patrick Flannery: It's only, you know, you're only storing more data. So once you, once you have all this data, it's, it's a, it's a pretty unformed data lake or data ocean. And so you have the process of maintaining it. Organizing it and accessing it. Right? So oftentimes there'll be an ocean of data and you're really trying to pull out the relevant portion of it.

Patrick Flannery: That's a big challenge, right? So, um, having that having that large collection of data and Once you've extracted the relevant portion of it, you need to put it in, deliver it to some set of workflows, which oftentimes requires downstream workflow integration. Maybe that's databases, maybe that's tools like iPushPull are excellent for that because they become a conduit to integrating with a whole host of downstream applications, right?

Patrick Flannery: And I think that that, downstream integration. You know, you're saying data is a service, but it's kind of integration as a service in a lot of ways. You know, you're, you're, you're, you're, you're making data useful. So those are questions, you know, what do we see the challenges? There's a lot of challenges and some of them are technical, but many of them you know, are really frankly because of the structure of the business.

Host: That's that's really interesting. One thing I was going to ask you about is the this change in the buy versus, buy versus build mentality. Does that is that something you'll say? Oh, for

sure.

Patrick Flannery: I think, you know, we see, you know, post 2008, a good number of firms, particularly ones that weren't necessarily digital leaders become way more cost efficient.

Patrick Flannery: sensitive. I think there's certainly the potential to accelerate that even further with whatever happens post COVID. You know, bank earnings today, we saw this last week have been good from trading revenues. But that's not to say that though, you know, the bank as a whole is not, you know, sell side institutions are going to be a lot more cost focused, but also buy side institutions.

Patrick Flannery: If we don't know what the market's going to do, whether it goes up or down, but if your business model is charging us assets and you have fewer assets, you're about to become a lot more cost

Host: sensitive. No, that makes complete sense. Actually leads us very nicely to our next poll for the audience. So I'll launch this 1 and it's absolutely on that topic.

Host: So if you would implement DAPID as a service, would you? Bill, bye. Or are you currently undecided?

Patrick Flannery: Just as a purveyor of these things, I hope the answer is bye.

Host: Yeah,

Host: yeah, bye at the moment is definitely out in the lead.

Host: We'll give it a few more seconds just to wait till there's a few more answers come in. But if I just close it out there. We've got by 58 percent bill 20 percent undecided 23%. My guess of the undecided, if they think about it, they're probably going to go to the by route. But interesting. Still 20%.

Host: On the build

Mark Woolfenden: perspective,

Host: but some of the greatest

Patrick Flannery: customers are the ones that try to build and then buy so, giving it a go yourself will actually Maybe provide the humility to to create a great

Host: customer Yeah, hopefully they don't actually use up their complete budget during the you know During the build phase and leave themselves exposed on that score, but

Mark Woolfenden: I can see that so

Host: if I can move on to you julian here, obviously you guys over at NatWest have embraced data as a service.

Host: Can you talk us through how you are actually using... Data as a service obviously very much part of your fixed income business now. And how you're optimizing AXE distribution using this new technology?

Matthew Cheung: Yeah, sure. So maybe before I start,

Julien Dugat: for those who don't know an AXE is is an indication that we send to our clients to just to indicate that we're interested in buying or selling certain assets.

Julien Dugat: And this asset can be a bond, an asset swap, an interest rate

Matthew Cheung: swap, so

Julien Dugat: anything else really. So for instance, we'd be sending an axe to tell our clients we're interested in buying a certain quantity of a certain bond at a certain price. And when it comes to distributing these axes it's very important that the data is timely so always up to date.

Julien Dugat: That it is

Matthew Cheung: it is actionable

Julien Dugat: that it is it's relevant to each and every clients and that a client can access the data easily. So in order to solve all of this, we've we've partnered with with and I'm sure Matt can can give you more details about his platform. But in terms of how we using, I push, boom.

Julien Dugat: We basically streamline data to the, I push pull controlled cloud. And then we let our client pull this data in real time into like Excel symphony and various other endpoints. But if I go back to access distribution, so as I said you know, this data needs to be, needs to be time end up to date.

Julien Dugat: The worst thing that could happen with Axis is getting a call from a client on a stale ax. So maybe we send an access morning saying, you know, we're interested in selling 10 million of this particular bond client calls us later To trade on that and we don't have the inventory So that's really bad experience for the client and really something that we try to avoid.

Julien Dugat: So how we do this is we avoid the like the manual ways of communicating this axis, so whether it's on the phone or emails or unconnected spreadsheets and instead we, we stream our data to clients directly. So everything that the client can, can see and access is live. So we know it's in sync with our internal view of the data.

Julien Dugat: The second aspect was about this data being relevant to clients. That's especially true for bonds. So each dealers would be sending and pushing tens of thousands of AXE messages every day. So being able to target specific clients with AXEs that we know they're interested in is very differentiating.

Julien Dugat: And how we've how we've implemented this in Atlas Market is we've got we've got an AXE monitor that our sales people use to monitor all our live AXEs. And we've integrated this with with iPushPool directly. So for instance, if you've got a client calling us and saying they're interested in buying some bonds in in the auto sector, uh, the salesperson can go into their monitor, they will filter the axes for the auto sector.

Julien Dugat: They will select the axes they want to share. And by clicking just a single button, they will start streaming these

Matthew Cheung: axes. To client you know, I push boom,

Julien Dugat: um, and clients will be able to then get this in real time straight into the desktop, which

Matthew Cheung: brings me

Julien Dugat: to the next the next aspect of it,

Matthew Cheung: which is accessibility of the data

Julien Dugat: is very important for, for clients to be able to consume our data where they want it and how they want to do it.

Julien Dugat: And for most users in front of this, that is on the desktop in Excel and more and more actually in the symphony chat platform. And what we really like with iPushPool is the fact that they do have all these endpoints. So we have we have an Excel add in that that, that we publish with iPushPool, which we make available to our clients for free.

Julien Dugat: That's enabled them to just load or X data in real time online into their spreadsheets. buT obviously they can also do this in Symfony, on the the iPushPool web app or mobile app. And, and all of these endpoints have a very, very low barrier to entry. So clients don't need to install anything on their desktop for this to work, which makes it really quick for us to start pushing this.

Julien Dugat: And the last point if I may, is we, we're really trying to get this axis to be actionable. So it's very important that a client can actually trace on this axis. in a very efficient fashion. So we've got we've got an execution bot on on symphony called scouts and scouts enables client to you know, to request life prices and to do execution workflow directly on symphony.

Julien Dugat: So what we've done is we've integrated scouts with I push boom. So what this means is that clients can seamlessly go from an which we've distributed on symphony. into an execution workflow with Scout. So they can in effect trade on Oaxis in Symfony. And that's probably quite a comprehensive answer

Matthew Cheung: already.

Matthew Cheung: I've got a question for you, Julian. Go for it. Why did you not choose to build this yourself? So

Julien Dugat: there's probably a few reasons. The main one is is speed to market.

Matthew Cheung: So

Julien Dugat: the, being able

Matthew Cheung: to buy a product off the shelf

Julien Dugat: like Appushpool, that needs very little integration to work was quite key for us.

Julien Dugat: So we, we, you know, once you bought it, you don't need to spend ages customizing the product and integrating it with your own data feed and all of this. You can get going really quickly. So you can do Excel to Excel very quickly. So we, we started by doing that and then we gradually. adapted the product to our needs.

Julien Dugat: So as I said, we started integrating our axe monitor to iPushPool. We started doing the integration on Symfony with Scout or Execution Bot but that happened gradually. So you can really get some views out of iPushPool out of the box which was quite

Matthew Cheung: important

Host: for us really. That's grand. Mark, can I pull you in at this point?

Host: You've been pretty quiet so far. I'd love to give you some, from your background Yeah, what are some of the commercial benefits to to data vendors or other others from data as a service? And how can it help facilitate, you know, potentially more flexible business models?

Mark Woolfenden: Cool. Thanks, Clive. There are a number of commercial benefits for both the suppliers of data and the end user, but most obviously only paying for what you need is a clear benefit for an end user.

Mark Woolfenden: I think, broadly speaking, today, many clients now only really want to purchase data that they have an active use for, so typically based on open positions that are being handled for clients to right size transaction activity and stay costs, although the degree of clarity on budgets related to on demand consumption are two conflicting things in general.

Mark Woolfenden: But if you view the, the economic model of supply and demand and, and the price volume curve of, Goods and services, the more granular the price volume changes are, the smoother the curve you get, and conversely, with larger price volume changes, with clunkier type graduations, the more of a sawtooth curve that becomes.

Mark Woolfenden: And I would suggest that many of the current market data contracts are designed for bulk data delivery for large investment firms who serve global products and markets with a clientele to match that. So data to service offerings may provide the opportunity for small to medium firms to access same or similar high quality data sets that they're currently unable to due to prohibitive costs.

Mark Woolfenden: And as such, data to service will allow such firms to

Matthew Cheung: match and scale operational

Mark Woolfenden: costs more closely with their, with their trading activity. thIs is, however, quite a big transition from where we are today with systems and contracts. More flexible business models can include placing data on demand in the trade life cycle, from pre trade risk validation with dynamic checking of, say, contract size and tick data, for example, through to post trade regulatory compliance, such as pre submission data field checks on transaction reports, and portfolio management.

Mark Woolfenden: such as dynamic symbol lookup to send repricing requests to pricing engines. Data on demand can streamline trading processes such as resolving trade breaks where currently well over 85 percent of trades are automatically processed with the further. Approximately 10 percent or so resolve using bots these days, leaving the last few percent of trade breaks that need actual human intervention.

Mark Woolfenden: And these residual trade breaks can be corrected using community networks, for example, such as Bloomberg, Messenger, Icon, Team Slack and Symphony, to name a few. where users can message co workers within their firm or externally to counterparties or within a secure environment in a pinpoint focused and data driven workflow at a very granular level to resolve the trade break.

Matthew Cheung: We should,

Mark Woolfenden: however, recognize that mass data publishing is still very effective by large funnel transfers. And I believe I'm right in saying that most of the world's payment processes are still done by large mainframes which represent less than 10 percent of global payment costs. So there are still great efficiencies in the bulk transfer of data, and I don't see that really reducing to any great extent.

Mark Woolfenden: in the foreseeable future. But what I do see is the combination of bulk file delivery being complemented with top of the datasets that aren't regularly used. So basically, you only pay for what you use outside of that bulk coverage required by large firms, for example. The the move to data on demand via APIs and other digital channels, such as Excel add ins, bots, and community meshing systems that I've just mentioned.

Mark Woolfenden: play to the fact that small data can be delivered to the point of absolute use on demand and in a live environment and so is complementary in the main to existing trading systems. However, I don't see a great shift away from Bolt data sets anytime soon because of the legacy systems that are still in place that have to process them.

Mark Woolfenden: But I do see a lot more investment into the use of real time reference data on demand as as part of the trade life cycle.

Host: Brilliant. Very, very interesting insights there, Mark. Thank you very much indeed. Just for the audience, we've got a number of bits of Q& A coming in so please keep those coming and we'll we'll move off to some Q& A in a little while.

Host: Brings us round to

Mark Woolfenden: our last

Matthew Cheung: poll,

Mark Woolfenden: uh, of the

Host: day before I bring Matt into the conversation. So, I'm just going to launch this one for you now, and it's what end user app would you use to consume data as a service? Excel, Symfony app, Symfony bot, API or other. Just get an idea of what the demand is for the for the different types of

Mark Woolfenden: consumer interfaces.

Mark Woolfenden: Clive, can I say that while people are polling, these these apps are actually complementary to existing trading systems, so they're actually additive,

Matthew Cheung: and

Mark Woolfenden: they can, they can provide a much needed injection of efficiency gains

Matthew Cheung: within legacy systems as they still,

Host: they still stand. Yeah, I, I, I can well imagine definitely.

Host: So if we look at the the responses to the poll there we've got the winner, API, 63%. Next favorite, interestingly, Excel, the good old favorite that's been around for years. And you can't, you can't get them out of your investment bank. It's going to be difficult taking Excel out, kicking and screaming.

Host: Then Symfony out, Symfony bot, and, and other. So thank you very much indeed to the audience for the insights there. Matt, if I can bring you in here, obviously we've talked about how data is being serviced, service being used, both on the buy side and by the likes of Matt West, some of the difficulties that Patrick was talking about.

Host: And picking up data service as, as, as a new service for your, for your business and some of the business and business implications from Mark. But from your perspective, yeah, can you give us an idea

Mark Woolfenden: of how iPushboard

Host: particularly helps clients to adopt data as a service? You know, how that is a nice, quick, easy win.

Host: And also, can you share some of your success stories with us, please?

Matthew Cheung: Yeah, yeah, sure. I'm really glad for the results of that poll. People wanting to hook up to APIs is a great thing. You know, it is the easiest, most seamless way to do it. And it's good that that type of technology is being adopted a lot more.

Matthew Cheung: Excel is always, I mean, we, we call Excel probably our gateway drug. Because, you know, if you walk around a trading floor, Excel is the thing that you see probably as much or maybe more than Bloomberg. Yeah. It is everywhere. It is ubiquitous. And as much as people want to try and move away from it, you know, big firms, you know, EUC policies and big firms are to move away from these end user computing applications.

Matthew Cheung: But you know, it's still very prevalent. And probably also on the buy side as well, because they don't have those same EUC policies. How does iPushpool kind of, how do you address some of these problems and how do we how do we help? I mean, we work with firms across the entire trade lifecycle, so from the front office who want to be, want to be able to embrace the ability to easily share real time data to clients, you know, across the desk or to the back office.

Matthew Cheung: who might want it, who might want it live and on demand. And just to clarify kind of the, the nuances in terminology, you know, there's, I talk about live, talk about streaming and on demand and they're all slightly different. By live, I mean, update, you can update data in one application and then you see it updating somewhere else at the same time, like you might have used something like Google sheets before.

Matthew Cheung: And you get that type of functionality, whether it is live by streaming, I mean, the typical type of streaming data you'll see in the markets. So ticking prices and quotes and so on where, where, you know, typically on sales and trading or an execution desk on the buy side, you'll need that type of streaming data.

Matthew Cheung: And then you've got on demand, which is you know, what we talk about with, with you know, your money and main street where they're doing the heavy lifting in. getting all this data together and petabytes of data and millions of rows in a database type data they're dealing with reference data and market data.

Matthew Cheung: But then those downstream applications, you can use something like a bot that's sitting in a Slack app, in a Slack chat or teams or symphony. And that bot can just query this big heavy database that's sitting in the background. So there's different types of different types of ways you can use the data and consume it in terms of success stories, I mean where customers already have the data And have the platform they might they might lack the distribution, you know to end user applications Our customers are using us for kind of the either the first mile or the last mile of connectivity depending on how you look at it The first mile being data that might be unstructured or sitting in an applicational system that doesn't normally have any external connectivity and then by connecting that into our platform that data can be you know, securely pushed into the cloud be made available to other people.

Matthew Cheung: By the last mile, and it's really something kind of picking up on Patrick was saying about these downstream applications, it means getting applications, sorry, getting data into applications that people are using, and they're already using. So you don't need to install anything new. That data can be coming directly into your spreadsheet, directly into your chat, directly into your internal blotter or platform that you might be using.

Matthew Cheung: It's that kind of last mile of distribution. That poll was interesting. Like I said, I mean, out of our 20 or so integrations that we have, I would say you know, 90 percent of users are probably pushing data into I push pull either through an API, a database or Excel, and then distributing that in real time into Excel.

Matthew Cheung: API and a symphony and in symphony, there is an app or as a chatbot. There's kind of different, different things you can do inside that environment. In regards to success stories, I mean, you know, really glad and happy to have this amazing panel on board. And there's a number of projects we've kind of spoken about here and all of them are utilizing the cloud importantly.

Matthew Cheung: You know, nothing is on prem with, with everyone on, on this panel, everything is utilizing the cloud. And I really think that digitization of these type of use cases is going to become more commonplace. You know, people are going to question why are they using either expensive developers to connect things together or why are we doing this manual process where someone manually and literally physically is having to copy and paste something from one application to another, to another, send it over.

Matthew Cheung: via voice or chat. Someone else on the other side also has to do something similar. Lots of manual processes, which is emails, spreadsheets, copy, paste, you know, by moving to this new way of data sharing, it kind of unlocks this technical efficiency and ultimately automation as well. So then humans can be spent.

Matthew Cheung: Looking at higher value tasks, you're already seeing a lot of that in in the electronification of the OTC markets. You know, you've seen, you know, trade webs and market access and companies like that electronify these historically, you know, manual type of workflows, but then why, why have it stop front office electronification?

Matthew Cheung: Why not? Why can't that type of workflow and data driven workflow be you know, Start organically growing across the organization. So I think we see gonna see live data sharing becoming ubiquitous, both internally connecting onto all these applications and then externally to your clients, your customers and also, you know, connecting into technology vendors who also can take.

Matthew Cheung: Take what we've built and utilize it themselves into their own networks as well. Definitely a network effect going on and we're not, we're not the tipping point yet, but we're getting, you know, we're getting near that and the cloud is just, you know, ramping, ramping that up. So it's a very exciting time to be looking at data as a service.

Matthew Cheung: Matthew, could I ask

John Macpherson: you a question? Yeah. Would you say the two biggest hurdles for adoption are around infrastructure and operational

Matthew Cheung: hurdles, I suppose? Yeah, a hundred percent. Yeah, yeah, typically we'll, we'll go into we, when we're selling, we're selling into, into business, into a desk and, and they'll identify problems and we're a solution for it.

Matthew Cheung: And we will go into the bank. It is more, this is more for banks as well. For your less so as you as you the larger the institution is, you know, obviously things are going to be slower. Yeah, we run into can we even though there's been adoption of cloud in the industry, then that's great. But then what happens is, you know, we've had one company say, Oh, we use Azure and you're on AWS.

Matthew Cheung: Can you, can you rebuild your platform in Azure? It's like, well, yes, we can, but it's going to cost you more money because we're in AWS, but it raises interesting questions like, okay, well, can you have a cross connect between AWS and Azure or there's another bank we're dealing with who they are, they do use AWS.

Matthew Cheung: However, they want to want us to deploy in their environments, which means we need to change things how we're set up So I think it's it's the the infosec barriers are changing I think they're becoming more adaptive and more agile kind of the the infosec departments in banks, I think that's been led by where you have like innovation departments or when there's been a top down strategic decision to say, right, we want to start using this technology, let's try to lower the hurdles a little bit, because if you're dealing with startup organizations, you're not dealing with IBM or SAP, when you're doing some big deployment, they need to do it quick and fast.

Matthew Cheung: And that's what makes the startup Yeah, that's that's a competitive advantage for startup to be able to move quick. If you kind of throttle that down so things are slower, then you're losing some of the benefits there as well. And lastly, I think around deployments, if you're deploying on premise or on someone else's cloud, and you can't have access to it, it just means that you can't have that really quick, agile, iterative approach when you're building.

Matthew Cheung: products and you know, maybe, maybe we're helping to build a GUI for someone and they want to, I want this button here and I want to click this and be able to do that. Oh, you're going to have to wait until next month when the firewalls open up and we give you access. Yeah, it's, it's, it's, they're, they're a little bit like that.

Matthew Cheung: I think some, some old legacy type of processes and then what the cloud is on kind of unlocking, but it's, you know, it's a, it's a time of change. I mean, we've seen banks and by side, you know, change the way they view these things, even in the last 12 months. You know, people's perspectives have changed quite significantly.

Matthew Cheung: I don't know, like, Mark, Patrick, Julian, have you got any comments on that as well? I would like to follow

Mark Woolfenden: on to John's really. Do you think the IPishpool technology can assist in big firms moving or breaking down their legacy systems almost brick by brick to eventually change the whole of their infrastructure?

Mark Woolfenden: Do you see a

Matthew Cheung: role for that? Yes I think it's, we can move. because we integrate into legacy technology as well as cloud technologies. It's a way that exactly brick by brick, you can start filtering some data through and then over time yeah, we kind of go in on a, on maybe on a tactical use case and you end up becoming a strategic part of the, of the infrastructure over time.

Matthew Cheung: So yeah.

Host: A few questions coming in for the audience here. If I could just. Put this one to you, Julian. Should be fairly quick. Does your client know that it is an axed quote when accessing the data stream from you? I'm assuming the answer to that is yes.

Julien Dugat: It is, yes. That's a good point, actually.

Julien Dugat: So the, the data basically we take through iPushPool is so it's stricter as it is. So we, we only push access today. But we are looking at other data types. So selling things like prices and started looking into into workflow as well. But for now, yeah, absolutely. And it's it's very clear that all streams are

Host: axes.

Host: Brilliant. Next one, I think probably for you, Patrick person here saying data service is not new. It's been around for years, especially by outsourcing middle and back office or pure play data service like Tech Mahindra. Former

Mark Woolfenden: prevalent

Host: however, the outsource managers have not bought into the pure play model.

Host: Do you have any insights as to why?

Host: Well, I think,

Patrick Flannery: I think yeah, I mean, one of the, one of the things that's cited in there is you know, market data licensing can definitely play into it. The sophistication of different sorts of firms I think, you know, there's kind of You know where you sit within the hierarchy and kind of a firm's own adoption kind of describes a little bit of How they relate to their customers so certain sell side firms have a little bit more sophistication than than a number of buy side firms.

Patrick Flannery: So I think I think you will see it and the problems are solvable, but it's you know Levels of understanding for for a market participant and their counterparties

Host: Excellent. Next one, if I could put this to a mixture of you Matt and and Patrick basically the tools offered by you to to your customers, are they actually affordable for smaller institutions rather than just the the large buy and sell sides?

Host: So you know, how, how easy is it for the uptake for, Some of maybe John's smaller members of the Investment Association to to start using data as a

Matthew Cheung: service. Yeah, I'll grab that first, Patrick, if you want. Yeah, I mean, again, going back to keep banging on about the cloud, but because you now have the cloud.

Matthew Cheung: the cost barriers. There's a lot more flexibility around that now, where previously, even us as a company three, four years ago, we were doing on premise deployments, which then becomes you know, has a lot of support, maintenance and overhead and so on. So that's helped in reducing cost for end customers who want to use our service.

Matthew Cheung: And we have, you know, talking about kind of buy sides and sell sides and big IDBs and so on. We also have kind of a raft of hedge funds and prop traders that use our service. Yeah, we got quite a few prop trading groups that use us for consolidating real time risk and then putting notifications on, on, on certain P and L levels that are going into Slack and going in as an SMS message as well.

Matthew Cheung: So yeah, we're dealing with plenty of firms that are on the smaller side. Patrick, you've got anything to add to that?

Patrick Flannery: Yeah, we, we deal with, you know, tiny firms up to, you know, government agencies you know, banks, you know, big, big and small firms alike. I think the cloud has, definitely does have the potential to offer, you know, sophisticated solutions although we're not entirely there, definitely moving in that direction of on demand.

Patrick Flannery: Right. So, you know, if you're accessing a small portion of the ocean of data you have the ability to pay for a small. So I think that that has a lot of potential and when done well, looks a lot more like the highly successful large technology companies that you're able to consume, you know, huge amounts of value.

Patrick Flannery: You know, at piecemeal, you know, kind of pay per drink sort of model. So, yeah, I think, I think that's from a question from someone I know who we've met before. So

Host: it sounds like you've you've heard that one before. Question here for Julian. Are you seeing more business shift to OTC and away from trading venues, given the ability for you to deliver data as as you are doing with those access and just, you know, does that really help you take more market share

Matthew Cheung: on on, on the book?

Matthew Cheung: So, I mean,

Julien Dugat: you have, in terms of faxes, really. You can split it between bonds which are very very electronic today so we stream on our axes already onto all the electronic platforms onto bloomberg trade web market access on vision and so on And then you've got everything else, which is not electronic at all.

Julien Dugat: So there is no platform that offers access for IRS and, you know, asset swaps and all of this. So a lot of the business where we use iPushPool is for, for enhancing our voice business. So it's providing clients with a much better experience for voice business. It's what we call high touch basically, so it's a digital high touch business so better quality flows for clients but but focusing mostly on Resilience or salespeople basically in this workflow,

Host: right?

Host: That's that's brilliant. Thanks june. John, can I bring you in again here, please? One here on security of data distribution and control given your experience, both, you know, Goldman's BML and with the Investment Association, you know, that control of governance was obviously a very, very big hurdle for larger firms with legacy platforms.

Host: Are you still seeing this as a major hurdle for your members today? And if not, what's bringing about those changes? I think comfort with the

John Macpherson: cloud is the primary thing that's changing people's perception on this. I believe that it has been, it's constantly been an issue. The biggest hurdle really is the sort of crown jewels each of these firms hold, which is their own trade data and how they have interacted with the market.

John Macpherson: The amalgamation of that with the actual market data, if you like, is the Holy Grail. Some people are trying to do that on premise. Some people are trying to do it in the cloud, but it is, it is becoming more normal. Let's put it that way. Asset managers are, in my opinion, the most conservative trading type out there.

John Macpherson: And so they have been slower at this, but it's getting there.

Host: Is Mark, are you, have you got any insights to add to what John was just

Mark Woolfenden: saying there? I think the The key thing

Matthew Cheung: is, is that the the requirements of

Mark Woolfenden: a, of a smaller user in, in a marketplace are that, that they don't necessarily have time to, or the resource and staff to, and IT staff to um, you know, deal with data fees.

Mark Woolfenden: They need a platform or a working solution that that wraps around. The, the data. And I think that the, the choice and availability of such solutions to the, the smaller user you know, can be complimented by the the artificial product. For example, the Excel addin and sym bot solution as well.

Mark Woolfenden: So I, I think it's, it's about how you can access high quality data that was maybe in the realms of, of, you know, the big market data, consume multimillion pound contracts. How do you. Get the data for the smaller firm is actually more focused on a

Matthew Cheung: specific market or or or product,

Mark Woolfenden: and I think this this is a naming technology to do that, but it is complimentary, and I think I think you know there are a lot of prospects and opportunities for the buy side with this technology as there is for sell side in different

Matthew Cheung: types of areas.

Matthew Cheung: Can I follow on that question to maybe John and Julian, like, John, you mentioned that asset managers historically have been a bit slower to adopt new technology and the question probably goes to both you and Julian. Is it, is it, I mean, taking a step back, you would just kind of. safely assume that perhaps that's because asset managers, because they are, you know, they're the real money, they've got the flow, they, they, you know, the banks will run around to provide whatever technology, the asset manager demands.

Matthew Cheung: Is that one of the reasons why historically, they haven't kind of made as much efforts to keep up with With technology whereas on the cell side the cell side needs to be Out the cutting edge all the time with every single different type of distribution The mechanism and connectivity so that whatever customer asks for it.

Matthew Cheung: They can connect them up easily And

Mark Woolfenden: in

John Macpherson: short, yes the buy side has relied on the sell side forever for information flow, etc. That that that is that is not an argument. I Think it speaks volumes The investment association in its most recent cohort adopted. I push pull. So that tells you from all the people who made that decision that I push pull represents something they want to learn a lot more about and become more savvy.

Matthew Cheung: The other side, I think

Julien Dugat: it's probably it's probably a balance, right? So if you look at the market, uh, if every, if every asset managers were talking to each of the dealers and asking them to adopt their particular API, and if it was the same thing the other way around, you can imagine the number of connection and the huge amount of work that they could create.

Julien Dugat: So having an abstraction layer in between is. That's the obvious solution, right? So something like iPushPool you can put in between. So we can push using whatever API we want into the cloud. And then our clients can start loading this data using either an API or Excel or whatever means they want.

Julien Dugat: It's just making everybody's life easier, really.

Host: Question, I'm going to start off with you, Mark, if I can. We've talked about distribution of data, consumption of data, flexibility, ease, but how critical is the quality of that data and managing that in the first place? You know, if you could give us some insight on that, that'd be great.

Host: I'd be interested to hear thoughts around that from other, other members of the panel as well.

Mark Woolfenden: Thanks, Clive. I think data quality is even now more critical in today's modern trading environment especially when it can be the basis for AI type of processes and machine learning processes, which require almost perfect input to achieve anything, anything meaningful.

Mark Woolfenden: And also the cost of fixing trade breaks is becoming disproportionately higher. And whilst the number of trade breaks is actually decreasing, the cost of exceptions becoming quite a common concern for businesses as they, as they recognize high quality data can, can essentially minimize train breaks occurring.

Mark Woolfenden: I, I think you go back 10 years and the tolerance on quality could have been in double digits, but today zero tolerance

Matthew Cheung: of, of, of quality required.

Mark Woolfenden: And we have. Basically, in terms of our firm,

Matthew Cheung: we have modified and enhanced

Mark Woolfenden: our data management

Matthew Cheung: and data quality processes.

Mark Woolfenden: We have ISO 9001, which is a quality management standard that really is a minimum requirement to be able to publish data with the highest quality at the most cost effective price.

Mark Woolfenden: So it does have a sort of price point. But the requirement for high quality data. It's absolutely essential. We've heard a lot of talk about technology, new and legacy, but without the high quality data or the right data in the system, it's, it's, it's, it's, it's not meaningless, but it means that you, you get more trade breaks, you get more points of failure.

Mark Woolfenden: So it's something that will prevail through whatever technologies arrive over the next two or three decades. You still need to be confident in the, in the data set is the highest possible

Matthew Cheung: quality.

Host: Great. We've got not much time left, actually. We've we've managed to fill that lot of time pretty well.

Host: We've had a few questions, or quite a few questions I haven't been able to get get round to. I think, Matt, before I... Ask you to give us a sort of wrap up. What I'd love to do is go around the round the room. I know John has to leave us fairly promptly at 4 30. So come to you first, John. And how do you see data as a service?

Host: Growing within your organization, I suppose, within your members, given given your role over the next year

Mark Woolfenden: or

Host: so.

John Macpherson: Thank you, Clive. It's how is it going to grow? I think, you know, there are a lot of firms now using it actively. I think there's an awful lot more exploring it at different levels of that, but I think everyone realizes that rather than going back to what we were talking about earlier, rather than the, the build first and buy last option and the way that is completely reversed, people are now prepared to look at how it will help them streamline, how it will lower costs create a faster path to innovation, allow them to make more agile decision making processes creating a data driven culture lowering risks and leading to better revenues.

John Macpherson: So once those dots are connected, even some of them it will become more prevalent without a shadow of a

Host: doubt. That's great. Thanks very much for that. John Patrick, can I put the same question to you, please? You know, over at Mace Street, how does data as a service grow over the next year or so?

Patrick Flannery: Yeah, so I think, you know, for each for each of the consumers as data as a service It's interesting to think about what their objectives are as a business Is it to reduce costs is to reduce errors to make better investment decisions?

Patrick Flannery: So, you know the the the ways that it grows are definitely you know different in the front middle and back office right and reflect the sophistication of the firm so I think we you know, you'll see firms all kind of moving up the stack. So firms that are just looking at it are trying to do some of the simplest things, prevent errors, right, make Data more available at the time of the decision, whatever that decision may be.

Patrick Flannery: So incorporating the data into broader processes is always the, is the trend, right? So, you know, maybe you have one input today and in, in a few years you have 10 and, you know, 100 sometime after that. So we'll, I think the, the trend that we'll see more than anything is, is using more data in more workflows.

Patrick Flannery: For, you know, whatever, whatever those original business objectives are and there are some really, you know, phenomenal opportunities over the scale of years of how, how do you, how do you get the right data at the right moment in the right format so that people can make automated better decisions?

Patrick Flannery: Which is, you know, non trivial problem.

Host: That's an interesting one. Yeah, Julian, over to you. Same, same question. I suppose you'll, you'll be looking at people automating, executing against Axis and how you move that sort of thing forward. But over and above your use of you know, ice push pull services to get your Axis out to the to your clients at the moment, how do you think things will change for you over the next year or so?

Matthew Cheung: Well, I think

Julien Dugat: perfectly spot on. It's really about. It's about getting the right data at

Matthew Cheung: the right time

Julien Dugat: to the right client for us. So, it's making all these data more relevant we got loads of data, we distribute loads of data we, we now have a really good way to distribute it and, you know, get it to the right place.

Julien Dugat: So it really is about making sure we make this more and more relevant that we. We basically send to clients what they want, basically. So rather than just inundating everybody with, you know, generally data, really make it, make it relevant, make

Matthew Cheung: it to the point.

Host: Okay, last but not least, you Mark, and then we'll hand it over to Matt to just give us a bit of a wrap up on the, on the conversation that we've had

Mark Woolfenden: today.

Mark Woolfenden: Sure, well as a traditional data vendor having used, been used to providing bulk data feeds to many tier one banks. We come up with a phrase which is we've used for many years, which is, we only do what only we can do. And we're more concerned about, curating the data set and understanding the complexity of the futures and options universe that we're actually active in.

Mark Woolfenden: So that is why when we looked at our strategic plan, we wanted to attract more second and third tier clients, the route to market. was either through partners or through directly through technology. And that's the reason we're, we're a classic buy from iPushball to the extent that we, we are using their products for Excel add in and Symfony app and bot to access the, the, the lower tier markets or the tier three markets in particular.

Mark Woolfenden: And so I hope over the next year or two at least. is to increase our presence in that market and to essentially publish this, this factory of high quality data that is used and proven by Tier 1s 2s already into the rest of the market. So we're looking for an active engagement with iPushPool from

Host: here on in.

Host: That's great. Thanks very much indeed for that, Mark. Matt you know, we've had quite a bit of conversation about, you know, what is data as a service? What are the economic benefits? What are the efficiency benefits? What are some of the difficulties in, in applying? Can you just give us some sort of, some wise words just to round off the conversation a bit and and some final thoughts,

Matthew Cheung: please.

Matthew Cheung: Yeah, sure. And yeah, just wary of time. So I'll kind of wrap this up quickly. COVID has transformed the landscape, you know, for the majority of those that work in financial markets who are now working at home. Yeah, there's no new norm anymore. There is no norm. There's only a future state. So that means our working practices will change.

Matthew Cheung: Workflow needs to be more efficient. Data needs to be easy to access as well as being secure and access controlled. As we move towards live data driven workflows, people need to seamlessly connect to data in an, in any application in real time. And also at the right time and the right place, like we've just been talking about.

Matthew Cheung: And from any location. You know, now that people are not working in office anymore, that needs to be easily accessible. So we're seeing data as a service being adopted across sales and trading between sell side and buy side across technology vendors, all of whom are providing a better, more efficient experience for their clients.

Matthew Cheung: So moving away from manual processes, emails, spreadsheets, copy paste, not using expensive development projects to connect things yourself. Instead, you know, look to incorporate data as a service into your digital transformation projects or as a new digital distribution channel. So on that note I'll kind of round up here.

Matthew Cheung: So, so thank you everybody for listening. Really appreciate it. I hope you learned something new today. Thank you to the panelists. I mean, John had to leave early, but thank you, Mark. Thank you, Patrick. Thank you, Julian. Appreciate your time and thanks for hosting Clive. Thanks very much,

Host: Matt. Just a reminder to everybody.

Host: Please go on to the iPushpool website to download the report, give you a bit more information there, and also reach out to iPushpool if you've got any more questions around their data as a service offerings. Again, thanks very much Riz Rehmat thanks for everybody to attending, and also the panelists.

Host: Have a great rest of your

Matthew Cheung: afternoon. Thank you. Thanks, everyone. Bye. Thanks. Bye.