Sayula's President and CEO, Cristian Alzati, tells us how Sayula's Anima is bringing together Real-Time Analytics, Excel, the Web, and the ESB by deploying all logic and data where they belong: on your Linux and Windows servers. Dan Tudball reports.

CRISTIAN: It all started in Tokyo. The original vision really started from my own work experience as the IT support manager of Lehman's FX and Fixed Income desks. What happened was that we had to support dozens of Excel add-ins that were in active use by traders, sales people, and their quant analysts. The main problem was that although the same add-ins were used globally, for practical reasons they were supported regionally. Thus you can imagine the kind of problems we were running into on almost a daily basis. Spreadsheets using both global and regional add-ins will be consistently inconsistent. So it was routine on a Monday morning to get completely meaningless numbers only because somebody in London had done a global add-in release over the weekend. It was a real nightmare. I then started to think there's got to be a better way of doing this. And that is how Sayula and the Anima concept were born. The idea we came up with was indeed a straightforward one. It's simply deploying the analytics on the server side, while at the same time providing Excel with a generic mechanism to access the server-deployed functionality in a completely transparent way. No matter what you put on the server, be it data, analytics, or even real-time market quotes, you will be able to access it in Excel same as a web browser can render any content type, only better because it's in real time. So what Anima has managed to do is to convert Excel into a thin real-time client and we believe this is a great achievement. A true paradigm shift, we like to say. Not only have we been able to solve the spreadsheet inconsistency problem in a single stroke, we have also completely opened up the Linux, Solaris, and Windows servers world of data and real-time analytics to all the until-now miserable Excel users of this world!

CRISTIAN: It all happened over approximately one and a half years, from concept to a working prototype. Last year when we first presented the prototype to UBS they became immediately interested. For them Anima was especially attractive because of the number of home grown and third party analytic libraries they use. Since the libraries have been developed in different programming languages and by different teams or companies, they all have different interfacing mechanisms. They wanted to have a standard mechanism whereby they could make all these libraries' functionality available in Excel. So it was quite an amazing thing for them (and for us too!) when we first came to UBS and were presented with one of their libraries in the shape of a dll and its header file. Could we make that available in Excel right there? We could not really look inside the library since it is obviously highly proprietary information. But there was no need for that because by simply adding the library function's signature information from the header file in Anima, and copying the dll into a specific directory, all the functions became immediately available to all the Excel users on their network. It was done in minutes, literally.

CRISTIAN: Absolutely. Just think about it. We can almost instantly connect all your existing analytics to Excel, even the legacy ones for which the source code is not available any more. With Anima there is no need to write a single line of code, let alone to recompile anything. There's not even need to restart Excel! And since it is server based, everybody gets to use the same version of the code. It's the kind of functionality the front office wishes for and the middle office dreams about. Now, wouldn't it be nice if both of them can be guaranteed to always use the same logic to calculate sensitivities, mark positions to market and recalibrate models?

CRISTIAN: Yes. This of course has to do with not only data consistency, but equally importantly with procedural and algorithmic consistency. With Anima you know the numbers you are seeing in the spreadsheet are the actual numbers your audited analytics are calculating and not the result of an old version of the library, hidden somewhere in the dark corners of your Windows box. That will definitely make your compliance department happy.

CRISTIAN: Well the fact that we have managed to shift Excel's analytics execution to the server affords us tremendous power and flexibility because the benefits can now be easily extended to the full enterprise via the corporate web and the enterprise's real-time bus, the ESB in Gardner's terminology. So what we have done is to go one step further in the abstraction chain to implement a framework to deploy full-fledged analytical models on the server side. The mechanism could not be simpler: create the model in Excel, upload the Excel file in Anima and that's it. The model is now available on the Web via our Excel-like Flash client. You want your model to publish data into your ESB? Just define a real-time Anima context and you're done. From your standard web browser you can now enter and change the model's data and formulas just like in Excel. And just like in Excel the formulas can be either standard spreadsheet functions or your own proprietary functions. The Anima server keeps track of cell and formula dependencies for you and gives you access to market data right there on you browser, ticking away in real time!

CRISTIAN: The vision is to make available all this power to the quantitative finance community at all levels; high management, front office, product controllers, finance departments, etc. And we of course see the Web as the perfect vehicle for fulfilling that vision. Sayula's main marketing message "Your analytics anywhere, anytime" conveys this philosophy well. Wouldn't it be nice if you could make available to everybody authorized in the global enterprise your own market data and analytics results, no matter where they are? If your quants have already done the hard work of coming up with analytic models that can determine your enterprise's global trading strategy, you want to make that work visible globally and immediately without having to worry about the ugly IT stuff. These are the kind of benefits that we are seeing hitting home again and again with our potential and current client base, and so far the response has been tremendous.

CRISTIAN: Yes, that's quite interesting. You know, those were the exact words I received in an email from them, together with a kind invitation to attend one of their developer's conferences. What happened was that we presented Anima to Reuters in Zurich at a pretty high level. They were very impressed with the functionality and the degree to which it can be extended. And in a way some of Anima's capabilities overlap with some of the functionality they provide on the desktop, especially for accessing real-time market data in Excel. One of the constant themes of our experience with potential clients is precisely their great desire to improve on the cost/benefits aspects of that functionality; thus the interest from Reuters, I guess. It is clear they have now realized the potential. So now the challenge for us is to leverage that interest and make it work for our business collaboration strategy. We of course don't want to be a threat to Reuters; we much rather be partners with them.

CRISTIAN: Yes, absolutely. We're very excited with Wilmott and Paul joining forces with us. First of all Paul has been a great inspiration and example for me personally, from my CMU days when I first learnt about him as the co-author of one of the few good textbooks we used. Then about a year ago when I approached him to explain and show what Sayula and Anima were all about, he became positively interested, and we of course were positively thrilled. We are now working together on our first joint project that should go live by the time this article is published. We are calling it Wilmott's "Live Models". It is quite an exciting project I must say. We want to make available in wilmott.com a simple web-based mechanism whereby members of the Wilmott community can contribute and share their own pricing models and analytic libraries. Contributed models will then become available for immediate interactive use on Sayula's Excel-like Flash client. Model parameters and formulas can be changed directly by users of the model, and if we can reach a data distribution agreement with Reuters, real-time market rates should also be seen ticking and feeding the online models. At first we will do a controlled deployment, with some of Paul's volatility forecasting models as pilot examples. But the real target is to allow wilmott.com members to be able to deploy the models on their own. We think such service will be of great use for the quantitative finance community, and so are very much looking forward to seeing it live soon!

CRISTIAN: I cannot even start to give you a ballpark figure, but if it could be quantified it will clearly amount to millions if not billions of dollars. Things have changed substantially from the times when Nick Leeson could bet the whole of Barings capital resources based on a single spreadsheet. But even so, to this day we are still seeing an ever increasing number of mission-critical spreadsheets producing numbers on which vital investment decisions are made. And there is not indication that this will change even in the long term. Excel is the most ubiquitous application ever to see the light in a financial institution, and this is a fact we should all try to happily live with. Firms are very aware of this and thus have been proactively taking measures to lessen the risks. One of the first steps taken on this regard has been to move most of their existing applications to the corporate web, a very effective measure of course. But Excel is different and although not impossible, with spreadsheet-based applications this has not been simple to achieve; unless you use Anima that is. Because here is exactly where Anima colors truly shine as it can effectively convert Excel into a web-like centralized application, with all the resulting risk-controlling advantages that it entails.

CRISTIAN: Well I'm glad you asked because we definitely have some big plans here. Together with Paul we have identified a clear need from the thousands of small hedge funds and other investment firms who will have a high interest on a web-based service whereby they can either use their own pricing models or the standardized models that we will make available as part of the service. Most of these funds and firms have neither the IT nor the quant analysis labor force or budget to implement an equivalent functionality. Such service will mean huge savings to them. And in order to broaden the asset classes that our models could support we are starting to develop partnership with specialized analytics companies.

CRISTIAN: Well, since we are not attempting to, at least initially, provide our own analytics models or libraries, we have started looking into partnerships with companies like Numerix, Xenomorph, or NAG, just to mention a few. We have already started talking with Numerix, but will also like to start talking to other companies like FINCAD and Xenomorph for example. We know they also have a close relationship with Wilmott and Paul so it makes a lot sense for us to join forces. It is a win-win relationship with them providing the specialized content and we providing the immediate time-to-market capability.

CRISTIAN: Well, our target is to have a base of at least 25 institutional clients by the end of 2007. The way we're planning to achieve that is through regional partnerships with system integrators. It's happening already in Belgium with ITG Consulting and it's also going to happen very soon in Japan with Simplex Technology. For them it's much easier to distribute the product with their established customer base, we then provide the specialized support and the intensive training for them to be able to offer first-level support on their own. We see this as an effective way to reach a wide client base while at the same time keeping a comfortable growing pace, a pace we can manage.

CRISTIAN: I'm the founder and the main architect of the system. My background includes Stanford, CMU, and almost 20 years in the investment banking IT. On the business-happening side we have Henri Lastenouse, our shiny Chairman of the Board. Henri's background is quite interesting because he is a lawyer with a very successful entrepreneurial track record. Henri was a co-founder of Hi-Media, a very successful internet-marketing agency, and now a French public company. On the European sales and technical side we have Jacques Van de Putte, Marc Dugimont, Bruno Flinois, and Patrick Labbe. And last but not least the seventh of the seven is Olivier Chambord who is behind Sayula's intellectual property issues.

CRISTIAN: Absolutely. Singapore and Hong Kong being the bustling financial centers they are, it makes a lot of sense for us to look for business and partnerships there. And given that the technical headquarters of Sayula are in Tokyo, it makes a lot of sense to partner with not only Japanese but also most definitely with Singapore, Hong Kong and Australian based companies.