Cesario: I graduated from UC Berkeley and went into business. I started at Wells Fargo Bank where I became a manager focused on the development of technology to manage business systems. I focused on database client server management and statistical analysis in Wells Fargo's fraud portfolios and then took that expertise to Pacific Bell, where I in essence created mechanisms for analysis of the marketing database of information on the customer base.

That led me to a start up opportunity at a company called Neomorphic. I, in conjunction with several of my college friends, started this company to do visualization of genomic analysis and statistical analysis of genomic sequence. We started that company in 1996 and ended up selling that company for $70 million to Affymetrix in 1999. I was COO of Neomorphic and now I am VP of IT for Affymetrix. That is to say I am the CIO and the person who runs all of the IT for the company.

OV: How would you describe your role and that of your organization at Affymetrix? How is your role at Affymetrix different from other roles you have held in the past? What do you like about it?

Cesario: This is the first time I've ever had the end to end responsibility. Everything from applications that facilitate business process that you would typically find in any company like ERP, which is finance and order-to-cash processes, customer relationship management, supply chain management or manufacturing and R & D processes like ECO, change order; all the way through to very business specific things like our website and our integrated database of genomic information that our clients use to do analysis on the Affymetrix change hits that they may run. Issues of R & D. Having this perspective has driven me to a view that my role and that of IT is to create efficiency and partnership with the business; deliver on the promise of IT; what's its enhanced functionality and competitiveness and value for the business that can come from this. As a case in point, we've created a portal - a web portal - in Affymetrix which allows our customers to get to biologically relevant information directly once they have results from the Affymetrix gene chip platform. That creates competitive advantage. Our ability to do that highly efficiently is through a web based interface that is a data integration play that takes our specific information and makes it central to the biologically relevant information that is available in the public sector.

OV: There's a huge push about CIOs now looking at their business as less of a cost center and more of a profit center; definitely a service center with customers being the business unit. Are you thinking along those lines? How does that play in with your own vision about your organization?

Cesario: When I took this responsibility two years ago, I built the IT organization on a consulting model. That is to say, every business group within Affymetrix is an internal client of mine such that we can partner with those individual groups to satisfy their business needs. It's a service based model and interestingly enough you find that same trend occurring more broadly in terms of outsourced services. I consider our ability to compete with outsourced services at the right price point in terms of creating efficiency, value; to be part of our value-add to our internal customers. This very much is the model we're building the IT organization on.

OV: What are the top challenges and priorities that your organization faces today? What is your vision for IT at Affymetrix?

Cesario: We talk about IT as a service so you've got different layers of service which can be provided. First you have core services like providing email and access to online storage for purposes of managing customer data. That's just kind of a baseline set of functionality; a secure network that is protected from the external world, from viruses and things like that. That's all service-based and customer support for peripherals like PCs and all that; that's just services. Then there are applications. Applications are where the service model, business model is nascent in its development and I'm not sure that I'm convinced that organizations are ready to accept a service model for applications and by that I mean outsource the ERP, outsource supply chain, etc. Ultimately I think that's where it will go; but what you see is, if you go and you look at a major applications implementation within any given company and Affymetrix is a good case in point, for instance we use JD Edwards on the ERP side. Our JD Edwards ERP instance is fairly uniquely customized and that makes it very challenging for us to outsource it effectively because then we would have to rely on an outside group to understand the nuances of that particular implementation. In essence what we've found is that you can't really find that for a very complex application like JD Edwards.

OV: We've seen a lot of companies that are targeting enterprises with hosted services such as ERP. It seems like that may not be a successful strategy because organizations similar to Affymetrix would have the same concerns in terms of the customization required for their particular needs.

Cesario: Right, and so I think I'm saying yes to that and the reason is that it's a cultural issue. When we're talking about applications, we're talking about processes, we're talking about people and individuals, which you normally find in any company, large or small is that the processes that have been developed there are to some extent unique. What I think will happen, the trend will be in order to gain efficiencies, in other words, the price point for the service there will be a vanilla core and applications that will be developed. For instance, Affymetrix, while we do have a highly unique product, we are in essence a manufacturing and distribution company. We manufacture an instrument, a consumable for that instrument, we have a client base which accesses that technology, and we provide that technology to them. We manufacture it and we sell it. You can generalize the end/end processes that go around that but you see when you look at Affymetrix is that our internal processes that grew up around that manufacturing/distribution process are unique from others out there. I don't know that there's a good reason for that; that's just a state of affairs. When you have a business model for out sourced services, you get into that problem of getting the infrastructure to comply with the generic process. That I think is a real challenge today but over time will diminish and make the model for a company that's doing outsourced applications services is looking to gain an economy of scale by having a very vanilla implementation across multiple clients. I think that's hard to get today. Customers will say, "well no I'm unique and I want you to make the application do XYZ." Customers see it as "it's unique to my business and I must have this functionality;" when in fact that might not be the case.

OV: Given the huge trend of outsourcing especially to India, China, etc., has this impacted your organization at all?

Cesario: Today we have a fairly successful CIS organization which is providing its core services and applications in a robust way that satisfies the business. We've spent the last two years right-pricing it; doing cost controls such that we can create a value proposition for the company and right now my cost of supplying internally those infrastructure services and applications is either the same or lower than an outsourcing option would be.

OV: Wow! That's very impressive. Most companies seem to feel that they need to have an "India IT" group.

Cesario: I very seriously looked at outsourcing major components of our infrastructure when I took this responsibility and what I found is that there are justifications for outsourcing if you have a very very large enterprise organization. We're a 1,000 person company with a global presence, which is to say we have offices at three different sites overseas. It's harder to justify when you're smaller than it is when you're larger so size matters there. The reasons for outsourcing were because in some way, shape or form a failure of the internal IT group to satisfy the business needs. I think it is understood and acknowledged that outsourcing is ultimately more expensive than in-sourcing. That's really a secondary issue, though. The primary issue is getting the functionality and the services that are required. My observation is that people say, "We're outsourcing because it's the most efficient way, the biggest bang for your buck, the cheapest way to get the infrastructure you need." But what I really think is going on is that internal IT groups have failed in most of the situations.

OV: What are the factors that have contributed to Affymetrix and yourself succeeding where others have failed in terms of not being required to out source for whatever competitive advantage? Do you see those factors changing over time such that you may eventually outsource?

Cesario: Right now I can compete. I have an organization which is run very efficiently. I have a high level of client satisfaction and IT is still inherently a fairly complex endeavor. You have unique end/end processes today. I've got a group of people who are trained to facilitate those processes as they exist here. I would be in effect introducing too much risk, today, to try to outsource. Absolutely. There will be a time when outsourcing makes sense for IT. I think that's the future especially for services. In the future there will be a compelling reason to outsource because at some point I will not be able to compete with other companies in terms of the outsourced services and that's when I'll outsource.

OV: What do you think is interesting now in terms of new technologies? What areas of high tech do you like at the moment? Are you working with any interesting start-ups or are there any start-up projects that you're particularly excited about?

Cesario: I think one of the core challenges of any IT organization is management of data. The trends in data management have been to - if you're a company of any substantial size - go with some sort of consolidated data platform as a means to address that issue of too many disparate and not inner-connected databases that spring up within the company. For example you might have a separate CRM system, a separate finance system and a separate manufacturing system. Each has key components of information for your company for purposes of managing the company direction you may want to combine. Consolidated data enterprise databases behind an ERP have traditionally been the answer. The reason that made sense is because that simplified managing the stack which is the IT infrastructure behind. You don't have to get into complex databases and managing multiple models and instances of data and correlating them across different disparate databases. There are new technologies out there that simplify that stack and allow you to do that management of multiple disparate databases easier. The one that comes to mind that we are using is Cast Iron Systems as a solution to that problem. One of the others things I find; if you look at my large enterprise apps, those are the apps that most drive costs in my organization. I'm speaking of things like People Soft or JD Edwards. If I can get point solutions that will satisfy a large part of the functionality that I get from those larger applications then I can lower my costs to facilitate that functionality. I'm talking about point solutions for ECO vs. an agile end-to-end product life cycle management enterprise suite. An agile entry level price point for MMPLM is in the high hundreds of thousands of approaching a million dollars to do a decent implementation. But if I can get point solutions that satisfy the management aspects of product development; the R&D aspects such as ECO, I can find applications out there that will satisfy that need in the $30,000 to $50,000 space. If I can overcome the problem of integrating them I'll have a much lower price point infrastructure and I think a solution like Cast Iron can allow us to have the back-end data integration.

OV: How have the various pieces of legislation enacted recently, such as SOX, HIPAA etc, - affected your business and specifically your organization?

Cesario: I see Sarbanes Oxley kind of like Y2K - really, really overblown. Keep in mind the size of my organization. Compliance at the Sarbanes Oxley level means effective right down to the data capture point all the way up through the analytical aspects of managing data you need integrity. Integrity has to come from departmental operating procedures and standard operating procedures that allow for effective controls to maintain high quality data. You can address these concerns through manual or automated controls. We've found in our analysis that we've been fairly thorough with regard to data quality and data quality management. So while we have had a set of work to do with regard to cleanup, most of it's been about effective controls vs. actual changes in how we manage the data. In my view there is very minimal impact to us. I don't see it as an issue.

OV: Have the IT purse strings loosened or is it going to be as bad as the last couple of years? If so, what are the areas that you see yourself investing in terms of IT efforts? What are your thoughts across the board as to what IT would invest in?

Cesario: I think that my experience is that the funds are there if the business value can be articulated. I think that's a political issue and a value statement issue and one needs to be a technologist who can facilitate business functionality in a way that's successful. If you have that then you can get down to the business of partnering with my individual clients so they drive new investment in technology not me, the technologist. That's how we've been able to do that. The money is there for the areas of investment that will bring us value.

There are two groups. There are infrastructure based projects which are internal IT-based projects and external projects that my infrastructure will utilize. On the infrastructure side, strategically we're investing in reusable infrastructure architectures like SANSs and moving away from direct-attack storage, re-architecting to cull out physical tape back-ups in our environment. It's huge levels of efficiency and allows me to allocate resource utilization at the database level, at the storage level, the bandwidth level MIPS required for instance we do R & D research and we run statistical analysis algorithms. If I have an infrastructure for grid computing I have access to a huge amount of MIPS that I can allocate freely throughout the company in a very quick and efficient way. The reusable componentized IT architectures where we are making investments for our IT infrastructure and to that end we have a very detailed plan that is driven by a document that is called the Applications Roadmap which is in essence a manifesto that explains how we use It infrastructure here. We use that in conjunction with the current IT infrastructure which has grown up somewhat organically (which I think is the case in most IT organizations) to give us a plan or path forward as to where we will invest infrastructure wise.

On the specific business function side, this is the part that's driven by the customer and our areas of greatest need right now are in customer relationship management. That's because we really haven't focused on developing an infrastructure for the customer record before now. We're about a 13 year old company.

OV: What specific areas of CRM are you looking at?

Cesario: The customer record itself, the management of the customer record, the ability to get access to that information by anyone in the company from the executive management team all the way through to a rep who's doing cold calling. Consolidation of all the multiple and disparate sources of customer records that we have today. The integration component is the biggest challenge for us.

OV: As the CIO, you deliver IT services to the rest of your organization. Do you have a CIO platform or dashboard that you use to measure your projects and how you're delivering your own services to the rest of the business units?

Cesario: What you're referring to is an infrastructure project that I call business intelligence. We have created a dashboard here at Affymetrix with what we think are the requisite indicators that give us the ongoing operational health of our business in terms of status and reporting with regard sales and all that and also a dashboard about product R & D. We're putting in place infrastructure to be able to capture that data which feeds into a dashboard. Decision-making in the business in an efficient way. This in an area where we are looking at off-the-shelf products like Crystal Reports for reporting and frankly, struggling to find good off-the-shelf solutions at the FPNA side for ad-hoc cube-like analysis of data and information in the company. Again taking an enterprise approach in the concept of that dashboard really is the strategic driver there.

OV: What is the biggest challenge that you, as a CIO, face today and how do you plan to overcome it?

Cesario: Connecting the capabilities of IT to the business and having full adoption is really the largest challenge we have. Investments in facilitating automated functionality within the company can only be taken advantage of by a company that is fully behind any given implementation. That's why frankly you see software companies struggle. You're dealing with internal process change when you come in with a new product or software within the company. In my view, how you deal with process change really dictates your success in the technology spot within a company. That's why we focus on kind of a three pronged approach. The three pronged approach that we take in IT is in three areas of goals for us: partnership, relationship with the customer - it's how well we understand the business need (say if the groups within the company like finance or sales or marketing or R & D) and how well we can partnership with these individuals and create solutions which can give value to the company, to Affymetrix as an organization.

Two is efficiency. We need to create a value proposition and it has to be, you paid this much, you got this set of services and what is that relative to where you can get it elsewhere. That's the question about outsourcing. Now when outsourcing becomes clearly, from an economic standpoint, the right move then you do that and you buy the service.

Lastly is people. In order to bring technology to bear on the problems of Affymetrix as a business we need a very knowledgeable team of individuals who are motivated to execute on partnership and efficiency so we focus on the quality of our IT infrastructure in terms of the people. That's the three-prong approach that I'm taking in managing IT at Affymetrix.

OV: In closing, please feel free to comment on anything that you think would be of interest to the audience of the newsletter. (They include 1) limited partners who come from large corporations and are managing their pension funds and 2) seasoned CEO's and entrepreneurs from portfolio companies).

Cesario: I think highlighting a couple of things that we hit upon. The trend towards standardization of end-to-end process functionality. The view which is going to allow individuals like me to make decisions for outsourcing applications is going to be a key thing. I don't think the application suite in the market itself and the products I see out there and the services I see out there are ready yet. Too, this concept of adaptive componentized IT infrastructure is a key driver in terms of how I make decisions about the infrastructure I put in place. The examples there are reusable bandwidth; reusable, redeployable networking capability and the storage side and MPS side for anaylsis.

OV: Moses, thank you very much for your time today. I'm sure that our readers will enjoy reading about Affymetrix and yourself in the upcoming issue of the newsletter.

About Affymetrix: Affymetrix is dedicated to developing state-of-the-art technology for acquiring, analyzing, and managing complex genetic information for use in biomedical research. The Company began independent operations in Santa Clara, California in 1993. Affymetrix is a market leader in creating breakthrough tools that are driving the genomic revolution. By applying the principles of semiconductor technology to the life sciences, Affymetrix develops and commercializes systems that help scientists alleviate human suffering and improve the quality of life.