Thursday, August 17, 2006

Full Transcript of Dana Gardner's BriefingsDirect Podcast on IT Shared Services

Transcript of BriefingsDirect[TM] podcast with Dana Gardner, recorded July 28, 2006. Podcast sponsor: Hewlett-Packard.

Listen to the podcast here.

Dana Gardner:
Hi, this is Dana Gardner, principal analyst at Interarbor Solutions and you’re listening to a sponsored podcast, BriefingsDirect. Today a discussion about IT shared services. This is a trend that has cropped up in back-office applications and processes over the last few years and is now being increasingly used for IT productivity, within the IT departments at large organizations.

Joining us to discuss this topic, and dive deeply into its implications for such other trends as services oriented architecture (SOA), are two executives from Hewlett-Packard (HP). Joining us are Peg Ofstead, the worldwide solution lead for HP’s IT Shared Services and IT Consolidation Services unit within HP Consulting and Integration. Welcome to the show, Peg.

Peg Ofstead: Hi, thank you.

Gardner:
Also, joining us is Ewald Comhaire. Ewald is the director of the global practice for Next Generation Infrastructure and Technology at HP. Welcome, Ewald.

Ewald Comhaire: Hello.

Gardner: Well, let’s dig right into this subject. Because it’s a bit esoteric for those who are from a variety of backgrounds. Let’s define IT shared services and perhaps also offer a brief history of how this came to be applied generally -- and then perhaps more specifically -- to information technology (IT). Why don’t we start with you, Peg?

Ofstead: Great, thank you. Shared services have been around for 15 to 20 years, but they really were more popular with back-office functions like HR and finance. Some companies also called their mainframe data centers shared services over the years. But now that there’s really a resurgence of shared services, there is also more of a centralization of IT -- in many cases going in to try and save money. One of the problems we see is that often companies have centralized their IT, and they are calling it shared services, but they haven’t moved to what’s traditionally a shared-services model where IT is acting like a supplier to the line of business with much more flexible models.

Gardner: So it’s not just centralizing from a decentralized perspective. It’s really changing the role of IT from responsive to perhaps filling the role of a supplier and service-provider to the larger organization.

Ofstead: Absolutely, the whole idea is that centralization can save money and that’s why a lot of people are doing it. It also provides some standardization and other benefits. It takes control away from the line of business, and that typically is the rub. That’s where the line of business pushes back about centralization. Because when they have their own IT, they obviously can ask it to do whatever they want when they want.

Once you centralize, if you are merely managing to an annual cost budget, as a cost-center you’ll lose a lot of flexibility and responsiveness. The shared services model is trying to get the benefits of centralization, yet using a number of techniques that give a measure of control back to the line of business. If they act as a supplier, where they charge based on usage or subscription for services, they are much more flexible in terms of being able to respond to changes that the business needs.

Gardner: I can testify -- from personal experience at the number of companies that I have worked at over the last 10 or 15 years -- that relating to and getting satisfaction from IT can be a difficult process. Sometimes you get the run-around. Sometimes you get a nameless help desk; sometimes if you are in an application development mode you get put in a line that’s long and takes months for seemingly easy things to get done.

That fosters the notion that people won’t even go to IT, and start doing these sorts of things on their own -- almost on a loose cannon basis, department by department -- all of which is highly inefficient and expensive. How big a problem is the issue about companies getting satisfaction from their IT departments? And how much progress do you think IT shared services is going to bring to this issue?

Ofstead: I think you've put your finger on a major problem especially as IT is centralized. The key here is really two things: if IT is managing to an annual budget that they are handed, and that budget keeps getting cut every year (which has been the case for the past five years or so in many organizations), then basically all the requests from the lines of business get prioritized.

IT looks at their budget, they fund the items at the top of the list and everything else isn’t done. That’s one major problem. So, moving away from a cost-center model addresses that, because basically it comes down to the fact that if the line of business wants something, and if they can pay for it, IT can do it. That becomes a significant change.

The other key is really a cultural change, and a lot of people get caught up in looking at the technology aspects of shared services. Quite frankly it’s all these other things that are major transformations that need to be made in terms of governance, the operating model, culture, and staff. For example, in a shared-services environment the whole thing is about service, customer service. It’s a much more of a focus on treating the line of business as if it were a real customer, and customer satisfaction, as you know, is really, really important. Ewald, Do you have something to add?

Comhaire: A couple of things, but I think you already hit on them more or less. One, we see still quite a few surveys that show that IT has some issues with showing its value to the business. That's typically a clear sign. We also still see companies that choose to outsource their IT, which also means that they are not at the right level of maturity or cost effectiveness that the business would like. And, finally there are also complaints on the transparency of the costs. Why does IT charge me this amount of money and what does that money give me as a value, are still often questions. IT shared services provides a framework or a model to help answer these questions in a better way.

Ofstead: Those were wonderful points. The other thing about transparency is on service levels, which goes somewhat to Dana’s point that if you have specific service levels that you’ve agreed upon with the lines of business, you can measure those. You can be transparent about how well you’re meeting those service levels -- whether they are about help desk responsiveness or availability or response time -- all those things help build trust and show the value of IT.

Gardner: Clearly there are some imperatives internally for how IT can improve and project itself as more of a solutions provider, making customers satisfied internally, rather than being viewed as a bureaucratic cost center. What about external trends? Is there anything going on in the business world vis-à-vis globalization or vis-à-vis the technology itself that makes this a particularly auspicious time to begin embarking on IT shared services? Ewald, why don’t you take this one?

Comhaire: First of all, the service providers -- and HP is a good example of this – are modularizing our services offerings. For example, in the past we used to offer much more all-or-nothing outsourcing. Today, it's a very modular portfolio of what we call "managed services" offerings to our customers. This also makes it obvious that the customers have a lot of choice points, whether they want to run a service inside of their own company or eventually acquire it from the larger market. That’s definitely a trend that forces customers to modularize their offerings and make them more comparative to what the industry can offer.

Gardner: Do you mean that there’s more of an opportunity now for a hybrid approach, in which you can have some services provided internally, some provided externally, some as a service, or some as simply an off-premises technology infrastructure benefit? Do we have a spectrum of choices now in order to approach this problem flexibly?

Comhaire: Exactly, and it’s also easier for the businesses to compare, because in the past it was all-or-nothing. You would outsource all of IT. Today, for example, they can look selectively at one service and see that there’s a capability outside.

Let me give you a practical example. Let’s say in the oil and gas businesses they need to do bid on an oil field. In order to get the price as close as possible to the value of that oil field they need to a reservoir modeling. In the past, obviously they would need to buy all the equipment and all the software, bring it inside their own house, run the whole model and when the money is used up, then the modeling is basically done. Today, they can look at different offerings on the market -- and HP offers some of these -- where they can do the modeling on a per-week or per-day basis. They don’t need to own the equipment and they don’t need to pay for the software for the whole year. They can just rent it for a smaller period of time. So, there are a lot of choices. And as you have choice, you can compare. This drives the modularization of the services offering into shared, modular-type services.

Gardner: Are there any other global or mega trends that are underpinning the value and momentum toward IT shared services?

Ofstead: I think that the cost issues of the past years have forced a lot of companies to look at their models. They've been in a situation where lines of business would have a new application and they would buy servers dedicated to development and testing of that application, and then another set of servers for production. That tended to be a very constrained and ineffective approach. Utilization tended to be low on a lot of those servers. So, with new virtualization capabilities, companies are looking at models where servers are shared and virtualized. The only problem is: What does that do with the model where the line of business is buying the servers for specific application? They are at a point where in order to take advantage of technology, they have to look at different governance model anyway. I think that has led a lot of them to explore these kinds of alternatives.

Comhaire: I would just like to add one more -- the agility angle. Now that we're seeing business up-tick a little bit in the market, the need for greater agility is coming up again. We have been able to show both internally and externally that shared services provides a very agile environment in which the time-to-service is much shorter in a shared services environment.

Gardner: If you have a variety of services from a variety of sources, you can do a cost-benefit analysis on how to best go about them but also you can move them in and out with some speed and flexibility. Peg, you have mentioned governance. Are there any regulatory imperatives -- perhaps Sarbanes-Oxley -- that make shared services something to consider?

Comhaire: There are two ways to look at it. First, they could be an inhibitor. We often see that in the pharmaceuticals industry where they say, "Well, our environment is heavily regulated, so we cannot touch IT. It can never be shared."

Sometimes that is an excuse; other times, there is some validity in it. But often, there not a willingness to go through some of the procedures again, because then they need to re-document some of the procedures to meet regulations. On the other hand, if you take a look at Sarbanes-Oxley, it’s sometimes easier in a shared services environment because actually at the end of the day you have fewer environments. They are more shared and standardized -- and fewer things to therefore make Sarbanes Oxley-compliant at the end of the day. It saves you time and effort, and you can be compliant better. I can give an example: Often you can have some shared services that support some of the financial applications that need to produce Sarbanes-compliant reports. If they’re running on shared services most often, they have a better disaster recovery mechanism behind it, and thus helps you get more quickly compliant with Sarbanes-Oxley regulations.

Gardner: It seems to me that going to a shared services approach also provides the so-called "one-throat-to-choke" benefit. [Under shared services] there is a hierarchy within the IT organization, so you can go and get action [from the top]. The leadership can take the role of triage and prioritization -- and therefore provide satisfaction -- on a high-priority issue … more quickly.

Comhaire: Yeah! And with the clear accountability that shared services offers, this is adding to the point you just made.

Gardner: Okay. So, we’ve looked at some of the issues internally, we’ve looked at some of the major trends that are affecting this. What’s the pay off? How do we get people to make the plunge and say, “Wow! This is going to be disruptive; this is going to cause me to almost have to do a complete audit. And, you know what, I can’t transform what I don’t know.” It’s going to take quite a bit of activity, energy, and probably some cost up-front in order to take on this IT shared services benefits issue. What’s the pay off? What are the incentives that break through inertia against such change?

Ofstead: Well, there’s definitely cost savings. However lots of companies may say that they could make lots of those cost savings through mere centralization -- and that's the question that one constantly has to come back to.

But, the other side is, as Ewald mentioned before, as CIOs struggle to prove the value that IT has for the business, what value do they add other than being a back-office function? It’s proven through surveys that CIOs who have moved to a real shared-services model, where they are treating the line of business as a customer, are moving away from this cost-centric model. They are seen as adding significantly more value to the business. So, that’s a huge win for IT and it removes CIOs from being in this situation of constantly having their budget squeezed and putting them more into a partnership with the business.

Comhaire: I would like to add three others to the ones Peg mentioned. I think she mentioned the value and cost, which are clearly there in almost all the cases. Agility we've already mentioned. We can prove that a lot of customers value that tremendously, as well as quality-of-service improvements, because shared services typically are managed as a true service with end-to-end service management and a service-level agreements. That's a clear benefit to the business users, and sometimes if it’s implemented, it also reduces risk through better security combinations with the service. For example, compliance to regulations or a business continuity element added to the shared services also helps to reduce the business risk in general.

Gardner: There are obviously some qualitative and quantitative ways of viewing this. One would certainly be the IT budget, going up or going down year to year. At the same time, there would be a qualitative view -- what’s the level of satisfaction and productivity from the perspective of the internal constituents? Are you familiar with any company that has undergone such a transformation and has been able to see its budget decreasing and, at the same time, seeing general satisfaction and productivity increasing? Are there some proof points that we can look to here?

Comhair: Unfortunately, I can't mention the name of the customer, but we have started a complete IT shared-services transformation with a customer in Europe. In fact, it was started more than three years ago -- and it's really important that it’s already that long, because the customer could measure the before- and the end-states, and obviously understand the budgetary differences. And, so we do have those numbers, which are pretty spectacular from an operational cost-saving perspective. In the meantime it has also been really good for the overall profitability of the company -- or at least that large unit of the company has done really well in the subsequent years. So, we can see the reflection, not just on the cost side, but also on the overall health of the business in question.

Gardner: You mentioned something important there and that was the operational cost. It’s my understanding that in many organizations it’s not the up-front investment in technology or even acquisition of skills, but it’s in the long-term operational cost -- making up somewhere between 70 and 80 percent of the total IT budget -- that is the major cost center and therefore the place to look for the best return on efficiency. Is that your understanding as well?

Ofstead: Yes, and the other really keen thing -- and Ewald has mentioned this before -- is that this allows people to spend less money keeping the lights on, as it were, in the operational costs, and spend more of their budget on innovation. So, IT can actually do more of the things that the lines of business want it to.

Gardner: So, even if the budget doesn’t go down, there is an opportunity to reduce the routine ongoing costs and then apply those funds to a new initiative that will offer even higher productivity benefits.

Ofstead: Right -- and higher value.

Gardner: Where does this work best first? Are there any particular types of company -- by culture, by organization, perhaps ones that are either more decentralized or centralized -- that work better than others? Are there particular verticals that this makes sense in? Where are you seeing the sweet-spot in the market for where IT shared services is accepted first and then has the biggest return?

Ofstead: Well, we're seeing broad-based interest across the industries. However, in terms of a broad transformation of all of IT to a shared-services model, I’d say that the financial services sector is particularly active in that approach, and also the public sector. However, other industries, such as manufacturing, seem to be very interested in implementing selected shared services.

Gardner: Is that because of low margins -- because they're in a tight and competitive environment?

Ofstead: That could a part of it. There’s a lot of change in financial services right now. The public sector obviously has a lot of pressure in terms of transforming itself to be able to provide information to the public, and there are other changes going on with manufacturing. I see them interested in implementing individual utilities, such as the shared infrastructure utility or shared database utility; Ewald, is that also what you are seeing?

Comhaire: Absolutely, I just would like to add the interest from the telecommunications and service-provider markets and what’s typically driving them. Traditionally, in a telco you had the network operations site, which was very much customer focused with an outside portfolio. Then you had the internal IT, and they were very different, because you had telco protocols and you had enterprise IT protocols and standards, all of which were different.

Today, there’s a significant convergence of these protocols. Mainly the enterprise IT protocols now become the telco ones with some additional flavors, and that means that those CIOs of the internal IT become now very active as well in adopting the more service provider-type of operations that their external colleagues were doing for a long time. So, we also see service-provider market telcos being very, very active in shared services.

Gardner: Let me move to the conversation a little bit to what I mentioned at the outset, the interesting intersection of IT shared services and services oriented architecture, or SOA. It seems to me that this is something that has a potential to be a whole greater than the sum of its parts -- that IT shared services is approaching a productivity issue set from one perspective, and then SOA enters from a more technological perspective. Can you perhaps delve in a little bit and help me understand the harmonious relationship between IT shared services and SOA -- and what this potentially means in terms of the magnitude of the transformation potential?

Comhaire: I would be happy to take this first, and, Peg, if you want to add something, go ahead. First of all, there are a lot of commonalities, but there are also clear differences. IT shared services is more of an operating model, a concept of how IT can provide services to the business. SOA is actually more an architectural approach and there are clear differences from that perspective.

But, there are also many things in common. For example, what’s definitely in common with SOA is a model of a consumer and a producer of a certain service. Obviously, the same applies to shared services where the service in this case is an IT-type service -- not an application but an IT-type service. That’s one place in common. The other area that’s in common and that is maybe a little bit more unique for our approach to shared services is that we have architected a typical IT shared services portfolio that the IT organization would offer to the business using SOA principles.

What does that mean? That means that services can be reused together, or architected in such a way that one shared service can be applied next to another -- to solve a bigger problem. That is something that not necessarily IT shared services mandates as a concept but it is something that, if you do a clever combination of both IT shared services and SOA you can get a lot of business value.

And the third location is as we implement individual shared services we also try to expose that functionality under a web services format, which means that again the service would be reusable at a higher level in a transparent way, which is another SOA concept applied to shared services.

Gardner: So, do you view this that SOA provides a new set of tools that can enable IT shared services? And then provides this larger benefit from productivity and agility? Is that how you see it?

Comhaire: Yes, absolutely. SOA is an enabling concept and enabling architecture for shared services, and if you apply those concepts of SOA to IT shared services you can get a lot of additional benefit over and above what the normal shared services model already would offer to you.

Gardner: Is it too much to ask of any organization to try to do these both at the same time? It seems that if you could do them simultaneously there would be a huge payoff, but it also sounds almost daunting.

Comhaire: That is absolutely true, and that’s why we are definitely helping the customer by doing some of that work for them. For example, we have our own internal templates -- starting points -- of an IT shared services catalog, and the way we architected those services is already SOA-compliant. So, in a way it may be too much to ask of a customer who doesn’t do this as a business all the time. Typically those customer do this once, but for us as a provider of IT shared services it’s kind of a must-do, and we need to spend just a little bit more time as we architect and build these shared services to make sure they are SOA-compliant.

Gardner: And, Peg, it seems to me that if you were to take two similar organizations -- perhaps operating in similar markets, of similar size, with similar resources -- and one takes the plunge and does SOA and IT shared services in some combination over a period of time -- and the other doesn’t. Well, the company that does would have a huge competitive advantage and a productivity advantage. Is that grasping at straws?

Ofstead: No, absolutely not. Again, the financial services market is very competitive at the moment and I think that’s one reason we’re seeing so much motion in that area. Obviously, we've talked about a cost advantage and an agility advantage, and the combination of two really could make a huge difference for a company.

Gardner: Do you have anything else to offer in terms of how companies should conceptualize of SOA and IT shared services as sort of top-down? That is, would the “C”-class executive work down from the CIO with an IT shared-services mandate? And is SOA perhaps more organic and needs to happen down in the developer ranks and really change the way in which the development takes place? Can we simplify it to that level, that one is sort of bottom-up and the other top-down?

Ofstead: Well, the concepts are extremely synergistic, and I think you want to do a good job of explaining that. But, both transformations are not small. There is a lot involved and so an organization would really need a strong commitment from the top to embark on these. They would need to understand that the full implementation probably takes a number of years.

So, a clear road map with specific projects and short-term demonstrable results, I think, is important. Having a clear vision of where you’ll be after nine to 12 months, where you’ll be after the second year or the third year, and understanding that this is not a small change that you are embarking on – these are key things for either one of these transformations.

Comhaire: But, just to add to what Peg already said, in our approach -- and that’s actually true for SOA as well -- we do offer top-down and bottom-up approaches. For example, a bottom-up approach in shared services would be when we help customers build out just one shared service. And, we try to do it well, guide them through the whole concept of what it takes to build out a shared service that reduces the scope to one.

In fact, you can do this for SOA as well. You can apply it to one application or a small set of applications that you redesign using SOA concepts. At the same time, you can have the strategic approach top-down to say, "I have decided to move to a shared services model,” or “I’m going to transform my whole application portfolio over the period of time to all SOA-compliant applications." For both, you can walk this continuum from totally strategic to a more tactical, and play-it-small, try-it-first, and then see which works and extend it to the more strategic direction.

Gardner: Well that brings up an interesting question. What do you see happening in the field? How do people do this crawl, walk, run approach? What do you recommend and perhaps what is HP doing in terms of putting some structure around that process?

Ofstead: Well, as Ewald said, we have customers who prefer a more broad-based transformation, and we have other customers who prefer to implement selected shared-service utilities one at a time. We have a methodology and process around both.

We’re helping customers with both of those approaches. Our methodologies are consistent. So, if you start, for example, from the bottom-up, as Ewald was saying, implementing one or more selected shared service utilities -- for example assured infrastructure utility or a shared database utility -- you can ultimately go through the transformation. But that’s a decision the organization will make, based on benefits they saw of implementing just selected utility.

On the other hand, for companies that have really looked at shared services carefully and have decided to take the plunge and do a broad-based strategic transformation, we have a reference architecture and the tools and methodology to help them with that process. We have a clear vision of exactly what’s needed, and we can help them assess where they are today, set their goal of how much of shared services they want to implement, in terms of the level of maturity they aspire to, and give them a clear road map on how to get there.

Gardner: Are there any lessons from other folks that have undertaken this sort of activity in terms of how to get started and not necessarily implement the shared services, but how to get yourself ready to begin implementing? Are there any first steps in preparation that you‘ve seen?

Ofstead: In working with our clients and also at the Shared Services Conference that was held in the U.S. this year, I heard all of the people who have embarked on this road say that it’s very important to have an external partner to help guide you through this -- a change agent. That was consistent among all the presentations: You need someone to help guide you through this kind of a change. Obviously, a key thing is helping people understand why you’re making the change, and that’s your people within the IT organization as well as in the lines of business.

Comhaire: I think you said something that’s really valid: It’s a people, process, technology change. So, that’s definitely a key. Depending on which level we talk to, some of our customers see this just as a technology problem, "How do we move multiple workloads on a shared platform?"

They forget about the fact there is a whole impact on process and, as Peg mentioned, a management of change on the people, on the staffing side that they need to take into account to guarantee success. That’s a lesson we have definitely learned: We have to make sure that we correctly value the interdependencies between the people, the process, and the technology. And even if you get those three right, at the end of the day it’s also about designing and building your services. Does your service meet the requirements of your business consumers? Is it well aligned with where your company wants to go? So, there are some additional elements besides just people, process, and technology that are also key elements for doing it right.

Ofstead: And long-term, one of the key things is that customers need to start is to begin benchmarking themselves. That means really comparing the cost of their service and the quality of their service to what’s available from external service providers. This means actually implementing continuous process improvements around their shared services. So, it’s very important that people understand all these things and they not just view the shared-pool of technology that they are using.

Gardner: Well, we are almost out of time. I think this has been a fascinating discussion. Let’s look to the future briefly. What are the implications for IT shared services, particularly in concert with SOA? It seems to me that we have an opportunity for a marketplace effect on a number of different levels. IT organizations that have this agility can shop around, can swap sourcing options in and out, based on the performance and payoff. And then, in a sense this sets companies and enterprises up for an even larger set of productivity goals, when they can bring competition in marketplace forces into IT -- how they source it, how they deploy it themselves internally, and also in providing a much more market-oriented approach to the internal customer. And, as we mentioned the public sector -- government for example -- how they actually provide services to their constituencies, or voters, or taxpayers. Can you elaborate a little bit on what is the potential, the future road map is for all of this?

Ofstead: I think for the CIO who does this properly and just doesn’t centralize, but actually changes the model from a cost-centric model to a service-provider model, it’s what makes them an entrepreneur within their organization.

Instead of being constrained by their annual budget, they’re able to really work as a partner with the line of business. It certainly puts them in a position where they can help be an innovator with the line of business, which is huge. There’s the cost savings we talked about, and the agility.

For companies that do this properly not only will the IT organization be seen as very successful, but the organization as a whole will too. They would be much more competitive, much more successful. It could make a tremendous difference in the success of a corporation. Ewald?

Comhaire: Yes, and to add to some of the things you’ve already mentioned, I think it will also bring IT closer to the business; because within IT you start to think like a business. “What do I offer? How do I price, not just on what costs I have [but on] planning for my demand and my supply?”

So, a whole series of notions that typically only business units within a company would think about -- now have IT thinking in same terms as the business. This typically will bring them much closer together because they start talking the same language, and so on.

And another aspect I think for the future is that IT organizations will become a little bit more like service brokers. I think you described this really well. That means they have to in-source, outsource, and combine these things much more dynamically than today -- and I think that will change the IT landscape over the next years.

Ofstead: To build on that, one kind of hidden benefit that a CIO would have is really true visibility into his cost, and an organization that’s much more service-centric than technology-centric.

Today, many IT organizations are organized around technology. They may know what technologies cost. They may know what that they spend on servers every year, for example -- they may not even know that. But they certainly don’t know what it costs to provide a service. For example, what does it cost for each SAP transaction?

Once they move to this new model, there’s lot of benefits in terms of knowing what their true costs are, knowing the quality of their services, really understanding that well, and being able to really make very sound business decisions about what they in-source and what they either outsource or out-task.

Gardner: Well great. Thanks very much. I think we need to wrap this up. It’s been a discussion about IT shared services, how it’s developing as a productivity and agility benefit and as a business alignment and transformation benefit. It really does encompass a lot in the modern corporation and enterprise.

Joining us for this discussion, have been two Hewlett-Packard executives: Peg Ofstead, the worldwide solution lead for HP’s IT Shared Services and IT Consulting Services organization, as well as Ewald Comhaire, the director of the global practice for Next Generation Infrastructure and Technology within HP services. Thank you very much both for joining us.

Listen to the podcast here.

Podcast sponsor: Hewlett-Packard.

Transcript of Dana Gardner’s BriefingsDirect podcast on IT shared services. Copyright Interarbor Solutions, LLC, 2005-2006. All rights reserved.