What is outsourcing?
Outsourcing desktop management
TCO and the desktop management problem
What is TCO?
Why measure TCO?
How to reduce TCO
From cost to risk
Management tools for TCO reduction
Identifying the focus of desktop management
Capital and revenue funding
Making seat management work
Keywords for this document: ASSET MANAGEMENT, CONFIGURATION MANAGEMENT, DESKTOP MANAGEMENT, DMI, ELECTRONIC SOFTWARE DISTRIBUTION, HELPDESK, NETWORK MANAGEMENT, OUTSOURCING, RISK ASSESSMENT, SEAT MANAGEMENT, SERVICE LEVEL AGREEMENT, SLA, STANDARDS, TCO, THIN CLIENT, TOTAL COST OF OWNERSHIP
Total Cost of Ownership (TCO) is a concept pioneered by the Gartner Group as a means of assessing the true cost of corporate desktop systems. Initially, TCO was often criticised for being wildly inaccurate and grossly inflated. The criticism could be justified by referring to the simplistic and potentially misleading basis of Gartner’s early methodology. Recent methodological enhancements have improved the situation but, in any case, Gartner’s TCO figures have always provided two important insights – first, that the purchase price of PC systems bears very little relation to their significance within enterprises; and, second, that ownership itself involves on-going costs.
In the attempt to reduce TCO (estimated at around $10,000 per PC today), Gartner has pointed out that the greatest problem is complexity and that management complexity is more significant than technological complexity. A lot of emphasis has been placed on simplifying the technology of desktop systems (for example, through the use of thin clients) and on using technology itself to reduce the costs of technology (for example, through the use of management tools). The big problems with these technology driven approaches are that they may themselves be costly, they risk increasing complexity rather than reducing it (due to the proliferation of management tools), and that, in any case, they miss the point that management itself is by far and away the most significant source of runaway costs. Even so, technology-driven approaches to TCO reduction can cut costs by as much as 15 per cent according to reliable estimates.
The US Federal government has pioneered an approach to reducing TCO called ‘seat management’ in which the purchasing, installation, maintenance, and management of desktop systems along with on-going end-user support and system management are all outsourced at a fixed rate per seat (or end-user system). Some remarkable reductions in TCO have been claimed, but we do not really have any reliable data. What can be said is that the transition from managing technology to managing a service, and the use of carefully formulated service level agreements, can produce significant benefits in terms of service delivery. Seat management – like most other forms of outsourcing – should perhaps be considered primarily as an approach to risk-reduced service enhancement. In fact, only certain risks are reduced (typically, the ones associated with investment in hardware) and the wished-for improvements in service delivery are dependent on negotiation, contract, and the existence of a good working relationship between service providers and users.
Recent Guidelines have looked at issues connected with in-house desktop management at some length (see, for example, Guidelines XXX, XXX and XXX on Asset Management, Configuration Management and Thin Clients, respectively). These Guidelines have covered total cost of ownership (TCO), desktop management technologies and standards, and administrative procedures, but they fell short of discussing the outsourcing option.
In fact, outsourcing desktop management (so-called 'seat management') is becoming an increasingly attractive option as the costs and complexity of supporting, maintaining and upgrading end-user PCs and workstations have rocketed. The leading lights of this development are the US Federal authorities whose desktop management requirements are immense.
This Guideline addresses the whole area of TCO – it justification and usefulness - and introduces the idea of seat management and something of its history, meaning, practice and probable future in both public and private sectors. It aims, among other things, to provide a set of principles for the implementation of seat management within a general strategy for IS outsourcing.
What is outsourcing?
It may seem a little late in the day to ask what outsourcing is, but the question is actually more relevant than ever now that 'business virtualisation' (to coin a phrase) is becoming a realistic commercial and operational strategy. There are a variety of different forms of outsourcing, ranging from simple contracting out to complex transfers of management functions, personnel, and corporate assets, but the traditional view is that outsourcing involves a change in the ownership and proximate control of costly resources. Yet it is impossible to draw a solid line between, at one extreme, the transfer of responsibility for managing business functions that remain otherwise unaffected and, at the other, the wholesale removal of a business function along with associated personnel and systems to an external supplier.
The definition of outsourcing offered by the troubled North American Space Agency (NASA) may be considered typical of traditional thinking. It focuses on the rationale for outsourcing - the transfer of risk and responsibility with the aim of enhancing business performance - and implies that the defining characteristic of outsourcing is the ownership of assets. 'Outsourcing,' according to NASA's definition, 'is a strategic business concept which is undertaken to transfer a function to the commercial sector which has previously been performed by NASA or its contractors primarily through reliance on the Agency's internal assets. The transference of risk and responsibility to the commercial sector, and the possible divestiture of Agency assets, are primary goals intended to ... improve the performance of any outsourced functions and ... to enhance the performance of the Agency's core functions.' (1)
Unfortunately, as NASA's recent history amply demonstrates, there are some risks that just can't be transferred.
In fact, there are no clear boundaries between outsourcing and other forms of third party provisioning such as equipment leasing, labour contracting, interim management or consultancy, and - indeed - the currently vaunted system of 'public-private partnership' - all of which may be considered approaches to the 'transference of risk and responsibility to the commercial sector'. The idea - implicit in NASA's definition - that there are core activities or resources that cannot be outsourced without destroying the company is coming into question. Indeed, NASA itself is a timely example of an organization whose undoubted core activities (such as landing spacecraft on Mars) may actually be better managed by an external supplier. An organization is neither a collection of physical things - buildings, desks, computers, people - nor a set of activities - making, selling, marketing, paying - but a group of relationships whose precise expression can vary from time to time, place to place and organization to organization.
Over the years, and in a variety of contexts, outsourcing has claimed payroll, mainframe operations as a whole, maintenance, staff development and recruitment, building services, procurement, design and print, PR and marketing, telecommunications operations, and network services. All of these are seen as activities which are either peripheral or problematic - or both. In practice, outsourcing has been seen as a nostrum for organizations suffering from runaway costs, managerial incompetence, or lack of expertise.
The idea that businesses outsource in order to focus on 'core functions' underscores the view that there are areas that suit the skills and accumulated experience of the organization concerned - areas with which it is familiar and tasks which it knows how to handle. Typical targets for outsourcing have therefore been activities which are either so banal that anybody can do them or, contrariwise, are complex or specialised enough to lie outside an organization's particular talents.
Outsourcing desktop management
Desktop management - by which we mean the purchase, configuration, and administration of end-user systems - fits into a third category and is, accordingly, a contentious target for outsourcing. Although effectively universal within organizations - like telephony or payroll - desktop management is not a routine or banal activity. On the contrary, the deployment of end-user systems is often seen as precisely characteristic of the organization in which they are deployed. That really sums up the problem: as organizations are increasingly dependent on the activities of 'PC-enabled knowledge workers', end-user systems can no longer be assumed to be part of the plumbing. They are strategic business tools. And yet the management of these tools is not itself a 'core function' - the core function is their use.
Desktop management may not be a core business function, but it certainly presents a set of strategic issues - from cost control to the empowerment of individuals, from the management of data in a distributed network to the preservation of authority at an enterprise level. The cult of 'total cost of ownership' has proved invaluable in demonstrating and sizing the hidden mass of the end-user iceberg; and, although TCO figures are of doubtful accuracy, they are a vital corrective to the blithe irresponsibility of the 'petty-cash PC purchasers'. In the corporate environment, there is really no such thing as a standalone PC; it always exists within the network of relationships that go to make up an organization.
Outsourcing desktop management is increasingly seen as a viable solution to the problems posed by the PC explosion. Seat management, as it is called, seems to offer a magic bullet solution to the problems associated with end-user computing:
It removes the financial risks associated with investment in rapidly changing technologies
It fixes the cost of end-user support through a well-formulated and enforceable service level agreement
It enables IS departments to regain their traditional authority by transferring their focus from technology to service delivery
TCO and the desktop management problem
Far and away the most popular reasons for outsourcing in general are reducing costs and increasing manageability. In seat management, the two are often seen as effectively identical, given that the purchase price of most desktop systems is negligible (and falling) compared to ongoing management costs. The realisation that associated costs were outstripping the acquisition costs of corporate PCs led the Gartner Group to develop a methodology to measure ‘total cost of ownership’. Last year, the group calculated the average TCO for a Windows 95/98 PC at $10,000.
Gartner’s original 1987 methodology was extremely flawed since it only assessed TCO for organizations using one platform and therefore avoided most of the real world issues of desktop management. More recently, Gartner has introduced TCO Analyst, developed with another US company, Interpose, to address some of the issues raised by the original flawed methodology:
The need to cover a range of client platforms, including – at least - network computers and Windows terminals, Windows 3.1, 9x and NT workstations, and portables (both Windows laptops and Windows CE or EPOC palmtops). This last category is increasingly important as growth patterns indicate that within two or three years more than a third of PC users within most enterprises will connect to the corporate network on a regular basis using dial-up from a laptop, palmtop or other portable device.
The need to cover a range of server platforms, including PC systems running Windows 9x and or NT server, Novell NetWare (in two or three versions), various systems running different flavours of UNIX, IBM AS/400 machines running OS/400, and mainframes. Gartner’s original methodology covered only NetWare LANS.
The need to cover the fast developing markets for low-cost and cost-saving products, particularly regarding efficient power-consumption devices, cheap network technologies, and the use of the Internet for WAN applications.
The need to weight system costs in order to more accurately measure costs of downtime or skills development, for example. Weighting actually takes some clear management factors into account – for example, the type of tasks being undertaken by a desktop machine, the status of the person using it, whether best practice principles are being followed, and how much redundancy has been built-in to the systems. TCO Analyst, for example, takes the following into account: adherence to technical and management standards, the organization of the IT function, the approach to training, and the use of change management and asset management techniques.
It’s clear that TCO is being transformed from a simplistic measure of ‘hidden costs’ and is well on its way to becoming the most widely accepted method of measuring manageability. As such, the criticism that TCO figures are often inflated may be misdirected. A more valid criticism may be that, while cost worked for the measurement of technological complexity, it is the wrong metric for manageability – one might, for example, get more useful information by measuring the time spent on maintaining the usability of desktop systems. That said, the price mechanism provides a useful, if far from perfect, way of identifying the value associated with time or other resources, as long as we do not confuse the resulting figures with actual payments of any kind.
What is TCO?
As a financial estimate, TCO only makes sense if it is understood as the estimated cost of a PC in terms of its impact on the whole organization. Thus, the components of TCO, viewed in this way, are:
Capital costs (hardware and software)
Technical support (installation and configuration)
Administrative costs (procurement, deployment, and decommissioning)
Operational costs (training, self-learning time, data management, peer support, and time spent on user solvable problems)
The last of these can represent almost half of TCO, depending on the exact type and configuration of the PCs involved and the general level of user competence. Its significance has led some observers to argue that end-user computing may actually result in a measurable loss of productivity (2).
Technical developments promise a reduction in TCO by the introduction of 'zero administration' systems, network management tools, operating system improvements, self-managing processors, and 'wired for management' technologies. All these approaches are targeted at reducing the complexity of technology management, yet they run the risk of actually increasing organizational complexity and undoubtedly increase the complexity of the technology itself.
Why measure TCO?
There are a number of reasons apart from obvious financial ones for measuring and - more importantly – understanding TCO:
1. Increasing awareness. General managers and end-users (and not a few IS managers, too) are frequently unaware of the hidden costs of their desktop systems and applications, and assume them to be negligible. TCO measurements can help to stimulate efficiency based on knowledge.
2. Avoiding error. There is often a widespread assumption that desktop PCs and package software are effectively free. TCO measurement can bring out the truth about the cost of packaged solutions, helping management to avoid making serious mistakes.
3. Assessing the options. TCO measurement can help managers make realistic comparisons between different solution-sourcing options – from in-house development to application service provision.
4. Opportunity costing. By understanding the nature of TCO elements, managers can make better decisions about expenditure in situations where a number of projects or proposed activities are chasing limited funds.
5. Calculating ROI and payback periods. TCO measurements allow managers to make more realistic assessments of ROI and payback periods for different investment options, so avoiding the misallocation of funds and helping planning over time.
How to reduce TCO
The key to TCO reduction is widely considered to be simplification One solution has been the introduction of thin clients. These are technologically simplified devices - in effect, terminals not standalone PCs. Vendor marketing messages suggest that thin clients can cut TCO by as much as a half or a third. Recent announcements from a number of PC suppliers revealing that the stripped-down, so-called ‘EasyPC’ is about to make a come-back tell us that the persistent dream of a ‘white box’ desktop device is still vigorous.
The impact of technological simplification can be gauged from Gartner's estimates of help desk costs. According to the 1998 figures, the help desk cost per user of a 'high complexity' system is $1,713 compared to a per user cost of $183 for 'low complexity' devices (3). That is a staggering saving, yet technological simplification itself is less important than the administrative simplification facilitated by thin clients or EasyPCs.
According to Gartner, 'management complexity' accounts for 75 per cent of TCO, while 'infrastructure complexity' accounts for only 25 per cent. Management complexity covers business processes, general office procedures, purchasing arrangements, and issues touching on corporate culture as well as the more technology specific issues such as help-desk administration. With this in mind, reducing TCO requires a combination of technological simplification and best practice management, with by far the most significant savings to be made in management:
Standardise on operating systems and applications
Rationalise and synchronise upgrade cycles
Remove obsolete hardware and software (limit number of generations)
Implement a scalable and flexible architecture
Reduce hardware models and variations
Plan-in fault tolerance
Automate configuration management
Use electronic software distribution (ESD) and metering
Automate asset management (particularly through the deployment of standards-based technologies such as the Desktop Management Interface)
Automate network management (particularly through the deployment of standards-based technologies and/or an enterprise management framework)
Centralise helpdesk administration and knowledge base
Simplify and control technology change
Restrict user access to configuration tools
Restrict user access to Internet sites
Restrict user access to critical software and hardware
Implement a disaster recovery/business continuity plan
Implement asset management, configuration management and helpdesk around a shared knowledge base
Invest in IT research and planning
Increase support staff to end-user ratio
Centralise and rationalise procurement and purchasing procedures
Fund upfront investment in the right systems
Invest in end-user training on systems and applications
Monitor utilisation of systems and applications and synchronise upgrade cycles with useful life
Consult end-users regarding the usability and utility of systems and applications
Create a culture of technological competence within a context of centralised control
Together, all of these represent a complex of financial, technological and human issues that are all related. For example, it is necessary to implement asset management before it can be automated, and it is necessary to finance asset management before it can be implemented.
A simplified model of TCO reduction will probably achieve a great deal, and there are plenty of advocates of one or two club approaches. Perhaps the two most popular methods are:
Desktop policy management – centralised control of end-user system configuration and user profiles
Electronic software distribution – centralised control of the delivery and installation of software
These two may both afford an element of desktop ‘lock-down’. Policy management tools enforce restrictions on users, while some ESD tools support ‘desired state’ configuration management, although most of them simply automate the processes of software distribution. Desktop policy management is associated with the largest savings – as much as ten or 15 per cent of TCO - while ESD is the most obvious form of TCO-reduction. The critical savings made with both these approaches are in terms of reduced dispatch calls through centralised control. The concept may be extended to include the use of so-called ‘healing tools’ in which centralised software is used to diagnose client systems and, if necessary, repair critical system files (using stored software images).
There is no doubt that controlling TCO can save money - or rather there is no doubt that controlling TCO can reduce budgetary expenditure within the IS area. Some of Gartner's figures may be hotly disputed, but the overall message is surely uncontentious: planning, investment, rationalisation, and automation have consistently proved to bring significant cost advantages in all areas of business. The real question is how you fund the cost cutting measures in the first place.
From cost to risk
The problem can be seen as twofold: calculating the amount required and finding the money. Within any organization, finding money is typically a matter of negotiating budgets. Negotiation skills can be given an edge by recourse to the basic TCO argument - money found today to finance a TCO reduction package will be more than amply repaid from savings. Yet it remains an uncertain argument, because it depends on the exactitude of the calculations. These are not only notoriously ill-defined (when exactly does a system or network become less or more complex?), but susceptible to unpredictable market effects. For example, Microsoft's failure to deliver new versions of Windows when promised is now almost legendary. It could almost be written into cost of ownership calculations, except that Microsoft's timing is as unpredictable as its publicity is upbeat. In other words, there is a significant element of risk in the process of calculation itself.
As a result, cost reduction in itself is no longer the most significant factor within the whole TCO debate. The focus has definitively moved towards risk reduction instead. The very fact that such a large proportion of TCO is due to management complexity justifies this development. Reducing the cost of ownership is undeniably a good idea and the right technological approach may save as much as 15 per cent overall according to best estimates, but the important issues around TCO (management complexity and risk) are not directly susceptible to technological solutions - one might even argue that attempts to introduce technology into these areas will only have the effect of increasing TCO. Significant inroads into TCO reduction will only be achieved if organizations are prepared to tackle the issues of management simplification and risk reduction.
The conclusion increasingly being reached by some very large organizations is that the task is ripe for outsourcing, effectively in order to replace technology management (an inherently risky business) with service level management (which can be relatively easily formalised). The outsourcers themselves have taken the view that TCO measures can be implemented in order to predict and control IS costs at the desktop level. For this to work, they must be given a task which is already well-defined and an environment which is already well-organized, and they must understand both the business they are serving and the likely course of development for it and the technology it uses.
It is perhaps no surprise that the first furrows in this field have been ploughed by Government departments and large technology vendors - notably, the US Federal authorities and companies like Lockheed Martin whose declared aim is to deliver 'fixed unit rate seat management services' (4). For such companies, the trick is to bring the unit rate down to below TCO and this trick is easiest where TCO is greatest - that is, in the largest and most complex of organizations.
Management tools for TCO reduction
Table 1 is based on a report by PC Weeks Labs, and indicates how technological approaches to TCO reduction are being taken up. In particular, it reveals the amount of attention being paid to the development of management tools at every level, from the individual desktop to the collective enterprise.
Despite providing a useful snapshot of the state of TCO reduction, some of this table is already out of date and some of it may prove to be rather more of a wish list than a forecast. These are among the reasons why simple technological approaches to TCO reduction are never completely satisfactory. In any case, it would be wise to remember that ‘a fool with a tool is just a faster fool’.
TCO: How low has it gone?
Standards in place for determining PC assets and managing them remotely.
Network interface cards support network boot.
Software distribution packages can push operating system and apps to new PCs.
Applications provide simple policy-based controls for the Windows 9x user interface.
Better support for managing PCs from multiple vendors from a single application.
Network boot support in notebook PCs.
Policy-based distribution of software and control of the Windows user interface.
Better automation of management tasks such as backup of failing SMART drives.
Applications quickly point out problems and some make automatic corrections to solve them.
Very little integration between products.
Web-based interfaces and better reporting capabilities are starting to be added.
Feature overlap as products gain scope.
Web-based management interfaces.
More crossover from NetWare and Unix to Windows NT.
Products require dedicated staff and equipment, and firm corporate commitment, to get off the ground.
The framework is only the beginning -third-party modules are the key to fully exploiting systems.
Some integration between competing management framework products.
More mature integration of third parties management frameworks, with products sharing topology and event information and seamlessly integrating into the management interface.
Web-based interfaces, not only for viewing reports and network maps, but also for implementing management policies.
Table 1: Developments in TCO reduction (source: PC Weeks Labs, 1998)
Nevertheless, management tools can bring significant savings, but at a price. A typical medium-sized installation in the US with an IS establishment of more than 100 and some 2,000 PCs spread over several remote sites predicts that it will save nearly two-thirds of its $2.6 million enterprise management costs by implementing a management tool (in this case CA Unicenter across an ATM backbone). Even without the cost of deploying the backbone, this solution would not show bottom-line savings for a year, and for larger enterprises, the initial investment would be even higher.
All-in-one management solutions such as Unicenter, Hewlett Packard’s OpenView, IBM’s Tivoli TME 10, or Platinum’s ProVision may be an unnecessary expense. They are particularly useful for improving remote management, electronic software distribution (ESD), and end-user support, but they demand careful attention to their own management, in terms of getting the right staff and putting the right procedures in place. Cheaper tools, such as Novell’s ZenWorks, or free ones, such as Microsoft’s Zero Administration Kit (included in Windows 2000 and an add-on for NT), may not incorporate all the functionality of Unicenter, but they are probably easier to deploy and – crucially – represent a significantly smaller financial risk in the event of mistakes.
Identifying the focus of desktop management
The biggest problem here is likely to be the proliferation of tools and the attendant complexity that it engenders. No single tool provides a total solution for desktop management. Unless the IS department has a comprehensive, inclusive and effective strategy for desktop management, there is every possibility that a surfeit of tools will produce exactly the same symptoms they are intended to cure.
In fact, the major error made in implementing desktop management solutions is lack of focus. Although we have concentrated thus far on TCO reduction, there are three widely acknowledged reasons for buying-in a desktop management solution:
1. TCO reduction
2. Reducing the risk of catastrophic failure
3. Improving quality of service
The problem is that too many people try to achieve all of these at the same time. This is to court failure. If your objective is TCO reduction, you must focus on TCO reduction. The other goals can be pursued once you have achieved the first.
Seat management is the obvious way of avoiding the expense and problems of implementing your own desktop management solution. Seat management is the long-term outsourcing of the provision of end-user computing facilities, and may include procurement, installation, configuration, maintenance, helpdesk and other support services for desktop systems, servers, LANs and other communications facilities. A seat management contract typically specifies what services will be delivered at a fixed price per seat or end-user workstation.
Under the US Federal government's Outsourcing Desktop Initiative (ODIN - a so-called 'Government Wide Acquisition Contract' or GWAS), the customer defines the required computer and communications capabilities and purchases a particular bundle of hardware, software and communications equipment for each 'seat.' The price for each seat is fixed according to its type. Each location or business unit orders exclusively from one vendor, ordering for a period of up to three years of deliveries. 'The period of performance for each fixed-price, Indefinite Delivery, Indefinite Quantity contract with each ODIN vendor is nine years. It is an entirely flexible option for government, allowing for mixed and distributed computing environments. It is customizable for special needs, and provides a single point of contact and accountability.' (5)
The purpose of seat management, according to NASA, is 'to deliver cost effective services to meet NASA's mission and program needs using commercial practices and allow the Agency to share risks and rewards with the private sector. This will allow NASA civil servants to focus on the Agency's core mission. ODIN also will allow NASA to better account for the funds it spends on local computing products and services.' (6)
A cynical view is that this can be summed up in six words: 'control and effective monitoring of expenditure'.
To date, NASA has awarded two seat management contracts under ODIN, but it is notoriously difficult to find out whether they have had any significant impact on TCO, partly because of NASA’s lack of openness and partly because no-one is quite sure what was included in each contract – hardware, furniture, software, services, and telecommunications were all available but not always included. However, probably ill-informed estimates of ODIN TCO came up with the ridiculously low figure of $99 per month per seat, while a similar Federal project was said to have produced an annual TCO of some $3,400 per seat. (7)
Capital and revenue funding
From an accountancy point-of-view, the biggest benefit of seat management may be the fact that – like the defunct Private Finance Initiative and its successor Public-Private Partnership in the UK - it turns capital into revenue expenditure. As with the transition from managing technology to managing services, this has a profound effect on the way things are (and can be) done. Typically, it is easier to raise new money for capital expenditure, but revenue expenditure may be funded out of existing budgets and is certainly more flexible. Revenue funding removes financial risk, but replaces it with ‘service delivery risk’. If properly implemented, it should allow more accurate budgeting over the period of the contract, although it is critical to be precise about what is and is not to be delivered within the budget.
NASA is clear that seat management does not involve capital expenditure: 'Like other types of outsourcing, what the ODIN contract provides is essentially a utility like the phone service. You determine your provider and your service level, and the rest is provided. It’s bundled into a monthly price per seat, which is pre-determined and predictable. The assets are provided to you based on the requirements of your determined service levels. Federal Data Corporation is sensitive to agencies with existing asset inventories and recognizes there are assets with residual value and we are prepared as your partner to develop a customized solution taking those assets into consideration.' (8)
This means that seat management would be funded as revenue expenditure, however that is managed within an individual organization. In particular, seat management could involve incremental funding - particularly where growth targets involve the addition of new seats.
In terms of the balance sheet, such an approach shows up exclusively as a liability, although it is entirely possible to opt for a ‘hybrid’ or partial version of seat management, in which assets remain in the hands of the user organization while services are provided under contract by the outsource organization. This arrangement has the advantage of spreading risks, and may be easier to finance from new money, although it makes administration more complex.
Making seat management work
The single most important factor in making seat management work is to define objectives clearly. This is more precise than fixing the focus of desktop management, as discussed above, but should be seen within the context of corporate objectives for desktop management and for outsourcing. Reducing TCO may not be the ultimate goal of a seat management deal – although it is likely to be a major objective for both desktop management in particular and outsourcing in general. If reducing TCO is your objective, then failure to do so is simply failure.
Seat management usually has a positive impact on TCO through:
Identifying and eliminating hidden costs
Facilitating continuous improvement at a fixed cost
Unifying and rationalising service delivery
Creating economies of scale
But seat management can serve a number of different specific purposes, notably:
Simplification of the end-user environment
Freeing effort for self-defined core activities
Centralising and improving user support
Centralising asset management
Simplifying and centralising software delivery and upgrades, network administration, and configuration and LAN management
These may all correspond to corporate objectives to a different extent and any practical seat management strategy should be dictated by considerations of ‘best fit’. That said, there is one precondition for successful seat management agreed by practically all writers on the topic: at the outset, there must be a comprehensive inventory of hardware and software, identifying individual items and their physical locations.
As we’ve observed before in earlier Guidelines, such an inventory is at the heart of the databases which enable configuration management, network management, asset management, business continuity planning, and helpdesk support. It is absolutely critical. In seat management, it serves an extra purpose – the identification of systems ‘as-built’ to assist accountability procedures and the eventual return of systems.
Those US Federal agencies with experience of seat management also suggest that, even before an inventory can be undertaken, it is vital to identify the aims and objectives of the seat management strategy. If cost reduction is among them, then clearly there must be some agreed procedure for calculating TCO.
Before writing a service level agreement (SLA) for a seat management provider, the user organization needs to identify required service levels and acceptable costs of ownership at the level of individual desktop systems. These will vary with job function. Some systems (for example, graphics workstations) are required to deliver higher performance than others; some (for example, workgroup PCs running financial applications) have more stringent limits on downtime than others; some (for example, PCs connected to the Internet) need higher availability levels than others; and so on.
All this can (and should) be undertaken before the transition from in-house to outsourced desktop management is completed. But the processes of prioritising objectives, producing inventories, and writing SLAs can have significant benefits even if no actual outsourcing ever takes place. In other words, it is entirely possible for an IS department to provide its own structured seat management service, should the resources be available.
If an outsourced service is to be adopted, transitional arrangements will be required to ease the passage from one regime to another. Typically, these will include end-user awareness training, technology upgrades and deployment (and associated training), promotional activities, the preparation and disposal or redeployment of old equipment, and the implementation of new clerical procedures.
When, at length, the service provider steps in, the responsibilities of the IS department do not cease. Unlike situations in which specific jobs or tasks are outsourced, seat management involves an on-going dialogue between customer and supplier within a very changeable environment. Targets and service levels are critical, but they are not the whole picture. In practice, seat management is a 24/7 business in which the ability of customer and supplier to work together to meet daily challenges may be more important than meeting predetermined objectives.
1) Federal Data Corporation, ODIN FAQ, Washington, 19 August 1999 (www.nasa-odin.com)
2) Zona Research recently estimated that PCs had the effect of decreasing productivity in many corporate environments by around one per cent
3) Jeffrey R. Phillips PhD, Lockheed Martin Services, ‘Total Cost of Ownership (TCO) - Best Practices and Seat Management’, Presentation at AIMC’98 (The Annual Information Management Conference), 23 October, 1998.
5) Op cit. Federal Data Corporation
7) Wanda Smith, ‘Be wary of agencies’ per-seat price figures’, Government Computer News, Washington, January 11, 1999
8) Op cit. Federal Data Corporation
In practice, outsourcing has been seen as a nostrum for organizations suffering from runaway costs, managerial incompetence, or lack of expertise. It is usually intended to reduce or spread risk and to cut costs
Seat management is the outsourcing of desktop systems, usually covering purchase, installation., maintenance, end-user support and ongoing development. A contract typically specifies what services will be delivered at a fixed price per seat or end-user workstation
Seat management can be problematic because of the increasingly central role of the PC in corporate management, yet it is seen by some as a viable solution to the problems posed by the PC explosion
The concept of total cost of ownership (TCO) was introduced in 1987 to provide a simple measurement of the real impact of PCs on the bottom-line by including ongoing and associated costs over the life time of a system in the price calculation
Technology driven approaches to TCO reduction include the use of management tools and the simplification of technological platforms to improve security, software distribution, and end-user support
Three quarters of TCO is estimated to arise from inefficient or over complex management, not from technology
Simple technological approaches to TCO reduction are never completely satisfactory
The most obvious method for attacking management complexity is outsourcing
The major error made in implementing desktop management solutions is lack of focus – typically demonstrated by the attempt to deliver lower costs, improved service and reduced risk simultaneously
Estimates suggest that TCO can be reduced through technology enhancements by as much as 15 per cent, and through simplifying management by twice or three times as much
Seat management converts technology management to service management and capital expenditure to revenue expenditure
Successful seat management demands carefully constructed service level agreements and ongoing cooperation between users and outsourcers
The single most important factor in making seat management work is to define objectives clearly
1) Invest in PC hardware which supports remote management and network boot.
2) Implement policy-based management of end-users and electronic software distribution
3) Centralise all helpdesk, network management and configuration management functions
4) Prioritise desktop management outcomes and buy tools or select service providers accordingly; concentrate on cost reduction, service enhancement or risk avoidance, but not all three
5) Large companies should consider purchasing so-called ‘framework products’ for managing distributed desktop environments, and be prepared to make further investments in skills. Others may obtain as much benefit from cheaper, targeted and easy-to-use tools.
6) Do not invest in too many new management products – they are bandwidth intensive and can create confusion rather than simplify procedures
7) If you decide to go down the seat management route, be clear about your objectives and have a comprehensive and mutually understood service level agreement tied to an ‘as-built’ inventory
8) Seat management requires on-going collaboration between users and outsourcers.