1. Issues
2. Audit
3. Identifying items supported under an agreement
4. Reconciling data from Information Services
   
5. Interaction with the service provider
6. Total cost of ownership
7. Benchmarks
8. Summary
9. References

Issues

Although outsourcing is a valid approach, it should not be seen as a silver bullet, there are many issues needing to be addressed. One of the major issues is expectations. In an area as difficult to define as IT support, agreements can be reached that, when implemented, cause major dislocation. Sometimes, outsourcing is seen as a way of getting rid of the problem of how to provide effective IT support. This is quite the wrong way to approach developing effective support arrangements. Responsibility for these services cannot be shed and ultimately it is still the responsibility of the unit in question.  Outsourcing is most successful when there remain competent and knowledgeable in-house staff who take responsibility for seeing that the contract is carried out effectively.

Contracts

An outsourcing contract needs to identify a number of deliverables. These should be decided before any contract or agreement is signed as this is where there have been some major misunderstandings.

The agreements are usually in the form of:

• Service Level Agreements (SLAs), comprising service descriptions defining what service(s) are involved
• Service levels, the performance levels to be attained
• Agreements, stating what will happen if Service levels are not met and other related terms and conditions.

Service levels

Service levels are usually couched in terms of percentages and hours. The hours cover both the hours of the day and days of the week during which service is provided as well as the time to either respond to a problem or to fix the problem. The Response Time is usually the only one stated because the time actually to fix a problem can be difficult to determine and therefore to specify in advance. The percentages are measures of the number of times the response is less than the agreed upon time.

Agreement

An Agreement may state that for 99% of the time the response to a problem will be less than 4 hours. It is worth noting that it is not possible to guarantee 100% availability of response within any particular time frame. The best targets are 99.999%, which means that out of 100,000 problems only one was outside the agreed time. However, this can be very expensive to achieve.

Details of the hours of coverage also need to be carefully thought out. If help is required outside normal working hours there can be considerable costs and this may include continuing work on a problem already being addressed.

The support Agreement also needs to cover what is being provided and at what cost. A 'time and materials' agreement can be OK, but if there is not very careful monitoring and management of the times allocated the monthly invoice can be excessive. If people think they can ring the support desk on any issue and get it resolved in a short time they must be aware that there is a real cost involved.

The Agreement should also cover reporting of what has happened and who initiated what. It should be detailed to an agreed upon level and be delivered at agreed upon times.

The outsourcing unit must make it quite clear who has responsibility both for evaluating proposals, and for monitoring performance once an agreement has been established.

Responsibility

Outsourcing does not absolve the unit from responsibility for the service, indeed it usually makes it all the more important to have clear responsibilities set out. If that person does not have substantial experience in handling IT services, then it is vital that a third party be engaged to assist in evaluating proposals any of the existing UWA faculty IT managers would be suitable (the preferred arrangement in view of the need for understanding of the IT scene at UWA), or an independent consultant may be engaged.

Penalty clauses are sometimes included but enforcing them and arguing over the actual reason for the invocation of the penalty can be very time consuming.

The Agreement should involve regular meetings between at least one Faculty/School representative and the service provider.

An Agreement should always have an escape clause covering the position where the support is not acceptable and you want to get out. It should also include a start and finish date.

Back to top

Audit

Before going out to any supplier of support services the following Audit should be undertaken:

1. The desktop machines that will be involved. This should include make, model, age, operating system, criticality times, locations.
2. Printers, as for desktops plus an indication as to the traffic through the printer.
3. The network switches and configuration. Include make, model, capacity, software versions, management software installed.
4. Servers, make, model, age, size of disk space, operating system, criticality times, locations; what they are used for. With the servers, the availability of stabilised power supplies is also relevant.
5. Rules and procedures, what rules and procedures are at present in place for the above systems; how well are they documented and enforced; this includes things like any quota systems, times to create new users.
6. The present systems, to see how they are working and whether there are any SLAs already in place.

Back to top

Identifying items supported under an agreement

Desktop and printer systems

This can be broken into a range of sub issues:

• Office and academic machines
• Research machines
• Teaching machines.

Within this there is also the issue of machine type, Macintosh, Windows and/or Linux.

What is being supported:

• Is the support just for the system itself? If the computer boots and can access the network and the software loads, is that sufficient or does support extend to packages not working as they are supposed to? This leads to deciding what packages are involved and what versions of the packages.
• Does the support extend to updating the versions of the operating system drivers and software that is loaded?
• Does the support extend to helping with formatting and printing problems?

Network systems

• Does the support cover the support for the network within the section or department?
• Who liaises with Information Services (IS) for network issues outside the section?
• Who identifies and fixes traffic problems when the network goes slow?
• Who installs new switches and analyses the traffic on the present systems?
• Who recommends network configurations and structures? Who checks it?
• Who is responsible for responding to hacker attacks, “take down” orders and other emergencies?

Server systems

• Who manages the servers, updates, etc?
• Who manages the users, creates new ones, manages their space etc?
• Who provides the backup management and tape management?
• Who decides on backup strategies?
• Who watches the capacity of memory, processor, disk and reports on requirements?
• Who fixes the servers when they fail?
• Who specifies the new servers when they are required?
• How much is charged for the recover of a lost file?
• Who will be in charge of the security of the systems? Such things as updating the virus software and identifying attacks at and from the Faculty/School need to be addressed.

Back to top

Reconciling data from Information Services

Each month UWA Information Services sends bills for internet use. Will this be handled by the support organisation or the Faculty/School?

Back to top

Interaction with the service provider

Questions that should be quantified beforehand and specified under any support agreement may include:

• How many support people are at present involved? Is this adequate?
• What are the qualifications of the people?
• There will need to be a confidentiality agreement? The agreement may be stored on servers and desktops maintained by the service provider.
• How much say do you have in the selection of the people provided?

Ongoing support arrangements

It is likely that a support centre will charge some kind of a per-call and per-time interval rate. If someone calls the centre and has a long explanation of the problem, it will cost; this is where a time-based versus flat fee based contract becomes important. The support organisation is there to make money, they will charge for what they provide.

The actual procedures required will need to be clearly identified. Can individual staff members contact the support centre? There will be a need to identify who can contact the support centre and who is responsible if someone other than those people contact the centre.

It is also important to ensure that students know to whom they can take problems or queries. For students, but also for staff, it is also important to ensure that the provision of IT services and assistance is coordinated with other providers of IT services at UWA, so that problems are directed in a straightforward way to others who may have responsibility for dealing with them, and no user is left not knowing who will be addressing their need.

What kind of response is required for each type of problem? Critical problems like the computer not working will require a different response to a query on word processing. Who identifies what is critical?

Regular reporting and billing

It is important to obtain regular reporting and costing information from the supplier, so that blow-outs in costs do not come as a big surprise and the causes of excessive service use are investigated early.

Escalation procedures

Ensure that the supplier has effective means of escalating its handling of problems, and that these are well understood by all those who may have direct contact with the supplier (for example, for reporting problems). It is also wise to establish internal escalation arrangements to match the supplier’s – such as from end-user, to contract manager (for example, School Manager), to Head of School (and beyond if necessary). The circumstances under which problems will be escalated should also be spelt out.

Problem closure

There must be clear and agreed arrangements and criteria for declaring a problem 'resolved' or 'closed'. This should not be left to the unilateral decision of the supplier, who will, of course, be interested in closing a potentially intractable problem as early as possible. On the other hand, it is not reasonable to expect a supplier to work on a problem till final resolution, given the non-perfect nature of most standard software and the rapidly changing status of technology. But care should also be taken to ensure that “blame-shifting” does not happen – a circumstance to which units may be particularly vulnerable given the range of suppliers of IT services at UWA, such as ACS, IS, AARNet.

Back to top

Total cost of ownership

Most people underestimate how much it actually costs to acquire and operate a desktop computer. The purchase cost is often the one that dominates thinking, though the biggest costs are usually in providing various support services. Outsourcing brings many of these ongoing, support costs into sharp focus, and quotes for service are frequently much larger than people expect, indeed, modest costs should be regarded with a certain degree of suspicion.

Gartner, for instance, has done quite a lot of work on 'total cost of ownership', see Heine, 2001, and Kirwin, 2002 whose analysis indicates that the total cost of ownership for every computer is between $US4,000 and $US18,000 per annum. These figures are probably excessive in the context of Australia, and of the university environment. One study undertaken in Melbourne suggests that the figure is more like $A2,000 per annum.

The UWA Technical Advisory Group intends to continue to research this area, in order to try to identify realistic benchmarks for UWA to apply.

Back to top

Benchmarks

There have been many studies on the ratio of the number of staff to the number of desktop computers. This is perhaps the most common way to determine the workload by both the Faculty/School and the service provider. Unfortunately the numbers vary so widely that these figures may not be very helpful. The factors affecting the applicability of such ratios in particular situations are:

1. The type of computer, such as Apple Macs are often considered easier to support than PCs.
2. The age of the computer. Older machines often have more problems. Many of the studies on applicable ratios assume a 2- or 3-year replacement cycle which is rarely possible in a university setting.
3. Whether the computer is used for administrative, teaching, student laboratory or research purposes. This may affect the amount of work it does, the variety of packages loaded on it, the care with which the computer is looked after.
   
4. The diversity of computers. If there is a wide variety of machines to be supported, the demands on the support organisation increase, often substantially.
5. The location of the computers.  Physically diverse locations are more difficult to support and travelling time between locations may be factored into the costs.
6. The level of training of the people using the computer. Untrained users often take more resources to support. Sometimes service providers will assume, expect or require that all computer users have undertaken basic formal IT training in every computer, printer and software package they use.
7. The number of packages being supported on the computer. Support costs increase exponentially, as packages interact with each other.
8. The diversity of the software on the computers. If there are multiple versions of a particular product the support difficulty increases. Some service providers will not support use of more than the latest and/or second-latest version of any package.
9. The number and diversity of the printers. This can sometimes be a disproportionately difficult area.
10. The configuration of the computers. Under-configured computers have more problems.
11. The availability of help from nearby colleagues as opposed to obtaining it from the support organisation.

Back to top

Summary

Outsourcing is a valid approach to support for information technology and communications systems but it must be approached with care. It should be remembered that outsourcing companies are experienced and skilled at negotiating outsource contracts and their business depends on being able to do it effectively. They will ensure that the terms will not disadvantage them.

Significant economies of scale can be achieved in the area of IT support. Small units can rarely afford to acquire the whole range of skills and expertise that they need on their own but by combining with other units a critical mass can be achieved which often represents the best value-for-money for each individual unit as well as for the University as a whole.

The best outsourced contracts emerge when there is a degree of stability in the service to be supplied. University IT requirements are rarely stable enough to make outsourcing very effective and this may make internal outsourcing a better proposition. Also, it may well pay to acquire the services of specialist in IT support when that may not be the core competency of the Faculty/School. There have been some very large outsourcing contracts set up to undertake IT services for organisations like banks and government departments. But it should also be noted that very few of these have been found to deliver exceptional value-for-money over long periods and there have even been some spectacular failures. There is even some indication of a trend at the present time to bring some IT outsourcing back in-house.

The final consideration is to ensure that adequate IT expertise and resources are still available to the Faculty/School to undertake strategic planning, especially as issues such as increased use of IT in teaching and learning are becoming so important world-wide.

Back to top

References

• Stone, Lisa (Gartner): Critical 'Success Factors for Outsourcing Relationships', 17 September 2002,
• Maltus, Richard, et al (Gartner): 'CIO Update: Benchmarking Helps Outsourcing Deals Stay Competitive', 3 July 2002,
• Cox, Roger Clift, Jonathan (Gartner): 'How to Reduce Costs in Long-Term Outsourcing Deals', 28 January 2002,
• Grigg, Jeremy (Gartner): 'Financial Engineering: The Outsourcing Challenge', 17 September 2002,
• Harris, Kimberly, et al (Gartner): 'FAQ: Outsourcing Alternatives — Distinctions and Forecast', 28 May 2002,
• Heine, Jack (Gartner): 'Where Cost Reductions Lie: TCO Lessons Learned', 2 March 2001,
• Gomolski, Barbara (Gartner): 'Implications of Outsourcing on Budgeting and Chargeback', 3 June 2002,
• Harris, Marti, et al (Gartner): 'PC Life Cycle Management on Campus', 13 November 2001,
• Kirwin, Bill, et al (Gartner): 'The Total Cost of Ownership Index: Defining the Database', 16 October 2002.

Back to top