|
|
5.3.1 Scope of IT Accounting
5.3.2 Business Perspective
5.3.3 Building the Cost Model
5.3.4 Cost Types
5.3.5 Depreciation
5.3.6 Apportioning the IT Services
costs
5.3.7 Level of detail required in
cost calculations
5.3.8 Variable Costs or Indirect Costs
requiring apportionment
5.3.9 Calculating the Cost-by-service
5.3.10 Calculating the costs of Cost
Units
5.3.11 Changes affecting costs
5.3.12 Investment appraisal
5.3.13 Total Cost Of Ownership
5.3.14 Budgeting, IT Accounting and
Charging cycles
5.3.15 Charging in Profit Centres
The basic IT Accounting principles (why do it, what to do and who it affects) are common to all Business processes. To a large extent, the implementation of IT Accounting is similar throughout the organisation but the detail of what to Cost, and how to cost it, is dependent upon the type of Service being provided.
To implement IT Accounting may require improved IT Accounting in many areas, for example in staff time and activity recording, supplier contracts, software licensing, resource metering or accommodation costs.
Other items which rely upon the information provided by IT Accounting and hence may dictate the shape of a IT Accounting model include:
IT Accounting can be very complex and if implemented at too high a level of detail, may cost more than the benefits realised. The IT Accounting systems described here should enable an organisation to:
IT Accounting and Charging are integral parts of IT Infrastructure Management responsibilities. The policies must be agreed by the organisation, i.e. at board level.
The difference between IT Accounting and Charging and the responsibility for each, has to be defined and made clear by the IT organisation. They are different but have linked sets of activities.
The most visible of the two disciplines, Charging, is concerned with the recovery of the cost of IT Services expenditures in a simple, fair, affordable way. IT Accounting is concerned with providing detailed information on where and for what reason expenditure is incurred within IT Services and is inward-looking.
The management of the IT organisation can choose to implement Accounting for IT Services (IT Accounting) with no Charging, or to charge for IT Services (either to break-even or to make profits). If IT Services have to make a book profit, the organisation may even create IT Services as a separate legal entity:
The IT Services organisation must be assumed to pay for its own upkeep, probably with some financial support for capital expenditures. Choosing to operate as a Profit Centre is usually the first step along the path to a truly commercial IT Services organisation - one in which the Customers (the people who pay for the services) have some freedom of choice to go elsewhere if they are not satisfied with the quality and/or price. The Profit Centre approach provides the IT Services organisation with a certain amount of autonomy, but it also carries risks, such as the Customer becoming aware that the IT organisation is 'making a profit' from them. This can be more serious if another Service provider can offer a cheaper service of similar standard (corporate loyalty disappears very quickly in this situation).
To calculate the costs of IT Service provision, it is necessary to design a framework in which all known costs can be recorded and allocated to specific Customers, activities or other Category. This is called a Cost Model. Most Cost Models are based on calculating the cost for each Customer but other models can be developed to show the cost for each service or the costs for each location. This Chapter concentrates on a Cost Model that enables the calculation of Costs-by-Customer, which is the usual start-point if a Charging system is to be introduced.
As discussed in 5.2.2 Estimating the cost of budget items, it is useful to categorise costs to ensure that they are correctly identified and managed. This categorisation should use consistent and easily understandable Cost Types. For producing a Cost Model, the suggested Cost Types are:
External Service costs and Transfer costs need further explanation. It is now common to buy in services from external parties (external services) that are a mixture of cost types, for example an outsourced service for providing an organisation's application development or the provision of a datacentre. It may be difficult to break down this cost (into each of the first four categories), as it is likely to contain elements that are indivisible or that the supplier will not wish to detail. It is easier and more usual to categorise this as an External Service Cost.
Transfer costs are those that represent goods and services that are sold from one part of an organisation to another (often within a multi-national or other large organisation that has a sophisticated internal accounting system). Transfer costs may be for:
Transfer costs should be visible in the cost model because people may forget that internal goods and services represent a cost to the organisation and are part of the cost of providing service. Hence a false figure may be reached when assessing costs if a service is dependent upon activity from another part of the organisation but this cost is excluded from calculations. Some organisations will insist on these transfer costs being accounted for in each part of the organisation while others may only use them when modelling costs and no money will actually pass across the organisation. In this publication, it is assumed that it is not necessary to separately identify transfer costs in the Cost Model examples.
The Costs-by-Customer Cost Model requires that all major cost elements in the current or proposed IT budget are identified and then attributed to the Customers who 'cause' them. To do this, the costs first have to be identified as either Direct or Indirect:
Any Indirect Costs, which cannot be apportioned to a set of Customers (sometimes called Unabsorbed Overheads), have then to be recovered from all Customers in as fair a way as is possible, usually by uplifting the costs calculated so far by a set amount. This ensures that the sum of all of the costs attributed to each Customer still equals the total costs incurred by the IT organisation - in Table 5.4, this is referred to as the 'balance check'. This 'balance check' can be applied to costs divided in other ways e.g. by service or by location; always, the sum of the parts should equal the whole.
If the Cost Model is being produced for the first time the categories and Cost Elements for it need to be developed first, to a level of detail that meets the needs of IT Accounting and of any Charging to be performed. Hence an understanding of Charging policies (see Paragraph 5.4.2) is necessary when the Cost Model is drawn up.
If costs are mainly Direct, perhaps because each Customer has independent hardware and software, the method of recording and of apportioning costs can be very simple. For example, if Finance are the only Customers of the General Ledgers and the system on which it runs, all costs directly associated with the General Ledgers, including purchase, maintenance and support, can be attributed to Finance department's code in the ledgers (often called a cost-centre or charge-code).
However, if Resources are shared, for instance a mainframe running applications for more than one Customer, the hardware costs may have to be classified as indirect and apportioned to each Customer, say by CPU-seconds/disk storage/print volumes/etc from workload predictions. To do this requires a model that allows these costs to be spread across a number of Customers.
In the example in Figure 5.3, it is assumed that there are 3 businesses or departments, who together are responsible for all of the IT costs. The three departments are Marketing and Sales, Manufacturing and Finance and all of the IT systems and services have been implemented on their behalf.
Figure 5.3 - Cost Model of Costs-by-Customer
An example of the calculation of a Cost Model for a simple Cost-by-Customer is shown in Figure 5.3. The same principles can be applied to calculating the costs of individual application services or even parts of a service e.g. support and maintenance.
If the cost of providing some element of a service is desired, for instance producing large reports for the Marketing and Sales department, this may require measurement of resource usage to apportion indirect costs to this one activity e.g. computer time consumed, printing and operations staff and facilities. This can become very complex and would normally be treated as a separate exercise rather than as part of the Standard cost Model used for calculating Costs-by-Customer.
To be able to derive cost information and report it in the formats required by different parts of the organisation, it is necessary to ensure that all costs recorded are classified to a standard system with a level of detail that anticipates future Changes, e.g. New Cost centres, new equipment types, new project codes.
If more detail is required in calculating costs, the chosen major Cost Types of hardware, software, people, accommodation and transfer can be further divided. For instance, hardware might be divided into Office, Network, and Central Servers. The purpose of this is to ensure that every cost identified in the IT organisation can be placed within a table of costs, by type. This enables analysis to be performed by type e.g. all Network costs.
The decision on whether to identify more detailed Cost units often depends upon whether more detail is required to apportion charges. In general, Cost Elements are the same as budget line items where the purpose of the model is simple recovery of costs.
If a more detailed analysis of costs is required, e.g. for organisations providing shared services, then more detailed Cost Elements have to be identified. Typical Cost Elements within each major Cost Type are shown in Table 5.3.
|
Major type |
Cost Elements |
|
Hardware |
Central processing units, LANS, disk storage, peripherals, wide area network, PCs, portables, local servers |
|
Software |
Operating systems, scheduling tools, applications, databases, personal productivity tools, monitoring tools, analysis packages |
|
People |
Payroll costs, benefit cars, re-location costs, expenses, overtime, consultancy |
|
Accommodation |
Offices, storage, secure areas, utilities |
|
External Service |
Security services, Disaster Recovery services, Outsourcing services, HR overhead |
|
Transfer |
Internal charges from other cost centres within the organisation |
Table 5.3 - Cost Element examples
For organisations providing services based upon central mainframes, the hardware costs may be the largest proportion but it is more common to see a rough balance amongst hardware, software and people. Increasingly, the proportion of costs attributed to networking devices and network services is becoming more significant and may be identified as a separate Cost Type.
Organisations that purchase software products, rather than developing them, find a higher proportion for costs categorised as Software. Organisations that use outsourcing services (such as offshore development or computing services) may see External Service costs as the largest proportion of costs.
When each Cost Element has been identified, it should be classified, as a minimum, as either Capital Costs or Operational Costs (also known as Current Expenditure or Revenue Expenditure).
For financial purposes, costs are classified into either Capital or Operational when the financial ledgers are reported (the 'books'). This is because Capital expenditure is assumed to increase the total value of the company, while Operational expenditure does not, although in practice the value of Capital expenditure decreases over time (depreciates).
This distinction affects IT Accounting because the Cost Model needs a method of calculating the annual cost of using a capital item (e.g. mainframe) to deliver IT Service. The annual costs must make allowance for the decreasing value of capital items (assets) and make for timely renewal of capital items e.g. buildings, servers, applications. Usually, this is taken as the annual Depreciation, from a method set by the Finance department (within the boundaries of the country's laws).
Capital Costs are typically those applying to the physical (substantial) assets of the organisation. Traditionally this was the accommodation and machinery necessary to produce the organisation's product. Capital Costs are the purchase or major enhancement of fixed assets, for example computer equipment, building and plant are often also referred to as 'one-off' costs. It is important to remember that it is not usually the actual cost of items purchased during the year that is included in the calculation of the cost of the services but the annualised depreciation for the year as shown in Table 5.4 (see Paragraph 5.3.6).
Operational Costs are those resulting from the day-to-day running of the IT Services organisation, e.g. staff costs, hardware maintenance and electricity, and relate to repeating payments whose effects can be measured within a short timeframe, usually less than the 12-month Financial year.
The following list gives typical examples of the Cost Elements, classified into Capital expenditure and Operational expenditure (revenue) items:
Organisations that measure themselves primarily on cash flow or budget adherence may have specific rules about the categorisation of costs in the accounts that differ from those who measure Return on Capital Employed (ROCE - see Paragraph 5.3.12).
The organisation's accountants explain the rules for identifying capital items and this depends upon a number of business decisions. Many organisations choose to identify major expenditure as Capital, whether there is a substantial Asset or not, to reduce the Impact on the current financial year of such expenditure and this is referred to as 'Capitalisation'. The most common item for this to be applied to is software, whether developed in-house or purchased.
The reason for this is that a business that is investing in a major software development, that provides service for a number of years, does not want to show all of the costs in a single year (and so, potentially, an operating loss). The board wants the value of the company and its shares to reflect the investment made but adding the cost of the item to the assets of the company without adjusting cash flow in some way would also give a false picture. The agreed method is to show Capital and Operational expenditure separately but to apply rules of depreciation, described later in this Section, to provide a balance. This system allows an organisation to spread the cost of a major purchase over a number of years although, as with all systems, many additional rules (and laws) have to be written to prevent fraud or misleading of investors.
Some costs can be directly attributed to a single Customer or group of Customers and these are referred to as Direct Costs. Examples of direct costs would be a server or application used exclusively by a single cost-centre. Usually, all direct costs are attributed to the Customer incurring them but there are occasions when these might be subsidised, shared, deferred or ignored, such as for the use of a new system which is subsequently 'rolled-out' throughout the organisation.
Other costs, such as operations staff in a Data Centre cannot be easily attributed to the running of a specific Customer's service and these are referred to as Indirect Costs or shared costs.
To fully attribute all costs requires some form of division of indirect costs by a fair method of re-apportioning. An example would be operations staff whose total cost could be apportioned to the businesses on the basis of the number of Users in each business. Often this apportionment is not completely accurate but the costs to the organisation of a more accurate calculation, is far too high. The system of apportioning has therefore to be affordable, clear, fair and in accordance with good IT Accounting practice.
It is usually necessary to be able to apportion Direct Costs to a specific business group, department or external Customer. Most businesses allocate Cost Centres to units within the organisation that relate directly to the general ledger system. The number and types of Cost Centres differ from installation to installation due to size and organisational structure. Occasionally, specific projects or initiatives may be allocated a separate Cost Centre to enable costs for them to be 'ring-fenced'.
Costs that do not vary even when resource usage varies are referred to as Fixed Costs. Examples of this would be a maintenance contract for a server or a corporate software licence (within agreed User limits).
Variable Costs are those that vary with some factor, such as usage or time. They are likely to be used for Cost Elements which cannot be easily predicted and which it is to the benefit of both supplier and Customer to determine the costs exactly, perhaps for variable charges to be applied. Examples of charges that might vary, because the underlying cost varies, are out-of-hours cover, major equipment re-location, and the production of additional quarterly reports.
A Cost Element such as filestore may be considered to be variable. If a Customer requires an additional 10Gb it may be possible to calculate that the cost of this is £1000 and hence the cost per Gb is £100. The danger of this approach is that there are often sharp changes in costs because they cannot be continuously scaled: the next disk drive may require another cabinet, an additional Process run on the server may cause queuing problems resulting in all jobs taking longer to run.
It is sometimes necessary to view a cost as having a fixed element and a variable element, for example, using filestore at all requires disk controllers and bandwidth, causing a fixed cost. The variable cost of additional disk drives can then be calculated and added to the fixed portion. This level of detail is not usually needed in calculating the cost of a service but can be useful when evaluating competing technologies or services.
Depreciation is the measure of the wearing out, consumption or other reduction in the useful economic life of a fixed asset, whether from use, passage of time, or obsolescence through technological or market changes. Depreciation should be allocated so as to charge a fair proportion of cost or valuation of the asset to each IT Accounting period expected to benefit from its use. This point can be a delicate balance and many IT organisations face a difficulty in funding the replacement of apparently useful items that have no capital value (i.e. are fully depreciated).
An example is a PC which cannot perform at the level required but has not yet been fully depreciated. The organisation may not be willing to replace it because it still is an asset with a value. Similarly, a PC may be perfectly functioning but has been fully depreciated and now has no value in the business accounts - strictly, it is no longer a cost in providing services but it is unlikely that costs will be recalculated to take account of this.
The assessment of depreciation, and its allocation to IT Accounting periods, involves the consideration of three factors:
The useful economic life of an asset may be:
The depreciation methods used should be the ones most appropriate having regard to the types of assets and their use in the business. The Finance department gives guidance in this. The most common methods of assessing depreciation are:
The Finance department may require IT assets to be 'written-off' before the end of their useful life, increasing the apparent cost of services but facilitating a charging system that generates revenue for the early replacement of systems.
Consider a company with three departments who require IT Services - Marketing and Sales, Manufacturing, and Finance. Each is asked to contribute towards the IT budget, based upon the services they require. Each Cost Element in the IT budget has to be identified; classifying them by type (hardware, software etc.) helps ensure that all such costs are found. It must then be decided whether these are Direct Costs or Indirect Costs and how they are allocated to Customers (in this case, other departments of the company).
In the example at Table 5.4, all the costs of providing the shared computer Infrastructure - cables, servers, routers, software, have been grouped into a single Indirect Cost Element called 'Infrastructure' for which a common apportionment method can be used; in this case, the number of Users of the Infrastructure. This simplifies the spreadsheet and enables a simple calculation of the cost of adding new Users to the network.
| Capital |
Annualised |
Direct |
Apportionment Method |
Customer |
||||
|
|
|
Marketing and Sales |
Manufacturing |
Finance |
||||
|
Hardware |
|
|
||||||
|
UNIX Server |
Yes |
£34,667 |
No |
50/50 split |
£17,333 |
£17,333 |
||
|
NT Server |
Yes |
£4,333 |
Yes |
£4,333 |
||||
|
Netware Server |
Yes |
£1,300 |
No |
Infrastructure |
||||
|
PCs (50) |
Yes |
£26,000 |
No |
By PC |
£5,200 |
£19,240 |
£1,560 |
|
|
Routers (5) |
Yes |
£1,300 |
No |
Infrastructure |
||||
|
LAN Cabling |
Yes |
£17,333 |
No |
Infrastructure |
||||
|
|
|
|||||||
|
Software |
|
|
||||||
|
General |
No |
£20,000 |
Yes |
£20,000 |
||||
|
ORACLE |
No |
£7,000 |
Yes |
£7,000 |
||||
|
Marketing and |
No |
£3,000 |
Yes |
£3,000 |
||||
|
MS Windows |
No |
£2,500 |
No |
By PC |
£500 |
£1,850 |
£150 |
|
|
MS Office |
No |
£3,000 |
No |
Licence |
£1,500 |
£1,050 |
£450 |
|
|
Netware |
No |
£3,000 |
No |
Infrastructure |
||||
|
NT |
No |
£2,500 |
No |
Infrastructure |
||||
|
|
|
|||||||
|
Employment |
|
|
||||||
|
Manager |
No |
£50,000 |
No |
Unabsorbed overhead |
||||
|
Technical |
No |
£30,000 |
No |
Unabsorbed overhead |
||||
|
Assistant |
No |
£20,000 |
No |
Unabsorbed overhead |
||||
|
Contractor |
No |
£100,000 |
No |
|||||
|
|
|
|||||||
|
Accommodation |
|
|
||||||
|
Computer Room |
No |
£10,000 |
No |
Unabsorbed overhead |
||||
|
Office |
No |
£10,000 |
No |
Unabsorbed overhead |
||||
|
|
|
|||||||
|
External Service |
|
|
||||||
|
Wide Area |
No |
£20,000 |
No |
Infrastructure |
||||
|
DR contract |
No |
£10,000 |
No |
Unabsorbed overhead |
||||
|
Total Costs |
|
£275,933 |
|
|||||
|
Direct and Apportioned costs |
£100,500 |
£31.867 |
£46,473 |
£22,160 |
||||
|
Absorbed costs (Infrastructure) |
£45,433 |
20%/74%/6% |
£9,087 |
£33,620 |
£2,726 |
|||
|
Unabsorbed costs |
£130,000 |
89.1% uplift |
£36,482 |
£71,349 |
£22,169 |
|||
|
£275,933 |
£77,436 |
£151,442 |
£47,055 |
|||||
| balance check for the 3 customers above: £275,933 | ||||||||
|
Notes |
||||||||
|
1. For capital items, this is 1/3 of the purchase price (the agreed depreciation) plus the annual operational cost |
||||||||
|
2. The M&S application will cost £100,000 to develop (one year of contractor), and the support contract is £3,000 annually |
||||||||
|
3. Includes NI, pension and other benefits, usually adding between 30% and 50% to salary |
||||||||
|
4. Contractor is employed to develop new system for Marketing and Sales and is funded directly by them |
||||||||
|
6. Marketing has 10 PCs with all software, Operations has 37 PCs but only 7 with Microsoft Office, Finance has 3 PCs with all software |
||||||||
|
7. Infrastructure costs will be added (absorbed) based upon numbers of PCs in each department, i.e. Marketing 20%, Operations 74%, Finance 6% |
||||||||
|
8. Unabsorbed overheads are added onto each cost centre by uplifting it by 89.1%, to ensure full recovery |
||||||||
Table 5.4 - Spreadsheet example of Full costs-by-Customer
Rather than trying to determine the actual usage of the Infrastructure that all departments rely upon, it is simpler to group all Infrastructure Cost Elements into one and determine a fair way of recovering those costs, e.g. by number of Users in each department.
In practice, the Cost Elements in each of the major Cost Types may be groups of items. For instance the UNIX server Cost Element may consist of a number of items e.g. central processor, UPS, filestore, peripherals but these rarely need to be identified individually unless the Cost Model has to show costs to that level.
Some organisations wish to calculate the cost of delivering individual services or activities. To do this requires a more complex Cost Model, with Cost Elements apportioned by Cost Units. These Cost Units are usually things that can be easily counted, such as staff numbers, software licences or things easily measured, such as filestore usage, CPU usage, network packets sent.
A higher level of detail is required to apportion fairly Cost Elements that are Indirect, but very few apportionment methods can ever be completely accurate. For instance, if central server usage is attributed by CPU-seconds, the issues of filestore and filestore transfers arise. Then, who 'pays' for the queuing and 'Paging' of virtual store? Large servers are rarely dedicated to a single function and the expense of apportioning their use to each Customer can be both divisive and prohibitive. Similarly, it is difficult to apportion the time of support staff to individual Customers without detailed timesheets, and the benefits of this level of detail may be less than the cost of obtaining it.
At all times, when performing a detailed cost analysis it is essential to ensure that the value provided by the answers is not outweighed by the costs of data collection and analysis. For most business-driven Cost Models, apportionment should be firstly simple, secondly fair and thirdly accurate (if possible).
Cost Units are the basic items of resource for which Customers are held accountable i.e. provide the method of apportioning Indirect Costs or calculating the actual cost of Variable Costs.
Cost Units should be chosen to enable simple, fair apportionment. As central hardware costs have become a small proportion of costs in many organisations, the use of CPU-seconds as a Cost Unit has decreased. It is used as an example here because it is one of the simplest measures, easy to measure on mainframe systems, accurate and apparently fair.
Staff hours are another easily identifiable cost unit, as is floor space. In the example at Table 5.5, PCs and software licences are both Cost Units.
After the Cost Unit has been set, the cost of each Cost Unit must be determined. Following the example above of CPU-seconds, if a business's IT organisation knows the total cost of the necessary Capacity, the average unit cost of a CPU-second can be calculated and apportioned to each Customer as it is used, in the same manner that an electricity company calculates the consumption of each of its Customers in KWh. In a Cost Model using Cost Units for every Cost Element, the sum of [units consumed X the cost per unit] should be demonstrated to be equal to the total cost of the IT Services (see also the 'balance check' in Table 5.4).
Figure 5.3 considered the Cost Model for calculating costs attributable to a Customer. To calculate the Costs-by-service, the Cost Model may require more detail, as discussed in Paragraph 5.3.7.
The basic approach is similar:
1. identify all those costs that can be directly attributed to the service being analysed, for instance any dedicated hardware, software, staff or contracts
2. decide how to apportion the Indirect Costs such as Infrastructure
3. adjust the total to allow for 'hidden costs' or 'Unabsorbed overheads' such as IT management or buildings - this must be the same uplift figure calculated for the whole model, or used from the Costs-by-Customer Cost Model, i.e. 89.1%.
It is best to consider the model as being one layer of a set of models, one for each service, so that, if the costs for each service provided are added together, this would again give the total of the costs incurred by the IT organisation. Even if only Service A is of interest, it is important to ensure that all costs are identified and attributed to the other services, as shown in Figure 5.4.
Figure 5.4 - Complete view of the Cost-by-service
The original spreadsheet in Table 5.4 can be used as a start-point. It is then necessary to identify those costs that can fairly be attributed directly to the provision of this service and to apportion those Indirect Costs in a manner that reflects the proportion of them that result from the provision of this service.
It may not be possible to identify costs in every category. For example there may be no Direct Accommodation costs and no Absorbed Accommodation or Staff costs - these may all be included in the Unabsorbed Costs (see Figure 5.5), which have to be included fairly in the costs of every service.
Figure 5.5 - Cost Model for a Cost-by-service
The remaining complexity is how to account for the £100,000 development costs (one year of contractor's time). Since this software is owned by the company (an asset), the Finance department value it. If its net value is less than £100,000 after development is completed, there has been some misunderstanding when the business case was made for spending £100,000. If the software is to be used for the next 5 years, it could be depreciated over that period, representing £20,000 pa costs (excluding maintenance) and this cost attributed to the cost of the service to the Marketing and Sales department.
| Capital |
Annualised |
Direct | Cost Unit |
Total Capacity or numbers |
Cost per Cost Unit |
Usage by M&S | Cost | |||||
|
Hardware |
|
|
||||||||||
|
UNIX Server |
Yes |
£34,667 |
No |
Average % CPU |
100% |
£34,667 |
37% |
£12,827 |
||||
|
NT Server |
Yes |
£4,333 |
Yes |
N/A |
£5,000 |
|||||||
|
Netware Server |
Yes |
£1,300 |
No |
Infrastructure |
||||||||
|
PCs (50) |
Yes |
£26,000 |
No |
PC |
50 |
£520 |
5 |
£2,600 |
||||
|
Routers (5) |
Yes |
£1,300 |
No |
Infrastructure |
||||||||
|
LAN Cabling |
Yes |
£17,333 |
No |
Infrastructure |
||||||||
|
|
|
|||||||||||
|
Software |
|
|
||||||||||
|
General Ledger |
No |
£20,000 |
Yes |
User |
5 |
£4,000 |
0 |
£0 |
||||
|
ORACLE |
No |
£7,000 |
Yes |
User |
5 |
£1,400 |
1 |
£1,400 |
||||
|
Marketing and Sales appl. |
No |
£3,000 |
Yes |
N/A |
£3,000 |
|||||||
|
MS Windows |
No |
£2,500 |
No |
PC |
50 |
£50 |
5 |
£250 |
||||
|
MS Office |
No |
£3,000 |
No |
Licence |
20 |
£150 |
0 |
£0 |
||||
|
Netware |
No |
£3,000 |
No |
Infrastructure |
||||||||
|
NT |
No |
£2,500 |
No |
Infrastructure |
||||||||
|
|
|
|||||||||||
|
Employment |
|
|
||||||||||
|
Manager |
No |
£50,000 |
No |
Unabsorbed overhead |
||||||||
|
Technical Support |
No |
£30,000 |
No |
Unabsorbed overhead |
||||||||
|
Assistant |
No |
£20,000 |
No |
Unabsorbed overhead |
||||||||
|
Contractor |
|
|
||||||||||
|
|
|
|||||||||||
|
Accommodation |
|
|
||||||||||
|
Computer Room |
No |
£10,000 |
No |
Unabsorbed overhead |
||||||||
|
Office |
No |
£10,000 |
No |
Unabsorbed overhead |
||||||||
|
|
|
|||||||||||
|
External Service |
|
|
||||||||||
|
Wide Area connection |
No |
£20,000 |
No |
Infrastructure |
||||||||
|
DR contract |
No |
£10,000 |
No |
Unabsorbed overhead |
||||||||
|
Total Costs |
|
£275,933 |
|
£25,077 |
||||||||
|
Absorbed Costs Proportion of Infrastructure Costs (£45,433 ) |
Employee using Infrastructure | 50 | £909 | 5 | £4,543 | |||||||
|
Total |
£29,620 |
|||||||||||
|
89.1% uplift to recover Unabsorbed Costs (see note 2) |
£26,391 |
£26,391 |
||||||||||
|
£56,011 |
||||||||||||
| Application Cost (annual) | £20,000 | |||||||||||
|
Total Annual Cost |
£76,011 | |||||||||||
|
Notes |
||||||||||||
|
1. The contractor costs is spread over 5 years of the application, i.e. £20,000 pa |
||||||||||||
|
2. The uplift is the amount calculated from the overall Cost-by-Customer model, discussed previously |
||||||||||||
When comparing the cost of internal provision of service against a bureau or packaged solution, a figure of £74,568 is reached, for comparison. The most contentious issue this brings to light is the 89.1% uplift (£26,391) that is mainly caused by the unabsorbed overheads of People and Accommodation cost. If the Customer feels that they could run the application themselves with the existing staff in their department administering it, they will argue strongly that the IT organisation is inefficient.
There is no immediate solution to this. All analyses that break costs down to the minute suffer both from not taking the whole picture into account (i.e. the provision of IT throughout the organisation is most likely to be more efficient if centralised) and from the fact that single Cost Units cannot alone be used to calculate the cost of service provision. However, the bureaucracy involved in more detailed calculation may be prohibitive.
The exercise is valid, however, providing that these issues are understood and discussed. Ultimately, strong corporate support is required for central IT systems.
Note that if the 89.1% uplift were ignored, the costs calculated would not be a true reflection of the cost of service provision and could mislead decision-making.
When calculating how to apportion Indirect Costs or how to assess the actual cost of a Variable Cost, it is important to select a Cost Unit that is a factor that directly and fairly represents the cost incurred. For example, using the PC as a Cost Unit for calculating the apportionment of Infrastructure costs.
When calculating the cost of a Cost Unit, it may be that the precise calculation results in an amount that is dependent upon when the item is used or upon how many Customers use the resource. For instance, if processor-second is chosen as a Cost Unit for determining the apportionment of costs across a number of Users, this varies with the total usage of the server. That is, the cost per Cost Unit varies unless the server runs at full Capacity all of the time.
To prevent absurdities arising, such as attributing all the costs during overnight running of an unattended server to the sole Customer, it is usual to set fixed rates based upon the calculation of the Capacity available in a fixed period. Hence a Standard Cost for a processor-second can be derived but this must be regularly recalculated to prevent anomalies from creeping in.
Table 5.6 demonstrates the calculation of some simple Costs of Cost Units but even this is beyond that which would normally be performed by an in-house IT organisation.
|
Cost Unit |
Total available in 1 month |
Total cost in one month |
Cost per Cost Unit |
Note |
|
CPU-seconds |
792,000 |
£100,000 |
£0.126 / CPU-sec |
Based upon 10 hours x 22 days i.e. excluding out-of-hours |
|
Filestore |
4,000 Gb |
£12,000 |
£3/Gb |
Based upon allocated filestore (not actual used space) and annual cost of £36/Gb |
|
Tapes |
10,000 |
£2,500 |
£0.25/tape |
Based upon tapes allocated, £30 each with a life of 10 years, hence annual budget for new tapes is £30x1000 |
|
Operator hours |
528 |
£30,000 |
£56/hour |
3 operators, working, on average, 22 days x 8 hours |
Table 5.6 - Calculating the Cost of Cost Units
Table 5.7 shows the cost of a given activity that requires 10,000 CPU-seconds of processing, 100Gb of filestore, 40 tapes and 3 operators for 10 hours can then be calculated by adding the costs of each Cost Element.
|
Cost Unit |
Quantity |
Cost per Cost Unit |
|
|
CPU-secs |
10,000 |
x £0.126 |
£1,260 |
|
Filestore |
100 |
x £3 |
£ 300 |
|
Tapes |
40 |
x £0.25 |
£ 10 |
|
Operator hours |
30 |
x £56 |
£1,680 |
|
Total |
£3,250 |
Table 5.7 - Calculating Activity Costs
It is possible to simplify this further. By using only the CPU-seconds as the factor and by dividing the £3,250 by 10,000 future cost calculations can be based on £0.325 per CPU-second, ignoring the other factors. This, of course, assumes that the other items are always used in broadly the same ratios as the example above.
To ensure true costs, 'Tapes' might have to include the cost of tape decks and maintenance, Operator hours may have to include Service Desk and support staff. In current practice it is rare to perform these calculations 'from the bottom up', i.e. by first breaking down all resource usage to its component parts. Most businesses are content with choosing a fair factor that is agreed to be the measure of usage and use this.
Even after calculating these detailed costs, many installations reduce all measures of usage to two Cost Units: processor time and filestore occupancy. By taking the total costs (including accommodation, hardware, staff etc) of storage and dividing this by the storage Capacity, a cost per Cost Unit can be determined. Similarly, the cost of providing the processor time (including accommodation, staff, Disaster Recovery etc.) can be divided by the cumulative processor-seconds of processing time to arrive at a cost for each CPU-second. Costs are then allocated to Cost Centres solely on the basis of processor time and filestore occupancy.
When systems Change, underlying costs also change. To avoid having to recalculate costs for Cost Units too frequently, IT Finance Management may choose to anticipate Changes or to defer the inclusion of the cost in the Cost Model until stable usage has been established.
An example would be the addition of extra filestore. The purchase cost of the extra filestore can be easily determined but this may affect the calculated cost per Cost Unit of filestore. This is because some overhead costs used in calculating the cost of filestore Cost Units may not change e.g. accommodation, people cost, and some factors may increase the overall costs for filestore, such as the need for additional disk controllers or accommodation.
When performing detailed cost analysis, a cost recovery plans is needed to provide a picture of under (or over) calculation. IT Finance Management may choose to leave the cost per Cost Unit unchanged until the error is greater than, say, 5% to avoid copious recalculations.
There are a number a methods employed to reduce the effect on costs of Changes to equipment and services. They are included here for completeness as they were widely used in the past by mainframe-based organisations and computer bureaux.
One common method involves establishing only the average cost of machine time and using this average as the sole Cost Unit. To do this, the total costs for providing the machine during a given period are divided by the total number of hours it is used and the resultant cost per hour of service time is used to calculate costs incurred by each Customer or usage.
The main advantage of the method is that all costs are attributed and a simple rate per hour of machine time is available for budgetary purposes.
The main disadvantages to a Customer, however, are:
The main disadvantages to an IT organisation are:
Standard Rate is calculated by estimating the costs of the resources necessary to meet an estimated Capacity, based upon the forecast Workloads and service requirements of all the Customers, and dividing this by a usage factor, such as number of Customers files or system Users. The resultant rate is then fixed for an agreed period.
The advantage of this approach is that a consistent rate is used for a whole period, allowing Customers to predict costs simply.
The disadvantage is that the calculation differs from true cost if the workload changes, for instance if a Customer changes their business process or submits increased volumes of work. Providing that Charging is not directly based on this rate calculation, the impact is minimal and can be addressed when the rate is next calculated.
The current installation life of the equipment has a direct bearing on the rate. Thus, recently installed equipment tends to be lightly loaded giving a high average cost, while a machine just about to be upgraded and, therefore, heavily loaded, gives rise to a low average rate. The major flaw in this method, therefore, is that as usage increases so the cost of machine time per hour decreases. The effect is to show reducing costs when, in fact, no such decrease is actually taking place.
The use of Standard Rate on a standard machine is a refinement of the Standard Rate described above and is an attempt to overcome the disadvantages that are produced by that method. The rate is not based on the actual forecast machine usage but rather on use of a machine at an average load. The effect is to fix the costs of machine time at what should reflect 'the market rate' - that is the rate charged by external organisations. Estimates can be made of a probable optimum configuration and the resultant additional costs and throughput potential added to the calculation.
The advantage of using this method is that a constant rate is obtained, which is not influenced by the actual equipment installed and allows Customers to match their usage to the anticipated costs. The disadvantage is that the IT organisation has to be capable of delivering service, at all times, to this Standard Rate. It may not be able to sustain this as business environments and Customer usages change.
Sound IT Accounting methods, including the Cost Models described previously, enable businesses to determine the costs of IT Service provision. These costs can be used in Investment Appraisal, a process of determining whether the business benefit from changes to IT Service quantity and quality. Techniques of appraisal have been developed mainly in the context of decisions on capital spending, but the general principles apply to any proposal for spending or saving money that involves changes in the use of resources. Systematic appraisal entails:
Used properly, appraisal leads to better decisions by policy makers and managers; it encourages both groups of people to question and justify what they do and it provides a framework for rational thought about the use of limited resources. It establishes the link between business operations and the cost of underpinning them; it enables the IT organisation to support the business in its cost/benefit analyses.
Investment Appraisal has long been recognised as an essential prerequisite to sound financial management both in the public and private sectors. The importance of Investment Appraisal has grown in recent years as an aid to decision-making in its broadest sense, as a means of identifying efficiency savings and controlling investment expenditure to maximum effectiveness.
There is often a trade-off between capital investment and running cost expenditure: i.e. between maximising effectiveness in the long-term and the risk of failing to achieve short-term goals. Capital investment decisions are essentially longer-term decisions, and thus it is more difficult to hold management responsible and accountable for such decisions. However, because the performance of a manager is often measured on the efficient and effective use of allocated resources within a budget period (of one year), there are only limited ways of holding managers responsible for investment decisions: this is why sound Investment Appraisal procedures are essential to an organisation.
The purchase of IT equipment and software is one of the most important investment decisions an organisation must make. It is therefore crucial that such decisions are properly included in the organisation's strategic planning. In all well-run businesses, there is a requirement to prepare and update a business IS Strategy, thus IT managers both at the strategic and operational levels have a central Role in helping to achieve organisational aims and objectives.
From the viewpoint of the business manager, IT investment and the supply of IT Services is the same as any other planned expenditure or allocation of resource in that it is measured in terms of its contribution to the effective, efficient and economic achievement of business goals. It must enable the business to determine whether, for example, the returns would be better from a new IT system or from increased advertising.
Many organisations now insist that IT projects, in line with other business projects, calculate a Return On Investment (ROI). This enables decision-making to be based on common business standards and allows comparisons between investment in IT and non-IT projects.
The factors classed as a 'Return' vary from organisation to organisation and from year to year. In concept, a return is a revenue or benefit, or the prevention of a lost benefit or revenue (Opportunity cost), which is attributable to the project. If the return is divided by the expenditure required to complete the project, a figure is derived which can be compared with the returns on doing nothing (often taken as the bank base rate) or the returns from other projects or initiatives.
ROI = average increase in profits(average taken over an agreed number of years) / Investment
The Investment figure may take into account the cost of borrowing money.
The accuracy of the estimates for ROIs is often challenged but that is the nature of estimating benefits. To improve the acceptance of these calculations, it is important to involve the business in determining the returns and how they are to be measured. It is essential that the Capital (one-off) and Operational (on-going) costs for a given project or initiative are identified and can be recorded in the manner previously shown.
Frequently, the benefits lag behind the expenditure and the Return has to be calculated over an agreed period. For some financially aggressive companies this could be 12 months. For others, it could be between 5 and 10 years. The exact expectations of the Financial Management may be dictated by board-level strategies, in turn responding to shareholder demands.
Shareholders and potential investors in a company look very closely at a number of 'ratios'. The actual ratios felt to be important are subject to fashion but a primary one is the ROCE.
ROCE = Net Profit Before Tax and Interest / Total assets less current liabilities
This ratio, of the Returns over the Capital Employed, is one frequently used by business analysts to judge the effectiveness of the organisation as a whole. Any changes to services or products would normally be expected to improve this figure and hence the ROCE calculated for a proposed project should be higher than the overall organisation ROCE.
The use of ROCE enables comparisons of different investment opportunities in a fair way and represents how effectively an organisation generates revenue from its assets (the capital employed). At its most basic, this can be compared with bank base rate that offers a ROCE that has minimal risk.
Adventurous organisations usually seek to employ their capital in furtherance of their business while organisations that are undergoing a period of stabilisation reserve the capital for future use.
The Gartner Group pioneered a method of calculating the costs of a product or service with the title of 'Total Cost of Ownership' (TCO). This referred to assessing the Lifecycle costs of an item rather than just the visible capital expenditure.
The most widely known example was that for Personal Computers. In an era where the price of a PC on a desk had fallen to $2,000, Gartner demonstrated that the 5-year cost of a PC, when taking into account purchasing overheads, upgrades, maintenance, a proportion of support staff and Service Desk costs, disposal etc. was closer to $35,000.
They argued that the cost per PC per year was thus $7,000. This is actually the same process as that described in Paragraph 5.3.9 but applied to a single product / activity rather than a Customer or service.
It remains a vivid demonstration of the difference between purchase price and ownership cost which many Customers of IT organisations still find surprising, particularly when IT budgets are discussed.
The current extension of this is to include Social costs such as recycling, environmental damage or work injury to produce Total Life Cost.
If the Cost Model is being developed during the IT financial year, the figures used are based upon the costs already incurred and those predicted for the remainder of the year.
If the Cost Model is being developed to enable the calculation of future IT budgets, it depends upon workload, Capacity and technology requirements from the businesses. It is likely that there will be some iteration or adjustment to both the model and the resulting costs during the budget preparation period (see Figure 5.1).
Table 5.8 shows that there are two distinct cycles associated with Budgeting, IT Accounting and Charging:
|
Budgeting |
IT Accounting |
Charging |
|
|
Planning (annual) |
Agree overall expenditures |
Establish standard Unit costs for each IT resource |
Establish pricing policy Publish price list |
|
Operational (monthly) |
Take actions to manage budget exceptions or changed costs |
Monitor expenditure by cost-centre |
Compile and issue bills |
Table 5.8 - Budgeting, IT Accounting and Charging cycles
It is important to plan the expenditure and cost recovery profiles periodically. The plans should cover each month in the financial year to facilitate monthly monitoring and cover peaks and troughs. Actual expenditures and recoveries are monitored and compared at the end of each month: the expenditures and recoveries are also compared with what was planned. It is also appropriate to plan Changes to the IT Accounting system to coincide with SLA reviews and any Change should be reflected in the SLAs.
In a Profit Centre the objective is to recover, through Charging, an amount greater than the costs incurred. To manage this, it is usual to be able to report the profitability of individual Customers and services at various points in time. The revenues and expenditures are controlled through the Budgeting cycle and operational management of the agreed budgets. The performance of the IT organisation and its managers may be evaluated by measuring both profitability and expenditure.
This may require the overall Financial Model for the IT organisation to include profit and margin information for individual services and the Cost Model must have sufficient detail to enable these calculations.
|
|
