Hidden Costs in ICT Outsourcing Contracts Reply


Why are IT outsourcing contracts almost always delivered over-budget and over-schedule?  Why do IT outsourcing contracts almost always fail to achieve their planned value? How come IT contracts seem to be afflicted with this curse more than any other area?


The common answer is that (i) the requirements change,  and (ii) that handovers from the pre-contractual phase to in-service management are always done poorly.  These are both true although hardly explain the complexity of the situation.  If requirements change were an issue then freezing requirements would solve it – it doesn’t.  The complexity of large ICT projects is derived directly from the fact that not all the requirements are even knowable from the outset.  This high level of unknown-unknowns, coupled with the inherent interdependence of business and system requirements, means that requirements creep is not only likely but inevitable.  Secondly, (ii) handover issues should be able to be solved by unpicking the architecture and going back to the issue points.  This too is never so simple.  My own research has shown that the problem is not in the handover but that the subtleties and complexities of the project architecture is not usually pulled through into the management and delivery structures.  Simply put, it is one thing to design an elegant IT architecture.  It is another thing entirely to design it to be managed well over a number of years.  Such management requires a range of new elements and concepts that never exist in architectural design.

The primary factor contributing to excessive cost (including from schedule overrun) is poor financial modelling.  Simply put, the hidden costs were never uncovered in the first place.  Most cost models are developed by finance teams and uncover the hard costs of the project.  There are, overall however, a total of 3 cost areas which must be addressed in order to determine the true cost of it outsourcing. 

True Cost of IT

1.  Hard costs.  This is the easy stuff to count; the tangibles.  These are the standard costs, the costs of licensing, hardware, software etc.  It is not just the obvious but also includes change management (communications and training).  The Purchasor of the services should be very careful to build the most comprehensive cost model based on a detailed breakdown of the project structure, ensuring that all the relevant teams input costing details as appropriate.

2.  Soft Costs.  The construction industry, for instance, has been building things for over 10,000 years.  With this level of maturity one would imagine that soft costs would be well understood.  They are not.  With project costs in an extremely mature sector often spiralling out of proportion it is easy to see that this might also afflict the technology sector which is wildly different almost from year to year. 

Soft costs deal with the stuff that is difficult to cost; the intangibles:  The cost of information as well as process and transaction costs.  These costs are largely determined by the ratio of revenue (or budget in terms of government departments) against the Sales, General & Administration costs, i.e. the value of the use of information towards the business.  Note that this information is not already counted in the cost-of-goods-sold for specific transactions.

Soft costs go to the very heart of how a business/government department manages its information.  Are processes performed by workers on high pay-bands?  Are workflows long and convoluted?  The answers to these questions have an exponential effect on the cost of doing business in an information-centric organisation.  Indeed, even though the cost of computing hardware is decreasing, the real cost of information work – labour – is increasing.  This is not just a function of indexed costs but also the advent of increasing accreditation and institutionalisation in the knowledge worker community.  Firstly, there is greater tertiary education for knowledge work which has hitherto been unaccounted for or part of an external function.  The rise of the Business Analyst, the Enterprise Architect (and a plethora of other “architects”) all serve to drive delivery costs much higher.  Not only are the costs of this labour increasing but the labour is now institutionalised, i.e. its place and value is not questioned – despite the data showing there seems to be limited economic value added through these services (i.e. no great improvement in industry delivery costs).

3.  Project Costs.  Projects are never delivered according to plan.  Requirements are interpreted differently, the cohesion of the stakeholder team can adversely impact the management of the project, even the sheer size and complexity of the project can baffle and bewilder the most competent of teams.  Supply chain visibility, complicated security implementations and difficult management structures all add to project friction and management drag.  There are many more factors which may have an adverse or favourable effect on the cost of performing projects. 

IT Transition Cost Graph

In the Defence community, Ph.D student Ricardo Valerdi created a cost model – COSYSMO – which isolated 14 separate factors peculiar to systems engineering projects  and gave these factors cost coefficients in a cost model.  Ultimately, each factor may be scored and the scoring then determines the effort multiplier, usually a number between approximately 0.6 and 1.8.  Naturally, when all factors are taken into account the overall effect on the contract price is significant. 

More importantly, for IT implementations, the “project” is not short.  IT outsourcing projects are generally split into 2 phases:  Transition and Transformation.  Transition involves what outsourcers call “shift-and-lift” or the removal of the data centres from the customer site and rear-basing or disposal of the hardware which allows the company to realise significant cost savings on office space. 

During the second phase – Transformation – the business seeks to realise the financial benefits of outsourcing.  Here, a myriad of small projects are set about in order to change the way a business operates and thereby realise the cost benefits of computer-based work, i.e. faster processes from a reduced headcount and better processes which are performed by workers on a lower pay-band. 

IT outsourcing  is not just about the boxes and wires.  It involves all the systems, hard and soft, the people, processes and data which enable the business to derive value from its information.  Just finding all of these moving parts is a difficult task let alone throwing the whole bag of machinery over the fence to an outsourcing provider.   To continue the metaphor, if the linkages between the Purchasor and the Vendor are not maintained then the business will not work.  More importantly, certain elements will need to be rebuilt on the Purchasor’s side of this metaphorical fence, thus only serving to increase costs overall.  The financial modelling which takes into account all of these people, processes and systems must, therefore, be exceptional if an outsourcing deal is to survive.

The Financial Value of a System: how to determine how much to pay for your ICT. Reply

Much work has been done in the field of Applied Information Economics (Hubbard, John Wiley & Sons 2007).  Most of the analysis goes towards how much a business should pay for a large commercial decision but glosses over the individual value of systems, human activity and infrastructure.



It’s easy to say that a company should spend no more than $375k on an advertising campaign, for instance, but that’s easy.  How much should the company spend on it’s technology?  How much should the firm spend on its ERP system?  How much should it spend on its CRM system? Is it even possible to measure the value derived from good customer relations management and if so how much of that value can be attributed, accurately, to the technology?  So, how much is it worth the company spending to come up with that figure?


If we take this analysis further, then what is the value of a back-office system?  What decisions do back office systems assist businesses in making? ERP systems assist in monthly financial reports and variance analysis.  Project modules of ERPs assist companies to determine whether project costs have overrun.   The core value of back-office systems, as opposed to operational systems, is that they reduce risk (cost) rather than create opportunity (profit).  Operational systems which can directly increase discounted cash flows, therefore, are better suited to NPV analysis.  Back-office systems, which are largely seen as sunk costs (the cost of doing business) are better appraised through NPC analysis.  Before the business gets to that stage it must come up with the detailed cost model.


Although the types of systems is the topic of another blog, the investment goal for a back-office system should be one which reduces uncertainty in decision making.  Therefore, what is the expected opportunity loss (EOL) from poor decision making?  How does one calculate the amount of revenue lost from poor decision support?  More importantly, how does a business calculate the value of the decision support which a back-office system delivers?

  •   Estimate Financial Value.  In operational systems it is far easier.  For instance, in an investment appraisal of new systems to automate certain plant and equipment experts may attest that system controls improve efficiency by 20%.  The decision is likely to be clear cut.  What about a new ERP system?

In such cases it is important to take a holistic view of the whole ‘capability’ (i.e. the technology, people and processes together).  Imagine that the new ‘system’ will enable an engineering services firm to quote and estimate proposals (in this example it is important to imagine that the new system will enable them to do so ‘perfectly’).

In this case, what is the current value of bidding and tendering information?  With the current information, for instance, a firm may have $10 million EBIT from tenders won, based on a 60% win rate as well as a 40% cost blowout. The firm wishes to improve their profitability by increasing their bid capability (which includes cost and schedule estimation).  If each project were tendered perfectly (let us forget for the moment about failures of project delivery) they could achieve almost $17 million in EBIT.  To this end they want to know how much to invest in ICT and how much to invest in people and process in order to achieve an additional $7 million profit?


For the system to work perfectly it must not only contribute perfect information but the information must also be perfectly usable, i.e. lost revenue should be a factor of human error not human input.  Note that the system will only increase the accuracy of bids/proposals (costs and schedules etc).  It may or may not increase the probability of winning.

Using the example above, how much should should the company invest in technology?  Firstly, in this example our ‘experts’ estimated that they had 90% confidence that the business would achieve $15,500,000 with a new capability.  This is largely because the problems seem relatively known to them.  They had a 10% confidence that the business would only achieve $100,000 largely because they don’t think their analysis is wrong.  They also estimate that the break-even threshold is about $375,000 which is roughly the equivalent of one new role (to account for capacity) and some new technology to improve workflow.

Without going through Hubbard’s detailed calculations that gives us an estimated $616,878,163.00 value of information.  This means that the company should invest no more than this sum in their bid capability to achieve the desired $7 million profit.

This means that the overall capital implementation and operating expenses (let us say out to 3 year projections) should be no more than $616,000.


How much of the approximately $600k should be spent on technology and how much on people (new roles/new hires and training), and process?

CAPABILITY CONFIGURATIONS – parametric modelling

The simplest and easiest way to assess how much to spend on technology is to develop a calibrated estimate of the configuration of the capability.  In summ, there is no system or tool that can authoritatively tell a company how it should spend its cash.  The business is the expert and the best way is still to calibrate its experts to give the best options to management.

With Capability Configurations one must note that any given capability may have multiple configurations.  There is no right configuration just one that is optimised for the given parameters.  It is up to the business to offer a variety of configurations with a range of costs.

After each configuration is developed it is run through Monte Carlo simulations to determine the probability of achieving the cost target within the desired range.  The success of this method is twofold.  Firstly, it simulates a range of costs knowing that static costs cannot be predicted.  Secondly, the cost ranges are determined by experts in the first place.

It is worth noting:

  1. estimates should be calibrated and given by experts.  These are not wild guesses but represent the true high and low ends of the likely (not way-out possibilities) spectrum.
  2. the cost models for the capability configurations must be decomposed to the lowest level.  To be effective, the Monte Carlo simulations run best on more detailed models.


The simulation shown above was performed on a cost model with high application, training and infrastructure costs.  The analysis showed that there was a 99% chance of achieving costs within the $617,000 range and a 46% chance of achieving ideal costs (here estimated at approx $450,000).  The second capability configuration included more people at a greater expense and estimated a lower probability of achieving the desired $450k goal.  Intuitively, one knows that when simulating technology scenarios in the low ten thousands as opposed to people in the hundreds of thousands (salary), the probability of success will be greater.  Intuitively we know that it is less riskyto invest in technology and systems rather than people.  People always cost the most so, ideally, a business will wish to spend more on technology than people.  Some money should ideally go towards training of high-performing staff who are difficult to replace.

This was a highly simplistic model.  A more decomposed parametric showing greater detail may have yielded a slightly different result.


The major criticism is that it only takes into account hard costs and does not account for the integration of the capability into the business.  How will the capability take to the business?  Will the firm be able to develop it? implement it? run it?  These use of qualitative factors and more will be examined in another blog soon.


The Polyvalence of Knowledge: using financial ratios to inform system choice (Pt II) Reply

For too long the language of CIOs and CFOs has not been confluent.  CFOs want hard returns in the form of discounted cash flows and better return on equity (ROE).  CIOs speak of ‘intangible’ benefits, better working practices and compliance.  Despite its poor reputation, information management, however, does show up in the balance sheet.  Intuitively, it is easy to understand that if a business not only increases the speed and accuracy of decision making but also decreases the number of highly paid executives needed to do it, then the effect on the bottom line can be significant.

The first part of this blog looked at the structure of information management that shows up in the balance sheet.  This part looks at how to calculate some of those ratios.

Unfortunately, International Financial Regulation Standards (IFRS) are almost entirely geared towards the performance of capital as a measure of productive wealth.  As the information economy picks up speed, however, capital is neither the scarcest nor the more valuable resource a business can own.

The difficulty is in calculating the value of managerial decision making.  Without going in to the detailed calculations of Return-on-Management (ROM) I have outlined below two new financial ratios which allow businesses to determine a financial indicator of both information effectiveness and technology performance.

  1. Information Effectiveness.  This ratio measures the effectiveness of corporate decision making at increasing the financial value of the business.  This is defined as the decision making value minus the decision making costs.  As described in a previous blog The value of information is calculated through Information Value Added (IVA).  Information Value-Added = (Total Revenues + Intellectual Property + Intellectual Capital) – (operations value-added – capital value-added – all purchases of materials, energy and services).   This is to say that once all labour, expenses and capital (that is not part of an information system) is accounted for, the cost is subtracted from the total of gross revenues (plus IP).  In other words, it is the part of the profit which is not directly accounted for by operations or increased capital value.  It is profit which is attained purely through better managerial decision making.  This might be achieving better terms and conditions in the supply of a product or it might be in the reduction of insurance costs on a given contract.   The cost of information management is somewhat easier.  Ultimately, corporate decision making is accounted for as ‘overhead’ and therefore shows up in Sales, General & Administrative expenses (SG&A) on the balance sheet.  The sum total of managerial activity – which is ultimately what information management is for – can be accounted for within SG&A ledger.  The more operational aspects of SG&A, such as R&D costs should be removed first, however.
  2. Technology Performance.  The measurement of the ability of corporate information systems to increase company value.  Ultimately, this answers the question whether a firm’s technology is creating value, as opposed to being simply value for money.  More specifically, how much value are the company’s systems adding.  This is shown as the total value added by information support (IVA) less the total cost of technology and management, as a percentage of shareholder equity.  Note that shareholder equity is chosen above short term indicators such as EBT because many KM/management systems will take time to deliver longer term equity, as opposed to short term cash flow.  This metric assists in determining whether the cost of technology is worth the value it is creating.

Financial ratios have benefits over other performance indicators.  For instance, there is a tendency within the corporate environment to benchmark costs against firms in a similar segment.  This is excellent where small and medium sized enterprises have aggressive cost management programs.  However, in large companies with sprawling ICT environments benchmarked costs become less relevant for cost comparison and more relevant for contract management.  The benefit of financially driven information management is  that it allows a companies to benchmark against themselves.  In compiling quarterly or yearly indices firms can benchmark their own information management performance.  More importantly, these non-standard financial ratios provide not only a means for the CIO and CFO to communicate using a common language but also the ability to refine the exact nature of the solutions.

In summ, financial ratios will not tell a business what brands to buy but they will help executives refine their choice.

The Polyvalence of Knowledge (Pt I): how financial ratios can influence system choice Reply

In this, the first part of, “The Polyalence of Knowledge” we examine the use of financial analysis to inform system choice.  In particular, back-office business systems and not operational systems.  Operational systems, such as the software used to tip a smelter are best analysed through NPV.  Back-office systems, however, cannot in any way be directly linked to the increase or decrease in revenue/cash flow.  These investments are much harder to appraise because there are few ways of determining exactly how much value they add to a business.  This blog looks at how to analyse financial statements in order determine exactly which systems are needed.

Does One Size Really Fit All?

Modern systems can be described as multi-valent.  One system can act on a number of critical functional areas but does one size really fit all?  Business and ICT believe they are achieving good value for money by purchasing a single inexpensive system to achieve multi-faceted roles.  However, what ends up happening is the system achieves little in each area and becomes a costly white-elephant.

What prompts the one-size-fits-all solution?  The primary cause of many of these implementations is (a) multiple business units have problems, coupled with (b) an inability by ICT to develop precise, accurate and complex business cases directly supporting the improved financial performance of the business.  In many cases a senior executive becomes nervous about the security of information, a separate business unit voices their frustration with their inability to collaborate and co-ordinate information and ICT says that it can solve both problems with one system.

Firstly, what are the primary back-office systems, what are they used for and what are their financial benefits?

  •   Electronic Document & Record Management Systems.  EDRMS are designed for the storage and retrieval of high-value records, such as contracts, patents and other documents containing intellectual capital which is hard to replace.  The loss of such material would be considered a security breach and would compromise current and future operations.  EDRMS, unfortunately, are usually only fully implemented in back-office units which have a culture of compliance and are therefore the least likely to need it.

Due to the nature of mature documentation EDRMS typically support contracts and supply chain & vendor management.  These systems assist in the search and retrieval of framework agreements for procurement as well as operational information.  A business with a well embedded EDRMS and developed supporting business practices could expect to have lower costs in their supply chain.  Supply chains themselves tend to be capital intensive and so a high-performing supply chain will empower a greater Return-on-Assets ratio, i.e. better better contract and supply chain management tends to support higher capital utilisation.  Service companies tend not to have capital intensive supply chains and are not, therefore, significant users of EDRMSs.

  •   Business Intelligence Systems.  BI systems exist in a variety of forms and have promised much over the years.  They can be as simple as reporting tools for standard data warehouses or may be implemented as complex artificial intelligence over multiple operational systems.  BI holds the power to reduce complexity in decision making making and good BI therefore holds the power not only to reduce management staff (overhead) but also to increase a company’s Information Productivity index (a ratio showing ‘value of information’, i.e. SG&A-2-Revenue, accounting for the cost of capital).  Good BI equals good Information Productivity index. Ultimately, if a company uses its BI systems well then it will show in their Information Productivity index.
  •   Project Management Systems.  PPM systems are designed to speed the efficiency and accuracy of resource allocation across a distributed enterprise as well as contribute to better project cost control measures.  Fluctuations in resource efficiency are not usually felt in the overhead but rather in project cost and schedule overruns.  It is important to note that PPM systems are only perfect when perfectly used.  This is to say that none are effective for significant analysis or project optimisation.  However, if a distributed enterprise does not have a PPMS then the likelihood of cost and schedule overrun increase significantly beyond the standard 30% risk factor.

It should be noted that I class CRM systems as a hybrid of project and risk management systems.

  •   Messaging & Email.  Little should be said about ubiquitous messaging and email systems other than that they are merely a cost of doing business.  Costs of these systems should be seen as sunk costs because the modern business simply cannot afford to do without them.  When building business cases for modern messaging firms could actually look further valuable social networking applications for the following reasons:
  1. The structure of SN groups of interest already provides valuable metadata to automatically tag messages for archive, search and retrieval.
  2. Parceling conversations by subject, group and associative images works more similarly to human memory than standard systems.
  3. The development of Communities of Interest (COIs) along with the web-based structure and storage of documents/non-critical records is both easier and more secure.

For these reasons and more companies should seriously look to SN apps to replace standard email systems.  More importantly, the security and storage issues taken on by SN apps remove the necessity for the rollout of the plethora of inappropriate SharePoint implementations.  In these latter cases greater attention could be paid to the development of more focused operational systems.

  •   Enterprise Resource Management Systems.  ERPS reduce the clerical burden of processing payroll and human resource transactions.  The value of ERPS is in the amount of overhead (finance and HR) labour they can remove from a business.  Good implementations of ERPS should show in reduced labour and lower SG&A costs.
  •   Knowledge Management Systems.  KMSs exist to store the non-critical knowledge capital and intellectual assets of a firm.  These may simply be the records and materials one needs for daily work.  For instance it has been estimated that it costs a law firm $100,000 in lost knowledge when a partner leaves.  Alternatively, KMSs may also store the accumulated knowledge capital of a firm, such as frameworks and intellectual property.  Businesses which use KMSs well have a higher Knowledge Capital value which is the difference between market value and shareholder equity, less estimated good will.
  •   Collaboration Systems.  Collaboration systems or team sites are generally smaller, simpler and locally managed KMSs.  MS SharePoint is one such example and also shares functionality with PPMSs.  Any financial benefits will be similar to standard KMSs.
  •   Enterprise Risk Management Systems.  xRMSs exist to reduce a broad spectrum of risks across the enterprise.  They provide a database for the documentation of risk although they offer little analytical capacity.  Separate systems often must be used for this.  xRM effectiveness will normally only show in reduced project cost and schedule overruns.  However, a large company may also be able to reduce its risk premium through the effective management of financial risk.

Why haven’t back-office systems increased information productivity, knowledge capital and asset utilisation?  Simply, the ICT and Finance functions do not generally work together to support cost reduction or revenue growth strategies.  Largely, back-office systems are implemented to satisfy the whims of personal functionality, security or broad-based compliance issues.  In addition, current financial ratios focus almost entirely on capital rather than information yet it could be said that the former is neither the scarcest nor most expensive anymore.

Read more in the next instalment of this blog to see how to analyse the financial systems in order to determine what applications the business requires.