Information Technology (IT) represents the largest portion of capital expenditures in many corporations, particularly in the service sector. And with so many IT investments having produced at best questionable value, it is not surprising that CEOs and CFOs insist on getting a complete business case for large IT capex. But, is a financial business case an appropriate tool to allocate capital to IT?
The financial perspective
CEOs and CFOs are judged by their ability to create economic value. Net present value (NPV) provides a sound basis for investment decision making. Future cash flows associated to an investment are discounted using the firm’s opportunity cost of capital and then algebraically added. If NPV is positive, economic value is added to the firm.
Calculating the NPV of an IT investment is relatively easy in the case of an application intended to automate tasks currently executed by people. Automation exchanges economic factors, providing capital and freeing human talent for more sophisticated tasks. Positive cash flows represent recurrent labor cost savings during the IT application lifecycle, whereas negative cash flows result from IT capital and operational expenditures (including of course the cost of IT professionals).
The problem with intangible benefits
In the case of automation, both positive and negative cash flows are numbers with relatively low uncertainty. Thus, NPV standard deviation is rather small. In less technical terms, NVP calculations from IT automation of business tasks are reasonably reliable. However, estimating economic benefits from other categories of IT investments could be more challenging. Some examples to illustrate this point:
- Processes coordination: reduced coordination costs, such as defects, rework, process cycle time, and excess slack resources.
- Operational risk reduction: reduced losses resulting from more effective controls.
- Customer’s value enhancement: additional revenues associated to a better value proposition, which in turn result from better customer intelligence and/or improved product quality and service metrics.
- Information for decision making: improved organizational effectiveness regarding goal setting, resource allocation and performance management, which all benefit from better decision quality.
- Strategic positioning: enhanced competitiveness resulting from boosting bargaining power against customers and suppliers, creating industry entry barriers, enhancing network effects and sustaining a superior business “ecosystem”.
All these cases have a common difficulty when compared with our automation example: cash flow estimation is subject to considerably more uncertainty. This list is somewhat ordered from lower to higher cash flow uncertainty. Some authors refer to this as the challenge of estimating the value of intangible benefits. In our automation example, the IT application directly reduces a cost element which is clearly identified in the income statement; that is why they call it a “tangible” benefit. In the other cases, IT investments improve intermediate variables (operational, marketing, managerial, and strategic) which in turn are supposed to increase revenues and reduce business costs. There is a recognized cause and effect connection between the intermediate variables and cash flows, but the financial model that is expected to give us hard estimations for the impact (how much and when) is difficult to construct and/or validate.
For example, everybody intuits that if you implement correctly an application aimed at enhancing the customer experience, service quality and customer satisfaction should improve. Thus, customer retention should increase and word of mouth should attract new customers, resulting in better revenues. The causality is clear, but if we are building a business case, we need a hard number for increased revenues. Getting to that hard number is a hard endeavor.
How financial modeling usually deal with intangible benefits?
Technically, the uncertainty challenge of intangible benefits could be partially overcome by introducing a probabilistic distribution for those cash flows. In that way, financial modeling software can calculate a probabilistic distribution for NPV, including its mean and standard deviation. Undoubtedly, this is a demanding exercise, in particular if it has to be performed for a portfolio of several significant IT investments.
When financial analysts are pressed (under time constraints) to construct financial models that take into account intangible benefits of IT investments, they tend to play safe. If they feel they are going to be accountable for a cash flow (benefit) with a large variance resulting from a multiple step obscure causality, they skew estimates to the conservative side. Instead of modeling uncertainty (probabilistic distributions) in cash flows, they opt for providing a very cautious expected positive cash flow. As a result, most intangible benefits of IT investments are under-represented in NPV calculations.
Applied systematically, this approach tends to pop up in the IT investment portfolio those initiatives with tangible benefits, pushing down initiatives rich in intangible benefits. Do this over a decade in an organization and don’t be surprised that IT maturity in that company does not grow and that business people start questioning the business value of IT. Organizations that practice this kind of conservative financial evaluation of IT investments end up with an IT portfolio over-invested in cost-reduction initiatives and support functions, and under-invested in revenue-generation, effectiveness-enhancement initiatives and business processes.
Getting it right
We agree with the core of CEO’s and CFO’s perspective: IT represents a large share of capex and opex in most organizations and IT investments should justify themselves; as any other investment, they are supposed to create economic value. But if the way we use financial modeling tools to evaluate IT investments creates a self-fulfilling prophecy of low business value of IT, then we need to re-evaluate our approach. Some suggestions:
- Start with the right modeling team: the best modeling team includes people who will ask the right questions (financial analyst) and people who can provide the best answers, regarding both positive cash flows (business owner of the IT initiative), and negative cash flows (IT leader in charge of developing and supporting the IT initiative).
- Consider both tangible and intangible benefits: Go over a list of diverse kind of benefits (like the one we discussed above) and identify how the IT initiative could drive each one of them. Weigh qualitatively the impact across benefit categories. Probably only two or three benefit categories will be relevant to each IT initiative. Concentrate your modeling efforts in those high impact benefit categories.
- Consider diverse kind of risks: An IT initiative could be affected by diverse type of risks. Some come from the business side: expected business impact may depend on the uncertain response of external stakeholders; organizational units affected by the implementation of the IT initiative may resist change; expected scope and functionality may be ill-defined. Other risks come from the technology side: the IT organization in charge of the IT initiative may lack the skills or experience to develop and implement it, or to coordinate an external provider; the project may be too big for the typical experience of the IT department; and even the underlying technology itself may be still immature and/or standards may be still unstable. Again, weigh qualitatively the impact across risk categories, and select the few relevant, concentrating your modeling effort in those high impact risk categories.
- Understand how risk affects cash flows: build a cash flow vs. risk matrix, including both positive cash flow types (associated to IT-driven business benefits) and negative cash flow types (associated to IT capex and opex). Each cell of the matrix should represent a qualitative assessment of how each risk type may affect each cash flow type.
- Model specific cash flows: for each benefit do your best at modeling explicitly positive cash flows. If modeling vagueness and/or risk categories creates uncertainty on a specific cash flow, represent it explicitly considering discrete values ranging from worst-case to best-case scenarios. A discrete probability distribution is a rough approximation to representing uncertainty, but much better than ignoring it or skewing value to the conservative side. A good starting point is a probability distribution with three values (worst-case, typical, and best-case) and probabilities for each of them assigned by the modeling team considering the risk matrix. Repeat for negative cash flows. Also, always remember to model the “incremental project”, i.e. cash flows resulting from implementing the IT initiative compared to doing nothing and maintaining the status-quo.
- Simulate and analyze the probabilistic distribution for NPV: use a financial modeling software to calculate NPV statistics (mean, standard deviation) and probability distribution. Play with the model to understand sensitivity to key variables. In particular, study those cases where NPV mean is positive but still there is a significant probability that it could turn negative. NPV statistics and sensitivities are critical inputs to an informed decision regarding approval or rejection of a proposed large IT initiative.
Of course, this is not a final word regarding IT business cases! Some authors have proposed practical multi-attribute decision making approaches to consider value and risk, from the business and IT domain. Real options represent a modern financial tool that can be applied to strategic IT initiatives. And creating a business case for each major IT investment is only an input to be considered when optimizing the whole IT investment portfolio. But this is a good starting point to escape the trap of simplistic financial modeling that creates a self-fulfilling prophecy of low business value of IT.
Recent Comments