System request and feasibility: ROI, NPV, payback

A project passes through two gates before any code is written. The first is the system request: a short document from a business unit saying “we need this system, because…”. The second is feasibility analysis: the answer to “can we do it, should we do it, will it be used if we do it?”. On the exam two types of question come from this week: definitions (the types of feasibility, the parts of a system request) and calculations (NPV, ROI, payback). The calculation question is almost guaranteed and carries high marks.

Golden rules:

1. Three feasibilities, three questions: technical = Can we build it? economic = Should we build it? organizational = If we build it, will they come?
2. Money has a time value. 1 TL today is worth more than 1 TL in 3 years. NPV takes this into account.
3. Build a table. Years in the columns; benefits, costs, net, PV, cumulative in the rows. Most calculation errors come from not building a table.
4. Write the unit: NPV is money (TL), ROI is a percentage, payback is years.

System request

A project idea usually comes from the business side: a manager, a department. This idea is put in writing with a system request and presented to the approval committee.

Business value is of two kinds. Tangible value can be measured and converted into money: “a 20% increase in collection of late fees”. Intangible value is real but cannot be converted into money: “a better student experience”. The point that gets mixed up on the exam: “customer satisfaction” is intangible, “15 minutes less staff time per customer complaint” is tangible.

Technical feasibility

“Can we build this system?” The answer is really a risk assessment. Four factors raise the risk:

Economic feasibility

“Should we do it?” This is a cost-benefit analysis. Four steps:

1. Identify the costs and benefits.
2. Assign values (give each one a monetary equivalent).
3. Spread the cash flow over the years.
4. Compute the financial measures: ROI, NPV, break-even point.

Costs are of two kinds. Development costs are one-time (software license, developer pay, hardware, training). Operational costs repeat every year (maintenance, cloud bill, support staff).

Formulas

Present value: the present equivalent of an amount $F$ to be received $t$ years from now, with discount rate $r$:

$PV = \frac{F}{(1+r)^t}$

Net present value:

$NPV = \sum_{t=0}^{n} \frac{B_t - C_t}{(1+r)^t}$

Here $B_t$ is the benefit and $C_t$ is the cost. Year 0 (today) is not discounted. If NPV > 0, the project is economically acceptable.

Return on investment:

$ROI = \frac{\text{total benefits} - \text{total costs}}{\text{total costs}}$

Payback period (break-even point): the moment the cumulative cash flow passes zero. Within a year the flow is assumed to arrive evenly:

$\text{payback} = (\text{last negative year}) + \frac{\text{shortfall at the end of that year}}{\text{next year's flow}}$

Worked example: the FitHub membership system

A gym chain is considering an online membership and door access system.

• Development cost (year 0): 150,000 TL
• Operating cost: 20,000 TL per year (years 1-5)
• Benefit (staff savings + fewer lost memberships): 70,000 TL per year (years 1-5)
• Discount rate: 8%

Step 1: Net cash flow. Each year 70,000 − 20,000 = 50,000 TL. In year 0, −150,000.

Step 2: PV of each year.

Year Net flow    1.08^t    PV
 0   -150,000    1.0000   -150,000.00
 1     50,000    1.0800     46,296.30
 2     50,000    1.1664     42,866.94
 3     50,000    1.2597     39,691.61
 4     50,000    1.3605     36,751.49
 5     50,000    1.4693     34,029.16

Step 3: NPV. Add them all:

46,296.30 + 42,866.94 + 39,691.61
 + 36,751.49 + 34,029.16 = 199,635.50
NPV = 199,635.50 - 150,000 = 49,635.50 TL

NPV is positive: the project is economically feasible.

Step 4: ROI (nominal). Total benefit 5 × 70,000 = 350,000. Total cost 150,000 + 5 × 20,000 = 250,000.

$ROI = \frac{350000 - 250000}{250000} = 0.40 = 40\%$

Some sources compute ROI with PVs. Then the PV of benefits is 279,489.70, the PV of costs is 229,854.20, and ROI comes out as 21.59%. Use whichever one the question asks for; if it is not specified, state your method.

Step 5: Payback. Simple (undiscounted) cumulative: −100,000, −50,000, 0. It breaks even exactly at the end of year 3: 3.00 years.

Discounted (with PVs) cumulative:

Year PV          Cumulative
 0  -150,000.00  -150,000.00
 1    46,296.30  -103,703.70
 2    42,866.94   -60,836.76
 3    39,691.61   -21,145.15
 4    36,751.49    15,606.34   <- sign changed

Payback = 3 + 21,145.15 / 36,751.49 = 3 + 0.575 = 3.58 years. The discounted payback is always longer than the simple one, because future money is scaled down.

PITFALL · discounting year 0. Money spent today is already today’s money: $1.08^0 = 1$. A student who divides year 0 by 1.08 inflates the NPV.

TIP · shortcut for equal flows. If the same amount comes every year, instead of adding the PVs one by one you can add the factors: for 5 years at 8%, $\sum 1/1.08^t = 3.9927$. 50,000 × 3.9927 = 199,635.50.

Second example: uneven flows

A campus food-ordering app costs 100,000 TL, with net flows of 20,000, 30,000, 70,000, 50,000 TL.

Year Flow      Cumulative
 0  -100,000  -100,000
 1    20,000   -80,000
 2    30,000   -50,000
 3    70,000   +20,000   <- crossover in year 3
payback = 2 + 50,000 / 70,000 = 2.71 years

In the division, put the crossover year’s own flow (70,000) in the denominator, not the average flow.

Organizational feasibility

“If we build it, will they use it?” Two views:

Strategic alignment. How well does the project match the organization’s goals? A project that directly serves a strategic goal has low risk.

Stakeholder analysis. Everyone affected by the project:

Resistance arises in a project where users are excluded: even the best system has zero value if it isn’t used.

Project selection

The feasibility result goes to the approval committee. The committee evaluates the project within the organization’s project portfolio: it balances risky and low-risk, short and long, strategic and operational projects. The feasibility study is updated throughout the project; it is reviewed again at the end of analysis and design.

Common mistakes