At the end of each sprint, the goal is to have a deliverable product that is sellable.
Obviously, we may not see a real product develop in the first few sprints since the product is moving from nonexistence to some level of functionality. However, the intent is to deliver a functional product to the customer with every release. What this means is that the organization that is providing the investment to produce the product will already be receiving payback on the development money even while the product development is under way (assuming we are receiving purchase orders for the delivered prototypes/products). This approach also results in a substantial reduction in risk since the product is always in some kind of condition for release.
Additionally, the customer/client can exercise the product to determine if it meets specifications or if it has any unsavory behaviors.
By now, the observant reader should notice the metaphorical parallels between the scrum approach to managing time and lean manufacturing. Daily sprint meetings and semi-monthly or monthly retrospectives and sprint planning meetings shorten the duration between iterations, very much like level loading and one-piece flow in lean manufacturing. If a production line is properly level-loaded, the manufacturer should always have some product to deliver even if it is not the entire lot (known as split-lot delivery), since something is usually better than nothing.
2.8.1 Payback Period
One measure of project success is how long it takes for the developing organization to recover their investment. When the product has multiple deliveries that can be sold, we see opportunities to begin receiving revenue for the product. This desirable situation starts the payback period even while the development work is ongoing.
Ultimately, these iterations will shorten the product payback period, often within a few sprint periods (see Figure 2.22).
6/9/2009 6/25/2010 7/1/2009 8/1/2009 9/1/2009 10/1/2009 11/1/200912/1/2009 1/1/2010 2/1/2010 3/1/2010 4/1/2010 5/1/2010 6/1/2010
Cumulative Investment
Traditional Payback Start
Scrum Payback Start
Figure 2.22 Example of payback impact of scrum.
It is important to note that scrum will only be expeditious if the objectives can be achieved in the short-term. Long lead-time items such as design-from-scratch hardware are more difficult to achieve using the scrum approach, although the prod-uct development can certainly be decomposed sufficiently to allow for a scrum-type approach to product delivery. These obstacles can be overcome through the use of commercial-off-the-shelf (COTS) hardware, modular hardware design philosophies, or extensive use of simulation. Alternatively, we can break the long lead-time item into smaller increments and apply the scrum approach to the subsystems.
If the project duration is substantial, we will probably want to use a more sophis-ticated approach to payback. The discounted payback takes into account the interest rates involved, although it makes the calculation of the payback period somewhat more complex. The two tables that follow show the difference between simple pay-back and discounted paypay-back. With simple paypay-back (Table 2.3), we see repayment in five years, whereas with discounted payback (Table 2.4), we see our repayment sometime in year seven.
COTS means using a hardware product that has already been developed. This requires identification and analysis by the customer and determination of the final application of the product. Even though the hardware is outsourced, it still plays a role in achieving the customer’s or the organization’s objectives so it must be given adequate attention.
A modular hardware design philosophy speeds the hardware development to match the scrum tempo. In this case, the designing organization creates the sub-assemblies of the hardwares so that the developers can reuse these portions of the
Table 2.3 Simple Payback
Cash flow 5,000
Initial expense 25,000 Payback, in years 5
Table 2.4 Discounted Payback
Year Cash Flow Multiplier Discounted Flow Result
0 −25,000 1 −25,000 −25,000
1 5,000 0.909 4,545 −20,455
2 5,000 0.826 4,130 −16,325
3 5,000 0.751 3,755 −12,570
4 5,000 0.683 3,415 −9,155
5 5,000 0.621 3,105 −6,050
6 5,000 0.564 2,820 −3,230
7 5,000 0.513 2,565 −665
8 5,000 0.467 2,335 1,670
design for subsequent designs. By eliminating some of the design cycles through reuse, we accelerate throughput. No matter which choice we make, the velocity may be dominated by the slowest design portion.
The lack of availability of hardware early in the process can be overcome through extensive use of simulation. The sprints themselves can be used to provide oversight for developing the models for the simulation. It is also possible to inject the simulation during the later sprints when we have assembled models. We can use simulation at any level, from the creation of the first working model to the final system. We recommend that scrum teams consider the benefit of simulation:
Early tryout of ideas
Damage-free exercise of the product
Quick turnaround
Coordination with the customer/client
Examination of extreme scenarios
2.8.2 Return on Investment
Return on investment is another measure of how successful a project is monetarily:
It is the ratio of the money generated as a result of the project to the amount of money invested to produce the product. Again, earlier payback can improve the business case for the development activities since we have revenue generated even as the product is undergoing development. Earlier payback goes beyond simple return on investment by improving on the net present value of the project by improving on the payout (negative cash flow) taken for non-recurrent engineering at the very
beginning of the project. Cash flow moves in the desired direction much earlier when we can invoice for existing development during the project rather than waiting for the terminus of the project timeline.
What follows is a simple example of a return on investment calculation.
Return
Investment = %ROI 20,000
100,000 = 0.2
2.8.3 Internal Rate of Return
The internal rate of return (IRR) is the net present value (NPV) of the project with the NPV set to zero and the discount rate calculated instead of the dollar value; in other words, we are calculating the interest rate rather than some dollar value. The primary defect of this approach is the lack of scalability when comparing projects with significant resource differences. Our example shows an initial investment of 100 units with a cash flow of 120 for one period of time. We calculate the IRR to be 20%. Often, enterprises will set hurdle rates for the IRR calculation to force a cut-off point where the project is considered to be unsatisfactory.
For IRR NPV = 0
−100+ 120
[(1+IRR100)1] = 0 120
[(1+IRR100)1] = 100 120 100 =
1+ IRR 100
1
1.2= 1+IRR 100 1.2−1= IRR
100 0.2×100= IRR
IRR = 20
2.8.4 Sunk Costs
Sunk costs are costs that have already been incurred and cannot be recovered. An example of this might occur during a staged gate approach, in which the end of each phase looks back and forward in time to determine if the project should continue.
The money already spent to the date of the review are sunk costs. The participants in the gate must consider if the spent money and project state warrants a continuation of financing for the next phase of the project. If the project is canceled, the money that we invested is lost—generally with no hope of return.
The problem with sunk costs occurs during some projects when managers attempt to remove the money already expended in the updates of the IRR or ROI calculations.
This action inflates the business case by allocating the previous costs as sunk costs and resetting the calculation from this gate forward.