A valid ROI calculation, according to guest author Steven Levy, includes the following factors on the benefits side:
1. Time value of money, usually at the corporate internal discount rate (often ~10%/year).
-
Assignment of returns to the proper fiscal period. You don’t start receiving returns until after the system is deployed, which means discounting for the time value of money.
-
Because few systems roll out all at once, the returns usually “ramp up” – i.e., if you expect a fully deployed system to return $10K/month, consider that in the first month after deployment it might return $1K, $3K the next, then $6K, and finally, after four months, the $10K expected.
-
User productivity goes down immediately after the deployment of a new system, as users struggle to learn it and grouse about losing the old and familiar tool or process; it takes time – say, a couple of months – to surpass previous productivity. Add this “slowdown” to the staged rollout described above.
-
Any delays in delivery push the benefits further out into the future. All ROI calculations should include a pessimistic schedule as well as an expected schedule.
-
Many technology projects don’t deliver all of the expected features in the first version. There’s no easy discount calculation for “didn’t get all the benefits,” but use your common sense. If the internal rate of return on a technology project works out to 15% on the pessimistic schedule, you can be pretty sure that in the real world it will have a negative IRR. Hosted solutions – those run by a vendor in their data center, requiring you to do little integration on your end – are the most likely to land somewhere close to the listed benefits. They also offer the least flexibility… but that’s often a good thing, since the attempt to be all things to all users is a sure sign of a troubled project.