If you shop at a warehouse club, you know the feeling well. If I buy a case of chicken noodle soup rather than one can, the unit price is significantly lower. But then the mental back and forth begins. In the long run it is cheaper, but it is an awful lot of money to spend on soup. Would I be better off paying a little more, but buying it as I need it? But if I buy it now I will have it on hand. What if prices go down after my big soup investment? What if prices go up? What if I have company over and everybody wants soup?
The same internal struggle goes on with many Information Technology investments. Whether you are building a data center, buying hardware or software, or even ordering paper clips for the office supply cabinet, more is cheaper by unit. The big question is how much do you need? Or the better question is, how much will you need? You almost need a crystal ball. You can make educated guesses, but when all is said and done, making purchases for future requirements will always be a gamble.
Is your business in growth mode, either organically or through acquisitions? Are you in the business of providing outsourcing or any type of XaaS type of services? If so, it might be worthwhile to buy in bulk and have some extra infrastructure resources on hand to absorb new business.
When purchasing a CPU from IBM, the dollars per MIPS typically decrease the higher you go in MIPS. Software costs are often the fly in the ointment when considering a CPU purchase including spare capacity. One attractive option that IBM offers is the ability to purchase a new CPU at a larger capacity setting than required at discounted rates, and then dial the capacity setting back down to fit your current needs. This is considered buying “banked capacity.” It is not uncommon for IBM to price a CPU at a calculated cost of $xxx per MIPS and offer to sell you additional banked capacity at a rate discounted by up to 30%. You can take delivery of the CPU at the lower capacity setting to save on software costs, and not activate the banked capacity until the predicted growth comes along. If you wait until you need it and try to add capacity to the CPU after the purchase, you can expect to pay more than the $xxx per MIPS price you originally paid instead of enjoying the 30% savings of an early purchase.
There are not as many nifty discounts available for peripheral devices such as tape and DASD, other than typical volume discounts where you pay less per unit as the amount you buy increases. Pre-purchasing of capacity on these devices can still bring savings in volume discounts, or provide payment and amortization options. Sometimes additional capacity will be provided at an agreed-upon rate and paid for 50% at purchase time and 50% at activation time.
Software vendors can often be less flexible than hardware vendors. A large percentage of mainframe TCO is tied up in software these days and many companies feel that they are being held hostage by software vendors. Unfortunately, the only bargaining chip software vendors sometimes respond to is volume. Larger environments get larger discounts. There is no avoiding it. This is where large operations have an advantage, and often IT outsourcing becomes an option as service providers can make the discounted volume purchases that smaller mainframe shops cannot manage.
So if you are predicting growth and can wrap some version of reality around it, it is possible to find some saving through careful planning and pre-provisioning of expandable resources. Not all acquisitions will offer the same savings potential, but it is always wise to go through the exercise and get those future requirements laid out as much as possible. Prepare for the growth and then make it happen!