In Part 1, I pointed at the use of the word SAVINGS in Marketing where using new technology to their extremes would save you money. The fallacy I pointed out there is that you actually would pay less than today but that the numbers used by Marketing point at future which you are not paying for today so would not be savings.

The fallacy I want to talk about today is the DOLLAR per UNIT metric where the cost of a technology acquisition will be calculated on the use to it’s largest extent rather than the real use in an environment.

Dollar per Unit fallacy

Basically the problem is when you calculate the dollar per unit metric, you are calculating to it’s maximum use. For IOps for example this goes to the maximum amount of IOps per array (note; this comes back in Part 3). The problem is that no-one is using any technology to use at it’s maximum capicity. I created an artificial example below;

In this case we are buying a storage array for VDI. If we look at the total IOps and calculate an average of 20 IOps per desktop (which is a wrong way of calculating btw) you would be able to have over 3000 desktops for vendor A and with a lower cost per IO and lower cost per desktop … even-though the appliance is more expensive?

However if we look at the actual use-case for an environment with 2000 desktops, these numbers go upside down! This is the reason Dollar per Unit metric is at least ‘obfuscative’.

Buying for Future Growth

In our previous example Vendor A would definitely play the “future growth” violin. I have had dozens of conversations like this and guess what: I still have to meet the first one with whom that future growth was perfectly predicted. Maybe you could read this previous blog when I used to sell EqualLogic. I truly loved EqualLogic at that point because I could buy a starting box in year 1 and in year 2 or 3 we could evaluate the usage and see if we needed extra volume or performance. Look at the picture below at what this does to the Dollar per Unit metrics for the same price result.

Here are more reasons why you DON’T buy for future growth upfront:

• buy the hardware for next year in the next year as the performance/power and so many more parameters have evolved
• software evolves so fast that chances are big you even don’t want to buy the same product next year

Summary

• The Dollar per Unit metric doesn't count for products at their largest capacity. It only counts when comparing the use-case numbers.
• Don't buy for future growth. Never make predictions, especially about the future!