Friday, April 20, 2007

You Can't Buy Success in Projects

In many projects managers try to buy success. The usual sequence is like this:

Management: We need a database. Which one should we use?
Engineers: MySQL
Management: Lets use Oracle because it is better and not a toy.
Engineers: But Oracle is expensive, slow, a memory hog, feature bloated, and vendor lockin.
Management: But Oracle is the industry accepted standard.
Engineers: It is a waste of money both in direct cost and more expensive DB admins salaries.
Management: We have decided to go with Oracle, live with.

Now why is the decision really made this way? The reason is that Management believes that paying lots of money makes Oracle more 'real'. They are trying to buy success on the project. Management also belives, but will not say so, that you the engineer are incompetent.

The reality is that successful projects are not bought, they are managed and created. So we get to the Thundefist First Law of Projects:

The key element of success on a project is Pay Attention.

Buying expensive solutions like Oracle, .Net, heavy metal servers, J2EE, ClearCase, or any other proprietary solution will more likely hurt the project rather than make it successful.

Throwing more bodies at a project is another style of trying to buy success. The worst version of this one is bringing in expensive consultants to 'fix things'. Then the Expensive Consultant leaves and unfortunately takes his brain with him.

A subtle form of trying to buy success is adding layer after layer of tracking, management, paperwork and such. This is directly counter productive.

A good manager will instead pay the real price of a successful project, which is to spend actual time on the project engaged with the people and details of the project.

There is a term for this management approach. Hard Work.

1 comment:

Anton Vishnyak said...

This is a good analysis of executives throwing money at a problem. You may also want to examine this issue as a form of risk mitigation. My point of view is that an executive looking to Oracle instead of mySQL is looking to transfer risk (a primitive form of insurance) hoping that Oracles bigger size and larger install base has allowed them to create a more resilient product.

Additionally, you may want to try the "How do you know?" technique when faced with this situation in the future.

Exec: We should use Oracle.
PM: Why?
Exec: It is more stable and a better product.
PM: How do you know?
Exec: It's umm... common knowledge
PM: I have several studies showing that mySQL will meet our requirements and cost us less.