Journal: News & Comment

Overbudget, and it doesn't work

Permalinks to this entry: individual page or in monthly context. For more material from my journal, visit my home page or the archive.

Bob Cringely has two new articles outlining the hazards when big organizations outsource IT—specifically focusing on EDS (Electronic Data Systems) and the U.S. Navy, which is a situation that isn't working out too well. He identifies a big problem with underbidding from the suppliers, and under-specified requests for proposal (RFPs) from the customers:

[The] amazing disconnect on the part of most bidders between their actions and the health of their company [is] that "get the contract at any cost" attitude. Some bidders just try to break even, figuring they'll make their profit on moves and changes not covered in the original contract. Some don't even think that far, figuring they'll somehow make it up on volume.

Bob's further comments:

Most outsourcing contracts don't live up to their service promises and the only ones that live up to their pricing and profitability promises are those that have an artificially-low labor component...

...there is no way Wal-Mart would entrust its IT services to an outside contractor or even to several outside contractors. Doing so would threaten the entire organization. If costs are out of control and services are inconsistent, that's something to be dealt with internally, not by hoping some outside organization is smarter or more disciplined.

So would it ever work? Well...

[Most modern outsourcing arrangments] aren't contracts to bring new technology to organizations that lack the internal capability to provide that technology for themselves. They are pure cost-saving plays, or at least that's what they are intended to be.

So, according to Bob, outsourcing that works needs to:

1. Bring in useful new technology that can't reasonably be provided internally.
2. Have contracts that are priced realistically.
3. Define the scope of fixed-price work, to avoid getting trapped like EDS, who "blindly signed-on to support at a fixed price an estimated 2,000 applications only to later find the actual number was 30,000."

Yikes.

 |