In this post, Michael explains the various acronyms used for the artifacts necessary to develop tech products. It's a really helpful summary of the most common terms. Many of the terms refer to the same concepts but by different names. BRD, MRD, and PRD offer various details versions of requirements; FSD and PSD refer to levels of specifications. And as Michael points out, the implementation of these documents differs from organization to organization.
In my experience, companies with too many types of documents are attempting to correct the emptiness of one document type by expansion into another. "The MRDs we're getting are not detailed enough so let's introduce a PRD." Product managers who write poor market requirements are often supplemented with Requirements Analysts who write good requirements. So an MRD is actually a poorly written PRD. A poorly written FSD is supplemented with a well-written PSD.
Maybe we would have fewer roles and fewer artifacts if each were well defined and appropriate performed. You get overlap and redundancy if you have too many types of documents. Read On Reqs and Specs for more on this point.
Perhaps this expanded set of documents isn't that they're poorly written but that the writers have smashed together the artifacts from a dozen different methods. Meetings today often take on a tone of folklore as each person describes a process and set of documents from the old days, from an old company, from an old (or new) book. If you're writing SRS documents, someone has read Weigers; if you're writing stories, someone has read Beck.
A requirement is a statement of the problem to be addressed; a specification fully describes it solution. In the end, I fear that many companies are reacting to poor execution in one group by introducing rigor in a downstream document written by another group. Alas, no amount of documentation can overcome incompetence or malevolence.
Be clear on the objectives and content of each document in the process and then hold the writers to that standard.
Always looking for a large unmet need
Intuit does many things right including seeing problems that others have missed. They introduced Quicken and Quickbooks despite dozens of competitors and became the dominate player by focusing on the problems that regular people were having. And now they're entering health care billing to help people get their arms around the ridiculous morass of bills and past due notices.
In Less Insult From Injury, BusinessWeek Online's Susan Berfield writes,
They visited dozens of people in their homes to see how they handled all of the bills and statements generated by the health-care system. Their conclusions were that people would spend hours to get even a few dollars out of their insurance companies, and that they wanted some way to assert control at a time when their lives had been upended by health concerns.
As Chief Executive Steve Bennett says: "We are always looking for a large unmet need that we can solve."
Posted on August 19, 2006 at 05:52 PM in Industry News & Commentary, Market Problems | Permalink | Comments (0) | TrackBack (0)