Definitions Must Be Localized
I am using a project management system that forces me to define feature requests as either High, Medium or Low priority. Has anyone got a good set of definitions that clarifies what those priorities should mean?
They mean whatever your organization wants them to mean. How any given organization prioritizes features or tasks is based on a variety of factors, including (but not limited to):
- ROI (Return on Investment)
- Marketing or public relations concerns
- "Squeaky wheel" customers
- Opportunity costs
In other words, your organization must develop a scoring system or a set of sliders that allow you to filter and prioritize features in a way that makes sense for your specific business.
Coarse-Grained Priority Buckets Suck
Even with the foregoing in mind, large priority buckets are a terrible idea. There's a reason that frameworks like Scrum require ordinal prioritization. Quite simply, if something is "most important," by definition everything else is subordinate to that objective.
If you sort into coarse-grained priority buckets, you end up with situations where you might have 37 "high priority" features. Not everything can be the top priority simultaneously, although it's common Bad Management Practice™ to think so.
Unless you have secondary filters for each of your priority buckets, your project is likely to multitask itself to death. Even if you don't, will you:
- Work through buckets on a first-in, first-out basis?
- Work on the low-hanging (and potentially lower-value) fruit first?
- Will you work on the hardest (and potentially schedule-destroying) stuff first?
In addition, unless you put a lid on your high-priority bucket you will never get to the medium or low priority stuff. Ever.
A More Sequential Approach
Generally, coarse-grained buckets are acceptable as a first pass, but you should strongly consider a second pass with some kind of ordinal prioritization queue. Mountain Goat Software has some nice web-based tools for helping teams use Theme Screening, Theme Scoring, Relative Weighting, and Project Success Sliders to help the stakeholders give truly sequential priorities to features.
Always remember CodeGnome's First Law of Prioritization:
If everything is a "top priority" on your project, run (do not walk!) to the nearest exit.
In my experience, undifferentiated "top priority" queues are generally the main source of project failures and death marches. Don't be a statistic.