Friday, April 11, 2008

11 Rules for Prioritizing Features in a Customer-Oriented Organization

Name: 11 Rules for Prioritizing Features in a Customer-Oriented Organization
Type: Rules
Status: final
Version: 2008-04-11

Gist: While prioritizing features, people people quite often take a ranking approach in order to balance different stakeholder's perspectives. This really may prevent satisfied stakeholders (or a shot time to market), and may lead to a mediocre release plan. Why? The math used in most ranking approaches is wrong. Here are 11 Rules for proper prioritization.

Source:, thanks to Scott Selhorst for yet another great idea. Go there if you need a more visual explanation. I added my bits and pieces.

E1: You have a list of features gathered from various stakeholders.

E2: Your stakeholders are of different importance. ' maybe one represents the customer of your product, another is development/manufacturing, yet another is some governing technical department. Read about Finding Key Stakeholders.

E3: Each stakeholder has ranked all features, with smaller numbers for the more important ones.

R0: Make sure you have a proper understanding of the goals (read Specifying Goals and Decomposing Goals) of the project or product. Make sure you understand your set of stakeholders in the light of the goals.

R1: Ignoring E2 above, the simplest approach would be to add the ranks for each feature and - voila - the features that are ranked best on average win.

R2: Not ignoring E2, don't make the mistake to give each stakeholder a weight and multiply the respective ranks with the weight before you sum the rank-numbers. This only works if you have lucky numbers.

R3: You can do like R2 suggests, but then you have to reverse the ranking numbers of E3, giving large numbers to the more important features.

R4: You should think hard what the stakeholder's weights represent. If you have one stakeholder that is more important than the others, it's paramount to make him happy, at least not to upset him. Do this by delivering his important features first, and only after that you think of features for the other stakeholders.
Note: Take the example from E2. I bet the purpose of your business is to make the customer happy.

R5: Make sure you do a Kano analysis in order to find features nobody thought of in the first place. This may be your advantage over competitors.

R6: If confronted with a large (> 20 items) list of features, let the stakeholders break it down into importance classes first. It's virtually impossible for the average stakeholder to truely rank so many items.

R7: Don't give in if your stakeholders say they need to know the costs before they can rank the features. You want to know what they want to have! You can always work out things to be cheaper, especially if your features are real (business) requirements, not solutions disguised as requirements. Bear in mind: "If you don't know the value of a feature, it does not make sense to ask what it costs ". (Tom DeMarco)

R8: If your features are Use Cases, go here.

R9: if all the above does seem to simple, revert to Karl Wiegers (First Things First), Alan Davis (The Art of Requirements Triage), Don Firesmith (Prioritizing Requirements), Lena Karlsson (An Experiment on Exhaustive Pair-Wise
Comparisons versus Planning Game Partitioning
), or QFD.

R10: Read Mike Cohn (Agile Estimating and Planning).


Tom Gilb said...

Here are my views on priority!
Choice and P…
Mng Priorities

Rolf said...

Thanks Tom,

I incorporated your useful material into the respective wiki-page