Sridhar Kakkillaya's Blog
I am trying to put together a single framework for feature prioritization. I have read a number of articles and blogs on this topic.
A number of product managers argue that it is very easy for new product managers to get sucked into thinking that feature prioritization is the most important function of a product manager. I agree with this argument. I believe that every product manager should have a broad view and have to be on top of a number of functions. But feature prioritization is still a function of product manager and there has to be a method for performing this!!
There seem to be a number of frameworks and approaches used by different product managers.
1) Assigning weightage to different attributes and prioritizing based on the score.
2) 2×2 matrix on the lines of BCG matrix with difficult of implementation (time, cost, risk, etc.) on x-axis and business value on the y-axis.
3) Matrix based on frequency of usage by holding up usage against personas.
Reference: http://contrast.ie/blog/swiss-army-knives/
4) 2×2 matrix with difficulty of implementation on x-axis and importance on y-axis.
Reference: http://misunderstooddesignalliance.com/?p=4
5)
6)
7)
.
.
So, which is the best method? I guess it differs from company to company. While one method might work well for a company, the others might not. If you were a newbie product manager, what method would you start with? Do you start with one and then cycle through the rest as you face difficulties with a particular model?
I would love to hear your thoughts on this!
In the next blog of the series, I will outline which model I would use and why.
If you enjoyed this article, please consider sharing it!
Great questions.
Wanted to add “bang for the buck” (linked to my name) to your list of options.
Personally, I lump all of the “value vs. cost” methods together. Explicit business-case-supported “business value”, risk-mitigation “business value”, frequency of use (as a proxy for importance-to-your-users) “business value” are all basically the same, and belong on one axis. Cost and complexity are both proxies for “(opportunity) cost of doing it” and also belong on the same axis.
Different organizations, in different domains, with teams of varying skill and domain-expertise will have different opinions about the _relative_ importance of different value measures. But all of them will be of some importance to every team. Another aspect can be timing/sequencing of how you (strategically) want to roll your product out – starting with a single market (segment) and moving to others.
I’ve had success by starting with
1) Prioritize goals (e.g. Segment X first, Problem Y first)
2) Prioritize the features that support the goals from (1) within each top-level sequence.
The aggregate cost-benefits from features can “roll up” to inform the value/cost tradeoffs for the sequencing of goals – which IMHO is where you should start.
Thanks for the good post!
Scott
A 2×2 Matrix of some sort seems to be one of the most popular methods. However it’s very limited and doesn’t depict everything a product manager needs to know to effectively prioritize.
Personally, I am very user centric. Here is how I go about prioritizing the backlog:
I start with a 2×2 matrix with “Effort” on one axis and “Value to End User” as the other (I find business value a bit too vague to be used on its own – business value to whom?). We are mostly familiar with this sort of matrix and its main problem that we never get to the low value and high effort tickets, even though they are required.
To resolve that, I add a new attribute. We add how each item in the backlog fits with the pre-determined product or company strategy. By knowing the role each item plays to meet our internal strategy, we are better equipped to prioritize. e.g. an item with low value to end users, and high development effort (such as backend refactor work) can be placed high up on the priority list cause it’s the perfect strategic fit at this time.
I wouldn’t make this a three dimensional matrix, even though it sounds like I have, I’d user colour codes if i where to create a chart.
–
This matrix that i described is in its simplest form. If the product is a B2B product, then we have additional elements that we must capture. For example in additional to capturing the value to end user we should consider the value to the business that is our costumer as well, or the business’s strategic fit in additional to ours, etc.
Thanks for the great post, very thought provoking.
“A number of product managers argue that it is very easy for new product managers to get sucked into thinking that feature prioritization is the most important function of a product manager.”
That was definitely me. My current method is based on a weighting between:
1) persona problems
2) customer demand
Thanks all for the comments and pointers. Sorry about the delay in responding. I have been down with flu the last few days. All excellent suggestions!!
@Scott. I agree with you that all “value vs cost” methods should be lumped together. Its just that different methods have different variables. These variables should be selected based on relative importance for the specific company scenario. Maybe, selection of variables can be accomplished with input from the team rather than self. This way you are involving the entire team in the process. I will be happy to add “bang for the buck” to the list. Can you send me the link please? Thanks.
@Shahed. Thanks for the explanation. Your comments makes a lot of sense. However, I would say that the 3 dimensional matrix can be force-fitted into the weightage method. Do you see any issues with that?
@Jon. How do you weigh persona problems? Is it based on customer feedback, market research or just gut? What happens when the intersection of persona problems and customer demand is not generating any value for the business? How do you go about it?