Project Backlog Priority Review

[richpomfret]

Following discussions on backlog 'sprawl' - need to assess process/forumula for how we prioritise stories across projects e.g. introduce a scoring system. To also evaluate how we add stories in relation to phase completion.

[Captainkirkdawson]

My suggestion had been that it was within a project. What I had been toying with was something like Legal 50-59, Security 40-49, Revenue 30-39, Researchers 20-29, Members 10-19, Maintainability 1-9. But that does not really work as there is no ability to cross group Hence is there another way?

[PatReynolds]

Do we need to do it cross-group? If all are scoring equally, we would be able to see that Project A had twenty tasks scoring over 45, with a total 90 developer story points, Project B had only six tasks, but 120 points, while Project C had fifteen tasks with 50 points. In each scrum, we'd be filtering by score (so the most important came to the top of backlog) but this would allow someone working across projects (whether salaried or volunteer) to see what they should tackle first (with the caveats of ability, dependencies, etc. etc.)

[Captainkirkdawson]

As an alternative use 4 factors then add the score for the 4 factors.Economic Benefit 0-10 (0: None; 5 Some 10: Significant), Urgency 0-10 (0: None; 1: feature affecting few; 2: Bug affecting few;3 feature affecting many: 4: Bug affecting many; 6: Maintainability of system performance; 8: Security; 10:Legal), Strategic Importance 0-10 ( 0:Does little for researcher or member; 2: Benefits few members 4: Benefits few researchers; 6:Significant benefit to management: 8: Significant Member Benefit; 10: Significant Researcher Benefit) Cost 0-10 ( 0: Extremely high effort; 10:low effort),

[PatReynolds]

A thought on Economic: this should be related to money - this could be either a removal of a threat (e.g. fine, meeting the requirements of Google so they don't take down adverts) or a promise (e.g. more visitors/more engaged visitors leads - presumed to lead to additional donations and advertising revenue, or larger donations from existing donors, or increased revenue from existing advert slots). Whole-life should be considered (e.g. a donation of £2 a month for 3 years is the same as a one-off donation of £72). If we think of the fine/extra donations as over 10 years £10,000 = 10 points, and £1,000 = 1 point, would that make sense? I can work out very roughtly how much money we get now from users, and guestimate how many new users would = 1 point on this basis.

[Captainkirkdawson]

My initial attempt at prioritization of in progress, sprint backlog and the first few product backlogs

https://docs.google.com/spreadsheets/d/1mFWZuWKRx7C-yTVwxRjTogfhAh4Yy9IGc7_OwDC2zx0/edit#gid=1458359714

[Captainkirkdawson]

@PatReynolds sounds good to me

[PatReynolds]

To cut down numbers ... could components of Epics be left out (if essential to achieve the epic). The epic itself and the 'Nice to have' components would be individually rated.

[Captainkirkdawson]

The issues/feature stories are available as excel for download https://drive.google.com/drive/u/1/folders/0BykgQKwJtk6faHhUWnRsWklNWWc or as a google sheet https://docs.google.com/spreadsheets/d/1FO6lh2lAX7IhUIu3K3IBrdwsvmfoAI4NTfpP5VJCB5Q/edit#gid=196723914

[smrr723]

I like Kirk's 4-Factors Score method above, and can't see any issues with it.

For cross-project scoring, it would take a bit of tweaking but you could give each project an importance weighting in which all internal project scores would be multiplied by their weighting factor, i.e. FR: 1, FC2: 0.8, FC1: 0.5, FBMD: 0.4. Just have two columns for internal project scores, and calculated cross-project scores. The weighting factors should probably be based on the amount of traffic each project gets.

[PatReynolds]

I like that, @smrr723 Urgency and Cost would be the same across projects, I think. But Strategic and Economic benefit is related to number of people affected.

[PatReynolds]

Any thoughts on Cost (currently doing the UCF epic, which has a total effort of 10. As waffle suggests that an epic over 100 should be more than one epic, can we count 1 -5 as 10 6-12 as 9 13-20 as 8 21-30 as 7 31-40 as 6 and so on and so forth to 81-90 as 1, and 91-100 as 0

(Sorry for forgetting to invert the count!)

[Captainkirkdawson]

I would prefer the waffle/Zenhub rating breakpoints for effort 0 is 10 1 is 9 2 is 8 3 is 7 5 is 6 8 is 5 13 is 4 21 (20) is 3 40 is 2 100 is 1

[Captainkirkdawson]

@richpomfret @PatReynolds as you know we use waffle as our sprint tool and it has served us well. It has 1 major deficiency that that is the ability to sort by size. There is a Github tool that is extremely similar called Zenhub that has the ability to sort by size and other measures. It is designed 100% for agile management and has some interesting documentation and ideas. Unfortunately it does not have a priority field by issue. It uses Releases (our Milestones) as well as short term milestones. The software is free for open source repos. So I signed us up to have a kick at it. There would be some effort and learning to use it but it might be worth the investment. I have invited the pair of you to look it over. Can easily delete if of no interest. There is a chrome/firefox extension that allows direct integration of github access to Zenhub

[Captainkirkdawson]

We could consider adding milestone completion as part of the Strategic Importance.

Strategic Importance 0-10 ( 10: Past Milestone; 9: NS Benefits many users; 8: Current Milestone; 7: NS Significant benefit to management: 6: Next milestone; 5: NS Benefits number of users; 4: Next but one milestone; 3: NS Benefits few users; 2: 2 milestones ahead; 1:NS Benefits one user) NS is Non Specific

[Captainkirkdawson]

If we accepted what is suggested for Strategic Importance we would have to review all 1.5 and remove those that do not directly related to the milestone and either push them into a future milestone or say they are NS. We probably should do that anyway!

[PatReynolds]

Thanks for the invite. Can you remind me how to invert the colours (if it supports this).

[PatReynolds]

That makes it easier, in some ways, When calculating an epic from componets, always up, always down or split mid-way (e.g. 10 points is 8, rather than 19).

[smrr723]

Worth giving a mention to JIRA https://www.atlassian.com/software/jira - if we are considering migrating from waffle. I've never used it but I believe it's the most used tool in this niche and they seem to have priority and sizing for issues.

They do have a free tier which charities / non-profits can apply for.

[Captainkirkdawson]

Summary of ranking proposal to date. Economic Benefit 0-10 (0: None; 5 Some 10: Significant) Pat to estimate if non zero Urgency 0-10 (0: None; 1: feature affecting few; 2: Bug affecting few;3 feature affecting many: 4: Bug affecting many; 6: Maintainability of system performance; 8: Security; 10:Legal) Strategic Importance 0-10 (10: Past Milestone; 9: NS Benefits many users; 8: Current Milestone; 7: NS Significant benefit to management: 6: Next milestone; 5: NS Benefits number of users; 4: Next but one milestone; 3: NS Benefits few users; 2: 2 milestones ahead; 1: NS Benefits one user) NS is Non Specific Cost 0-10 (0 is 10; 1 is 9; 2 is 8; 3 is 7; 5 is 6; 8 is 5; 13 is 4; 21 (20) is 3; 40 is 2; 100 is 1) Using size estimate

I am currently attempting to use these factors in reviewing 1.4. 1.5 and NS issues. There are well over 100!!

[Captainkirkdawson]

Have added P1 - P25 labels to FR. These are based on the priority calculation

[DeniseColbert]

@richpomfret can this be closed?