BT

A JIRA List Is Not A Scrum Product Backlog

Posted by Craig Larman & Tim Born on Jul 22, 2014 |

To start, this article is not an indictment of the excellent JIRA tool; we mention it merely because this widely popular tool, as others of its ilk (Trac, Redmine, VersionOne, Rally, ...) supports creating the list of all the stuff we wanted to write down, and people new to Scrum incorrectly confuse a list of stuff to do with a real Scrum Product Backlog. Thus, when a group starts Scrum adoption and we ask, “Do you have a Product Backlog?”, invariably the answer is, “Oh yes, we have our JIRA/... list!”

So, this article is about the contents of what should be in a Scrum Product Backlog, not the particular tool.

And as an example, we share a real-life story (part of a transition to Large-Scale Scrum -- LeSS) and technique of distilling an original JIRA list into a Scrum Product Backlog, from what was originally 508 JIRA items down to 23 Product Backlog items. That was a 20:1 reduction. If your group undertakes a similar distillation process and does not observe a dramatic reduction in the number of items, there’s a chance the group misunderstands what should be in a Product Backlog.

What A Product Backlog Should Look Like

A Product Backlog is an ordered list of (mostly) customer-centric features, called items (or PBIs).

Contrary to widespread misunderstanding, a Scrum Product Backlog does not contain ‘stories.’ One of the signature ideas in Scrum is that it is not prescriptive about practices (such as ‘stories’). To quote The Scrum Guide [Schwaber & Sutherland, July 2013 / Definition of Scrum]:

Scrum is not a process or a technique for building products; rather, it is a framework within which you can employ various processes and techniques. Scrum makes clear the relative efficacy of your product management and development practices so that you can improve.

And in any event, stories in agile development are a behavior of requirements by conversation (story = “Card, Conversation, and Confirmation”), not a thing in a list per se.

Therefore, in Scrum the approach to requirements and how they are expressed as PBIs can be any way that a group finds useful, and may of course evolve. Regardless of how PBIs are written, the bottom line is that it emphasizes the things of value to the users. Again to quote The Scrum Guide:

The Product Backlog lists all features, functions, requirements, enhancements...

A well-refined Product Backlog will have enough “ready” (small, well-understood, ...) work near the top of the (ordered) list to keep the Team(s) fed for at least one Sprint, although when the cycle time to properly refine ready or actionable a PBI is relatively long, you may want to aim for two Sprints worth of “ready” PBIs. Lower-ordered PBIs are less refined and often more coarse-grained, reducing the lean wastes of work-in-process and over-processing.

Therefore, even for big products, in a good Product Backlog most PBIs (those beyond the small “ready” set near the top) will be vague and coarse-grained placeholders. This makes working with even a “large” Product Backlog relatively tractable.

Initial Product Backlog Refinement Technique, Starting With a Big ‘JIRA’ List

To make things concrete, here is a real-life example of distilling a JIRA list into a Scrum Product Backlog.

To easily filter the 508 JIRA issues, we started by printing them out on paper, four per page, and cutting them up. This quickly gives us physical objects we can manipulate as we group and filter to find the PBIs.

We happened to have four people available, only one of which was familiar with the product. To accelerate learning, we had our expert guide us. To sort the signal from the noise in our initial attempt at a product backlog, we applied a novel technique to quickly bucket the items so we could focus on the real PBIs. We use a technique similar to affinity clustering to sort the JIRA issues into PBI, bugs, tasks and various other buckets as required.

  1. The initial cycle consisted of having our expert select an issue, consider it out loud and place it in a bucket (category), thus demonstrating his reasoning to us as he did the work. After a handful of issues were considered and categorized, we had some initial buckets and a vague idea of the criteria and reasoning process the product expert was using.
  2. We set a timer for five minutes while we all worked in parallel. Initially we were careful to place our issues adjacent to the buckets, rather than in them. At the end of the timebox our product expert reviewed the issues tentatively queued for each bucket. Correct assignments were promoted into the bucket, and errors were debugged out loud so we could all learn about some criteria or misunderstanding that applied.
  3. We repeated these five minute cycles and observed that each cycle our throughput went up and our error rate went down. In the end we were able to sort the entire JIRA list in about two hours, leaving us with 23 genuine Product Backlog Items.

Clearing the Fog of Development

As the bucket categories emerged and classified more issues, we noticed several interesting things. For instance, we uncovered issues in JIRA that the Team had lost track of, that should have been closed, or were otherwise orphaned.

Our all-time favorite discovery was a JIRA issue labeled, “create a JIRA issue”!

Let the issues guide your choice of buckets. In our case we had “requirements” (the real PBIs), “support requirements” (tangential issues, not actually PBIs), “tasks”, “fitnesse tasks”, “tech improvements”, “to be investigated”, “?” (complete nonsense), and “defects” (bugs).

In terms of distribution, we found that the vast majority of issues were defects (bugs). Individual tasks (presumably tasks for some Sprint) comprised the next largest bucket we found -- and Sprint tasks do not belong in a Product Backlog. Some issue were to investigate something, usually related to some apparent failure. Actual customer requirements and technical improvements were roughly the same count, with the remaining buckets even smaller.

Imagine being a Product Owner and having to wade through 50 bugs or Sprint task issues before finding something that looks like a customer requirement. While all these issues are important to someone, separation of concerns is a valuable technique to keep the Product Backlog tractable and interesting to the Product Owner.

As mentioned in the introduction, when the fog was fully cleared, the original list of 508 issues boiled down to 23 relevant Product Backlog items of customer-centric new requirements.

We have done many of these list-filtering workshops when first moving to LeSS, and a reduction down to 5 or 10% of the original list size is not uncommon. That said, there is a high standard deviation; we have also worked on LeSS transitions where 90% of the original pre-Scrum list of items were true high-quality requirements and made their way without change into the new Product Backlog.

As an example of the wide deviation, during another session where the list (nominal Product Backlog) was taken from a spreadsheet, after applying this filtering technique, the “Normal Features” bucket was relatively high, north of 20%. But that’s still a lot of noise.

In Scrum, Where to Record Product Defects?

A common question in Scrum is, “Where to record defects?” In small-scale one-team Scrum with only a few defects, recording them in the Product Backlog is a simple and sensible solution, and classic Scrum advice. But when first transitioning to LeSS, in a large product with many teams and years of pre-Scrum legacy defects building up in the JIRA list, recording them in the Product Backlog is not tenable; there are just too many.

In the case study we are describing, we discovered more than half of the total of 508 issues were defects. So as is common in such cases, our alternate solution was to keep the record of defects in JIRA.

And Where to Record Technical Improvements or Engineering Tasks?

As with the defects issue, a common question in Scrum is, “Where to record technical improvements and/or engineering tasks?” In one-team Scrum on a small product with a few improvements, and a seasoned Product Owner, keeping them in the Product Backlog is a reasonable option.

But context is king here: Once again, similar to the situation regarding defects, when (1) you are in a large-scale Scrum context and there are a ‘hundred’ miscellaneous improvements in the original ‘JIRA’ list that swamp the small set of customer requirements, and (2) there is a new Product Owner just joining from the business side who has no interest in getting drawn into “techy issues” and is only grudgingly getting involved in this new role, and is fearful of being expected to do traditional IT project manager problem solving, then you have to be sensitive to the situation in terms of making an attractive Product Backlog that speaks to the things that the novice business-side Product Owner cares about, and be careful to not “put them off” during this transition step.

So our solution was to keep the very large list of technical improvements and engineering tasks in JIRA.

A Happier Business-Side Product Owner

As you know, one of the major changes in moving to Scrum is to engage a business-side Product Owner. What you may not know is that when Scrum is introduced to the business community or Product Management (from where a Product Owner will usually come), that community of people often privately (if not publicly) voice this concern:

Why do we have to get involved in all that messy techy stuff? I’m not interested in doing project management of defects, engineering tasks, and so on. I just want to focus on the business features we need.

And that concern is exacerbated in the large-scale case transitioning to LeSS, where there aren’t 4 existing defects and engineering tasks to manage, but 304! To expose a potential new business-side Product Owner to all this ‘noise’ is very off-putting for many of them.

So we like to separate the signal from noise, and introduce them to an attractive clean Product Backlog that contains things they really care about (the new features) without all the unattractive myriad technical issues they are concerned they will be sucked into having to manage, as a traditional project manager would.

Imagine the delight of a new business-side Product Owner to discover that instead of having to manage a list of 508 issues, they only need to focus on a list of 23 small and relevant-to-them items.

The hanging issue of managing the hundreds or thousands of remaining defects and engineering tasks in a young LeSS adoption with a new business-side Product Owner is an important issue, but outside the scope of this article. Please watch for an upcoming article" on Handling Defects & Technical Improvements in Large-Scale Scrum (LeSS).

Simple Tools

Staying on the theme of attracting a new and nervous business-side Product Owner… What tool do they already know and are comfortable with? Probably not JIRA/Trac/..., but almost certainly a spreadsheet! (These days, probably a Google Docs spreadsheet). And it is the classic recommended tool for a Scrum Product Backlog. Therefore, consider keeping your practices simple and your changes simple for the new Product Owner by using a spreadsheet. We have seen LeSS adoptions involving hundreds of people and multiple sites quite effectively use a simple spreadsheet for the new Product Backlog. And be suspicious of claims that anything more complicated than a spreadsheet and wiki is needed -- even in large-scale cases.

Even simpler!... We even have Product Owners that (literally) tape one card per PBI on a wall and manage their Product Backlog that way.

The point is to keep the Product Backlog clean -- especially for the large-scale transition case with a new business-side Product Owner. Keep items of customer value in the Product Backlog and keep it accessible.

About the Authors

Craig Larman is the co-creator (with Bas Vodde) of Large-Scale Scrum (LeSS), and an organizational-design consultant for enterprise adoptions and very large-scale product development with LeSS. In addition to being one of the first Certified Scrum Trainers, since 2004, he has helped groups with LeSS adoptions, such as J.P. Morgan, Ericsson, UBS, Xerox, bwin.party, BAML, ION Trading, Valtech India, and more. He is the co- author (with Bas Vodde) of the upcoming book Large-Scale Scrum: More with LeSS, and the two existing LeSS books, Scaling Lean & Agile Development: Thinking & Organizational Tools and Practices for Scaling Lean & Agile Development: Large, Multisite, & Offshore Product Development with Large-Scale Scrum.

Tim Born is an agile coach with JPMorgan, helping to implement scrum in a large scale environment. Previously with Bell Labs, where he implemented scrum at scale for Alcatel-Lucent’s telecom products. Early experiences with scrum include being a team member on several development teams, scrum master and (fake) product owner before realizing he had morphed into an agile coach.

Hello stranger!

You need to Register an InfoQ account or or login to post comments. But there's so much more behind being registered.

Get the most out of the InfoQ experience.

Tell us what you think

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Email me replies to any of my messages in this thread

JIRA Agile? by Alonso Perez

Have you ever heard of JIRA Agile?... Move the tickets that are real user stories to JIRA Agile backlog and start working with the right tool.
Maybe JIRA Agile is not the greatest agile tool out there, but it has existed for a long time. It's worth knowing about it, IMHO...

Re: JIRA Agile? by Murali Dharan

+1 on this.

On Jira, we use a bunch of ticketing projects with Project Management workflow, but our Backlog project is a Jira Agile project with PBI tickets fully understandable by the customer, who help place them into sprints (in Jira parlance, Versions in Roadmap). Development tickets link up to these PBI tickets.

IMHO the PBI language/semantics must be preserved (and therefore not mixed up with development task tickets on issues, technical tasks etc) as the article rightly points out, but I'd rather leverage the extensibility of Jira to solve this than isolating the list elsewhere.

Color coding/tagging items in backlog by Richard Steele

Phillippe Krutchen makes a compelling argument for the exact opposite: features, bugs, technical debt, and architectural changes should all go onto the same backlog. Otherwise, it's impossible to get a complete picture of all the work that the teams need to do. See www.infoq.com/news/2010/05/what-color-backlog for more information.

Re: Color coding/tagging items in backlog by Beren Erchamion

Richard has got it. The product owner should know about all the bugs...all the dirty laundry...or as technical debt build the owner will not know. I don't know why a story can't be in the backlog either. Seems like a story is similar to a feature. No?

Re: Color coding/tagging items in backlog by Ryan Hartzog

It sounds more cumbersome and potentially problematic to keep separate lists. As Richard and Beren mention, how can you get a complete picture if you don't maintain the list(s) together in one location? From the Scrum.org Scrum Guide "The Product Backlog is an ordered list of everything that might be needed in the product and is the single source of requirements for any changes to be made to the product."

I don't dispute that keeping a single list presents challenges in a large environment, I live it every day. However, it seems counter to the goal of the product backlog to keep those items separate from the new features.

I have always understood a feature to be synonymous with a story as well Beren.

Re: Color coding/tagging items in backlog by Damien Corbishley

I had in mind that a PBI was an item that once complete would incrementally deliver value to the product. So technical debt and architecture changes don't (on their own) provide value to the user of the product. Scrum states that architecture emerges over time, so following that the architecture items wouldn't be a PBI. And technical debt - well we should all have a duty of care when we come to modifying code (be it production runnable code or automated test code), that should be factored in by the Development Team when sizing the PBIs.

With that I would agree with the article and leave the backlog smaller than a mammoth list, I believe that would allow the team to focus more when reviewing the backlog and for the "noise" to be hidden somewhat.

For the situation where you already have a list of bugs then JIRA Agile is great at that, the addition of quick filters on the JIRA Board allows anyone to remove from view these bugs leaving a clean, ordered list.

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Email me replies to any of my messages in this thread

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Email me replies to any of my messages in this thread

6 Discuss

Educational Content

General Feedback
Bugs
Advertising
Editorial
InfoQ.com and all content copyright © 2006-2014 C4Media Inc. InfoQ.com hosted at Contegix, the best ISP we've ever worked with.
Privacy policy
BT