<-- back to learn page

October 7, 2024

 - published by 

Robert Briese

6 Tips to improve your Product Backlog refinement

6 Tips zur Verbesserung des Product Backlog Refinements

When it comes to successful product development, there’s one technique that stands out: refining the work you need to do to deliver real value. In Scrum, this is known as Product Backlog refinement, and it’s a game-changer. Think of it as the process where big ideas get shaped into clear, actionable tasks that can be tackled within a Sprint. According to the Scrum Guide, refinement is “the act of breaking down and further defining Product Backlog items into smaller more precise items.”. This isn’t a one-time thing—it's an ongoing practice that happens throughout a Sprint. The goal? To ensure you're always moving in the right direction. But refinement is more than just breaking tasks down. It's about adding enough detail so that when the work is done, you know it’s done right. This includes setting acceptance criteria, estimating effort, and sometimes even prioritizing what comes next. It’s all about setting your team up for success, step by step.

As an Agile Coach, whenever I join new teams, I notice a common pattern during most Product Backlog refinement sessions: 

  • The Product Owner (PO) typically kicks things off by welcoming the team—a diverse group of product developers—and sometimes gives a brief overview of the items on the agenda. 
  • Then, they open up a Product Backlog management tool, like Jira or Azure DevOps, and start walking through the items they’ve created. Sometimes, the PO has created these items alone; other times, the business analyst or designers have contributed as well.
  • After each item is introduced, there’s space for the team to ask questions and clarify any uncertainties. A few team members raise their hands, ask their questions, and the answers are documented in the Product Backlog tool. 
  • Once the discussion wraps up, the PO asks the team to size the item—usually in Story Points or a similar estimation method. If the estimates vary widely, the team dives deeper, adding more clarity to the item or adjusting its scope until everyone aligns on the estimation. 
  • Sometimes, a "Definition of Ready" checklist is consulted to ensure nothing important—like acceptance criteria or additional tasks—has been missed before moving on to the next item. 
  • After all the prepared items are covered, the meeting concludes.

On the surface, this process seems efficient. The Product Backlog gets updated, and the team leaves with a list of items ready for the next Sprint Planning session. However, I believe that this approach often misses key opportunities for a truly effective Product Backlog refinement. In this article, I’ll share a few strategies that can elevate your refinement sessions, ensuring you unlock the full potential of this critical activity:

  1. Make Product Backlog refinement a formal event, occuring once or several times per Sprint.
  2. Have the Product Owner prepare and guide the structure of the event to keep things focused.
  3. Incorporate learnings from the most recent Sprint Review to focus on what’s relevant.
  4. Involve real users in the refinement process to ensure you're building what truly matters.
  5. Run multiple diverge-and-merge cycles, keeping product developers fully engaged in discussions and refinements.
  6. Educate developers in effective refinement techniques to enhance the quality of the results.

Let’s dive into each of these ideas and explore how they can supercharge your Product Backlog refinement process:

1. Make Product Backlog refinement a formal event, occuring once or several times per Sprint

This approach is particularly effective when multiple teams are working on the same product. If refinement is done informally during the Sprint, teams tend to focus only on items related to their current work, which they’ll likely select for the next Sprint. While this may seem efficient, it limits knowledge-sharing between teams and reduces the flexibility to redistribute work as needed. In agile organizations, adaptability is key. Teams should be able to help each other and take on tasks that may not be tied to their previous Sprint but are more valuable or higher-priority in the Product Backlog.

For different teams to pick the most valuable items, it’s essential that all teams are familiar with those key items. This is why frameworks like LeSS (Large-Scale Scrum) recommend holding Product Backlog refinement as a cross-team event. It encourages shared learning and opens up coordination opportunities between teams. One practical way to ensure teams have a broad understanding of the most important backlog items is to create mixed groups of developers from different teams to refine the items together.

Another significant reason for scheduling Product Backlog refinement as a formal event, as suggested by LeSS, is to consolidate non-Sprint related work into a timeboxed session. This frees up developers to focus exclusively on Sprint-related tasks during the rest of the Sprint. The key here is that the event should be strictly timeboxed—usually no more than 10% of the Sprint duration, which means about one day in a two-week Sprint. And importantly, you only need to hold the refinement session if there’s an insufficient number of detailed backlog items to last through the next two or three Sprints.

2. Have the Product Owner prepare and guide the structure of the event 

As I mentioned earlier, the key goals of Product Backlog refinement is twofold: (1) to ensure there are enough refined items in the backlog so that teams can move through Sprint Planning quickly, selecting the most important tasks from the top of the backlog without needing further clarification, and (2) to effectively distribute the most critical work across all teams working on the product. By conducting Product Backlog refinement as a shared event for all teams, as highlighted in tip number one, you’re already helping to spread knowledge across multiple teams. Since the Product Owner is responsible for maximizing the product’s return on investment, they are the best person to determine which items should be prioritized for the next Sprint and whether it’s necessary for all teams to work on a single topic or to proceed in parallel on different topics. In the first scenario, the PO may want to coordinate the refinement so that mixed groups from all teams break down a single topic into smaller tasks that can then be distributed among the teams. In the second scenario, the PO may organize the refinement such that mixed groups from some teams focus on one topic, while other groups tackle different topics in parallel.

Generally, I recommend that the Agile Coach or Scrum Master collaborate with the PO to prepare for Product Backlog refinement in advance by reviewing the current state of the backlog, considering the outcomes of the last Sprint Review (more on this in the next idea), and identifying additional business or user problems and requests to be discussed. It can also be beneficial to involve a few product developers in the preparation phase to provide input on technical aspects the PO might overlook. The preparation should result in a clear agenda with topics and timeframes for the event. One agenda item could be to gather and add any new topics (such as issues or requests) to the list of items for refinement, but the responsibility for leading the event should remain with the PO, with support from the Scrum Master or Agile Coach.

3. Incorporate learnings from the most recent Sprint Review

Agile product development is frequently equated with iterative and incremental approaches. However, if we merely take a large requirement or project, break it down once into smaller pieces, and implement those pieces iteratively and incrementally, we miss the true essence of agile product development. In such a scenario, we might as well have used the traditional waterfall model, potentially achieving even greater efficiency. The true strength of agile lies in its capacity to navigate the uncertainty inherent in complex product development, where our understanding of what needs to be delivered can evolve over time.

Tim Harford, in his book Adapt, references Russian engineer Peter Palchinsky, who highlighted three unpredictable dimensions—local, time, and human factors—that can lead us to misunderstand what customers really want. By embracing principles such as variation (exploring new ideas and testing them), survivability (operating at a scale where failures are manageable), and selection (gathering feedback and learning from it), we can take small steps to iterate and determine whether we're heading in the right direction.

As I noted in a previous article, one effective method for assessing whether we’re delivering the right product is the Sprint Review in Scrum. However, it’s crucial to dedicate a significant portion of the Sprint Review to gathering essential feedback that will inform our next steps. Unless we extend the time allocated for the Sprint Review to adequately sort, filter, and evaluate all feedback in detail, we need to postpone this important task. I believe that the ideal time for this is during the Product Backlog Refinement event, where we can intentionally discuss the work planned for future Sprints. This event provides an invaluable opportunity for the team to assess customer feedback, grasp customer needs, and clarify the purpose of the product.

4. Involve real users in the refinement process

Who better to explain the real problems and needs a product should solve than the actual users themselves? If you're developing an internal tool, such as a marketing or sales platform, the people who will use the product are the best to clarify functionality requirements and, more importantly, to discuss the underlying purpose of these features. Including these individuals in your Product Backlog Refinement (PBR) sessions could be one of the most impactful steps you take to improve developers' understanding of the product's true needs.

When developing a mass-market product, direct access to users may not be as simple. In some cases, your product developers may be part of the target audience, which allows their perspectives to be integrated naturally. By creating mixed groups of developers from different teams during PBR, you can incorporate diverse end-user viewpoints into the refinement process. However, for mass-market products where developers may not fully represent key personas, you have several alternatives:

  1. Create a customer group that accurately represents your target personas and invite different members of this group to participate in the PBR sessions.
  2. Involve marketing or business team members who have deep knowledge of customer needs and can act as proxies for important personas. This might also include customer support staff, who are on the front lines of customer feedback and requests.
  3. Leverage UX experts within your product development teams, who can represent the desires and expectations of your most critical user personas.

While bringing real users into PBR can yield highly valuable feedback and feature suggestions, it can also lead to conflicting opinions. In these cases, it's essential that the Product Owner (PO) remains involved in decision-making, as they are ultimately responsible for maximizing the return on investment (ROI) for the product.

Additionally, it’s critical that the entire discussion during your PBR events is conducted in the language of the customer. If you notice the conversation becoming overly technical—using jargon the customer wouldn't understand—this is a red flag. It may indicate you're working with component teams rather than true end-to-end (or feature) teams, which are necessary to deliver complete value to customers. In such situations, the value of real Product Backlog refinement as described here may be compromised, and an organizational change might be needed before you can fully benefit from the approach and ideas presented.

5. Run multiple diverge-and-merge cycles, keeping product developers fully engaged in discussions and refinements

As mentioned earlier, the primary goal of Product Backlog refinement is to compile a list of work items for the team(s) to tackle during the next iteration, aiming to deliver maximum value or uncover what needs to be delivered to create that value. These items must be detailed and clear enough that the developers who will implement them understand roughly what needs to be done and how long it might take to complete the work in the upcoming Sprint. Some call this having the items "ready” for the next Sprint. From my perspective, achieving this readiness is truly an art! It involves finding the right balance between over-discussion, which can lead to excessive detail, and not diving deep enough into understanding what needs to be done. Over-refining an item could result in waste if the Product Owner (PO) decides the team no longer needs to work on it (which is always possible). On the other hand, under-refining can lead to wasting time by building the wrong thing, or taking much longer to complete the Sprint than originally planned, leaving insufficient time for other work.

The best way to ensure understanding isn’t just through reading or hearing but through meaningful discussion with others. One highly effective technique for refining items is “Specification by Example” by Gojko Adzic, where you articulate clear examples of how you will know that what you are building is exactly what is needed. This method can be done using Gherkin syntax, following the Given-When-Then format, or through tables that document the start and end state along with the action to be performed. By creating small, cross-functional groups of developers with different skill sets (such as testing, front-end, back-end development, business analysis, and design), and having them collaboratively discuss and document examples, you ensure that a wide range of perspectives is captured. Additionally, this approach reinforces everyone’s understanding of why the item is being implemented in the first place.

I recommend forming several smaller groups (each with 4-6 people) to work on refining the items. This way, (1) you incorporate diverse skills, viewpoints, and insights into the discussions, and (2) you avoid having groups that are too large, ensuring everyone participates actively. If multiple teams are working on the same product, it’s even more beneficial to mix team members from different teams (as outlined under idea number 2). This cross-team knowledge-sharing allows multiple teams to confidently pick up items during Sprint Planning, as each team has developers who are already familiar with them.

In some cases, it may be helpful to have different groups refine the same item for a timeboxed period of 30 to 40 minutes, followed by a session where each group shares their results. This process can reveal different approaches, enhance learning, and lead to a better understanding of both the problem to solve and the best way to address it. The sharing of insights can be done by having a representative from each group present to the entire team, or through group rotations, where one person stays behind at each "booth" to explain the group's findings to others.

By creating small, diverse groups of product developers, refining items within timeboxed periods, and then sharing insights, you ensure that developers remain fully engaged throughout the refinement process. This fosters a deeper understanding of what and why something needs to be developed, aligning the team with the ultimate goal of meeting customer needs—the true purpose of refinement.

6. Educate developers in effective refinement techniques

This is, without a doubt, a vast and crucial topic. If your Product Backlog refinement isn't done effectively, you could be wasting valuable time instead of leveraging it to drive success in your product development activities. It's essential that everyone involved in the product development process is educated in advanced techniques. This goes beyond just breaking down large requirements into smaller ones, but involves defining clear acceptance criteria, properly sizing and ranking items, and excelling at product discovery—all of which are critical for success.

I highly recommend diving deeper into some key resources: Inspired by Marty Cagen, Impact Mapping by Gojko Adzic, the Product Backlog & Product Backlog Refinement guides from Large-Scale Scrum by Craig Larman & Bas Vodde, and of course, the previously mentioned Specification by Example by Gojko Adzic.

Final words 

You may have noticed that the approaches discussed here don’t necessarily prioritize work efficiency. There may be more time-efficient ways for a single developer to grasp a requirement—such as simply receiving it from the Product Owner (PO) or another stakeholder. If the PO spends time understanding user problems and writing clear requirements, it could indeed be faster for developers to implement them without prolonged discussions. However, I strongly believe that while this might save time upfront, it often leads to a greater waste of time later by implementing features that aren’t truly needed by the customer or missing opportunities to deliver a better solution.

By encouraging product developers to spend more time thoroughly understanding customer problems and needs, I’m confident that you will enhance overall effectiveness by delivering the right things in the right way.

I’d love to hear your thoughts on this and any other suggestions you might have on improving Product Backlog refinement efforts.

Robert Briese

Robert is the CEO of Lean Sherpas and one of 25 certified LeSS Trainers globally and one of the few with hands-on experience in LeSS and LeSS Huge. As a change agent for 15+ years, consulting and coaching, he has worked with over 30 companies, including DAX-listed global brands such as Adidas, BMW, BP, Dr. Oetker, Henkel, Hilti, Hugo Boss, SAP, Volkswagen, and ZF.

From 2015 to 2016 he led a LeSS adoption at a Global Player in the Software Industry, scaling development from 3 to 7 teams and coaching the development and the management team on how to lead a scaled agile organization. In 2018 he contributed to significant improvements in product delivery at one of the biggest LeSS Huge adoptions in the world at BMW AG. Working with over 30 teams, including large groups of over 400 employees, he led them to deliver bi-weekly product increments consisting of hardware and software.

Robert Briese - LeSS Trainer LinkedIn profile.
MockUp of the Lean Sherpas Newsletter on a Galaxy S21.

Join the LeSSons learned Newsletter!

Early access to new content, real-world examples, and funny stories!

More content you might like

Published on 
February 21, 2022
Post by 
Konstantin Ribel

Watch Konstantin Ribel discuss organizational behavior and performance and how LeSS can help manage organizational complexity. The orginal Video is viewable on the TC LeSS YouTube Channel. The original video can be seen on the TC LeSS YouTube Channel.

Published on 
October 21, 2023
Post by 
Konstantin Ribel

Konstantin Ribel and Craig Larman discuss how BMW's driverless driving team was built based on LeSS Huge, and the successes and problems encountered along the way.‍About the conferenceLeSS attempts to define what is just necessary for large-scale development. It avoids adding additional roles