Gathering customer feedback, turning feedback into actionable inputs, and then prioritizing those inputs for your development team are some of the most important functions of a PM. They’re also the most challenging. Every PM has their way of balancing and prioritizing their products backlog, but we want to give you some additional tips and techniques that have helped us along the way. In part 1 of this series we heard from Jason Wong, Principal PM from Jira Software.

In this blog we’ll sit down with Sherif Mansour, Principal PM from Confluence and get his take.

Backlog grooming at Atlassian
Meet Sherif, Confluence Principal Product Manager:

Sherif has 12 years of experience in software development. As the Principal PM for Confluence he’s in charge of setting product vision, prioritizing product feedback, and helping  the whole Confluence PM team execute the company’s goals for Confluence.

Sherif, what channels do you use to collect customer feedback?

We’re quite fortunate at Atlassian in that we never lack feedback on any topic! The challenges we’ve faced in the past have centered on reconciling customer needs and our company’s business model and product vision.

For example, at Atlassian, we’re focused on low cost, high volume customer acquisition. With this in mind, we take a volume approach to feedback. For the Confluence team, I’ve counted 15 channels we have customer feedback. Everything from in-product net promoter scorecustomer interviews, contextual and in-app surveys, and testing workshops to our online forums and feature request trackers.

Our wide range of feedback inputs is helpful for volume and diversity. For example, specific channels are great for feedback from one persona. For instance, our public feature request trackers are used mostly by our product administrators. Contrast that with our in-app surveys where we hear from end-users directly. In the end, our volume-based feedback approach matches our volume-driven business model. The same thinking should apply to most other product teams: think about how your business model should influence your customer feedback strategies. In contrast to what we have at Atlassian, if your product or service has a smaller amount of high-value customers, your feedback process will probably have fewer channels and be more high-touch.

How do you determine where feedback gets placed in the backlog?

Backlog grooming tips

In terms of the overall approach, some take the “art” approach: heavily influencing their backlog priority by vision, emotion, and experience. Some take the “science” approach: heavily metrics driven, living in spreadsheets with formulas and weighting scores for different features to help indicate priority.

It’s not art or science, however. It’s both. Depending on where you are in your product and feature lifecycle, your backlog grooming techniques will vary. You always need to be balancing both sides  to build a great product.

As a general approach, I’m less concerned about specific item prioritization and more concerned about the overall goal of the sprint and how it fits into the problems your team has set to solve. Your team needs to have a shared understanding of the customer value each sprint will deliver (or enable you to). Some teams incorporate this in their sprint goals. If you’ve agreed on the value and the outcome, the individual backlog prioritisation should fall into place.

 Pro tip: Create a meeting notes page in Confluence for your sprint planning meetings. You can use the page to track sprint goals, and give the entire team visibility into what will be delivered and how that outcome contributes to providing customer value.

How do you use Jira Software and Confluence for backlog grooming and prioritization?

Jason and I shared a bit of our process in a webinar we did earlier this year.

We’re using Confluence a lot for high-level research and vision alignment. Below is a Confluence page that we call the “project poster.” It’s important for helping us define the overall project strategy, goals, success criteria, and building a shared understanding across the team. A lot of times our project poster will be our gut-check when we’re prioritizing features in the backlog.


We also use Confluence for sharing links of ideas and resources, blogging to communicate to stakeholders or teams, quick roadmapping, and requirements definition.

As Jason mentioned, we also convert a lot of requirements in Confluence directly into our Jira Software backlog.


Any tips for backlog grooming?

There has been a lot written about the value of backlog grooming. I’ve struggled with this myself and so have many of our teams at Atlassian and customers we speak to. Yes, it’s a never-ending stream of work – kind of like laundry. But also like laundry, there’s not much point in doing it every day.

If you’re grooming your backlog a lot, you’re probably doing it wrong.

The best technique that I’ve seen work well in most teams is pretty simple – the backlog is broken in three broad categories:

  1. Now (usually current sprint)
  2. Next (the upcoming sprint)
  3. Someday

The rough approach and principles:

  • Don’t spend a lot of time grooming or planning the someday list.
  • Agonise over the priorities of items in your now and next list.
  • Set a time interval (e.g. monthly) to prioritize the top of your someday list (not necessarily a complete groom). It’s super critical to timebox this time. Time-boxing forces you to be ruthless with prioritization discussions.
  • Every second or third groom bulk-close all the issues that you haven’t touched. If it’s important enough it will bubble up again.

By the way, if your sprints are super short (e.g. weekly) you might change the definition of this to have a slightly greater amount than the current (now) and upcoming sprints (next).

Now it’s your turn to weigh in!

As you’ve seen from both Sherif and Jason, delivering value to customers is one of the most important things to a PM. How you go about collecting feedback, and prioritizing that feedback to deliver value is completely dependent on your personal, company, and team processes. Just remember to take advantage of your unique feedback mechanisms, build a shared understanding amongst your team, and don’t spend too much time grooming your backlog πŸ˜‰

So, now that you know how we do it, how do you collect and prioritize customer feedback? Please weigh in via comments! And if you’ve found this post useful, don’t forget to share it on your social network of choice so your fellow PMs can learn too.

Are you ready to start managing and prioritizing customer feedback with Jira Software and Confluence?

Start a trial now

PMs weigh in: prioritizing customer feedback and backlog grooming