If we have data, let's look at data. If all we have are opinions, let's go with mine. Jim Barksdale, CEO Netscape

Whether you're building the first version of a product or you're in the Maturity Phase, teams usually have a gigantic backlog of work. The typical backlog is full of Important items, many with passionate stakeholders. Sales needs a feature to close a big deal to hit its revenue targets. Marketing wants to build a referral campaign. Customer Success has some onboarding A/B testing they want to run, and Engineering wants to change the way they're accessing the DB to reduce hosting costs and upgrade to a new version of their front-end JS framework to stop some recurring bugs. The debate isn't whether any of these initiatives are essential. It's what is most important to tackle now.

Ideally, you've already instrumented your product(s) with an analytics framework that helps you understand your users and if you're succeeding with your goals and objectives. Analytics also enables your decision-making to be data-informed.

Data-informed decision-making has numerous benefits:

  • Decision-making is less emotional; we're talking less about opinions and more about the facts we have to support a hypothesis.
  • We're also talking about the data we don't have and whether or not we're comfortable with that. Acknowledging that we don't know everything and using learning loops to adjust our position not only makes better decisions but it also helps with team safety -- it's ok not know something 100%
  • Making our success criteria explicit in our decision-making helps align our expectations on outcomes. It creates the understanding that while the feature is implemented and deployed (AKA Done), it isn't indeed done until it meets its success criteria.
  • Most importantly, our decision-making process and criteria are transparent. WIth the data that informed our decisions, people can understand the rationale of why something is most important right now, and they also understand our goals and how we'll measure success.

Google created HEART framework after UX researchers noticed that five categories were common amongst their work:

  1. Happiness - measures of user attitudes, often collected via survey. e.g., satisfaction, perceived ease of use, and net-promoter score.
  2. Engagement - how frequently and intensely someone uses your product. e.g., visits/user/day, average session time, and the number of times a user performed a crucial task in your app.
  3. Adoption - new users of a product or feature, e.g., new signups/week, feature use/users/week.
  4. Retention - Number of users who continue to use your product during a specific timeframe.
  5. Task Success - How well did a user complete a task, how quickly, and the number of errors committed on average when completing the task.

I like it because it has a more holistic view of the product and includes both quantitative and qualitative data. It also uses a simple model of Goal-Signal -Metrics to help teams define success criteria and think about how they'll know if they've succeeded. Setting Goal-Signal-Metrics is a team sport; you want diverse opinions at the table from cross-functional teams to accurately capture the needs of your organization and gain consensus on your goals. Refer to this overview of Setting Product Goals from Teresa Torres if you're not sure where to start.

The Periodic Table of Product Prioritization Techniques

Periodic table of Product Prioritization Techniques from Folding Burritos

With your goals in place, you can now prioritize your backlog. There are many ways to prioritize a backlog. Daniel Zacarias has compiled a guide of 20 techniques over at Folding Burritos, and I'm adding one more: Prioritization with HEART.

Step 1: Classify Your Backlog

The first thing is to assign your work items (whether you're working in user stories, epics, features, etc.) to one of the HEART categories. Classify a cross-functional team as well. Invest the time to understand differences of perspective and choose the best fit. Wherever possible, avoid assigning something to multiple categories. Ask yourself what needle will this most likely move: Happiness? Retention? There are no right or wrong answers, but sometimes backing into it from the signal, and metrics might make it more evident.

Pro Tip: label/tag your backlog items with the HEART category. Prioritization is easier when you can filter for the things that will help you get to the goal. Filtering also helps surface bugs relevant bugs that may be hindering you from reaching product goals.

Step 2: Prioritize Your Goals

You'll notice that a natural prioritization of your goals will emerge when you're having the conversations to develop them. Where you are in the product lifecycle will have an impact on your objectives. If you're early on, you might focus on signing up new accounts. A product in the growth phase might have the primary aim of reducing churn by increasing retention.

Have the conversations you need to gain consensus on the most important goals. The more viewpoints, the better your chances of not missing something because of a lack of information.

Step 3: Prioritize your Backlog

For each of your priority goals, filter your backlog items. Determine what backlog items are essential now to make progress on your goals. Discussion is once again crucial for this step -- embrace opposing views. When opinions diverge, drill into what information each person has that supports that view. I've been consistently surprised that even in small companies, silos of knowledge can be profound. Often people assume we all have the same information, most of the time, we don't.

Step 4: Confirm Success Metrics

Now that you have your backlog prioritized to ensure that you have appropriate metrics in place to track your progress. It isn't necessary or advised to have metrics at the feature level unless it's critical. Existing metrics may have you covered. Resist the urge to over-instrument, start by capturing only the metrics that you'll use to measure success.

Step 5: Evaluate Success

Once your backlog items have been released, take the time to review your metrics with your team. Are they trending in the way you expected? Are you capturing the information we need? Are you making progress on your goals? Do you need to refine our assumptions based on the new information you have? Document these learnings and share them as status updates with the team. Talk to your stakeholders about what you're seeing and get their insight into what is happening. Start sharing this information right away so you can course correct and adapt as soon as possible.

Step 6: Close the Learning Loop

Review your documented learnings and adjust accordingly.

  1. Review and update assumptions.
  2. Adjust goals-signal-metrics based on new information.
  3. Re-examine your prioritization based on learnings and business changes.

Rinse and Repeat

Make these steps part of your regular product cadence. The frequency will depend on the team's delivery cadence, but I'd recommend:

  • a weekly status updates of performance that includes the assessment performed in Step 5 above.
  • a monthly review of your assumptions along with adjusting goals and priorities.

Using data-informed decision-making will not only help you make better product decisions, but it will also bring a shared understanding of what you're building and why to all areas of the company.