Learn why starting with problems instead of features leads to better product decisions, with real examples from user conversations
Your backlog has 47 features. Sales is demanding to tick the AI box. Engineering is pushing for a search refactor. The CEO just forwarded an announcement from your biggest competitor with "Thoughts???"
Meanwhile, you're puzzling over two seemingly critical features, trying to decide which one to build first.
Sound familiar?
Most product teams spend 70% of their time building features that users rarely use. Here's why starting with problems, not features, changes everything.
Here's how it usually goes: You put two features side-by-side. You create a pros and cons list. You might even survey users asking "Would you prefer Feature A or Feature B?"
But you're missing the boat. You've fallen into the feature factory solution trap without a deeper understanding of the actual problems.
When I set out to understand how users make feature trade-offs, I made a deliberate choice: Don't mention features at all. Using Rooost, I asked two of my personas about their workflow frustrations. Their responses revealed why the "Feature A vs Feature B" framework is fundamentally flawed.
Instead of asking "Do you want AI features or better search?", I opened Rooost and asked my Head of Product persona:
"Walk me through the most frustrating part of your workflow today. What's slowing you down?"
The response was eye-opening:
"I find myself getting stuck in a few key areas. First, the lack of dedicated UX research resources means we often rely on assumptions more than I'd like. I spend a lot of time trying to gather and interpret user feedback with the team, but without structured processes, it's tough."
Notice what didn't come up? She didn't ask for AI. She didn't mention search. She talked about the meta-problem: making decisions without user insights.
When I dug deeper with "You mentioned spending time gathering feedback. Tell me more about what makes this so hard," she revealed:
"Without a centralized repository for feedback, we have to sift through disparate data sources to find actionable insights, which is incredibly time-consuming... This lack of validation not only delays decisions but can also lead to prioritizing the wrong features."
Here's where it gets interesting. When I asked the same question to my Solo Founder persona, I got a completely different perspective:
"What tasks take way longer than they should?"
"Coding new features takes forever because without direct user feedback, I sometimes go down a rabbit hole, implementing features based on what I think users want... The lack of validation beforehand means I'm essentially guessing, which isn't efficient."
Same core problem (lack of user validation), completely different manifestation.
This is where the magic happens. Once you understand the real problems, you can evaluate potential solutions through a completely different lens:
Instead of comparing features, evaluate problems:
1. Problem Frequency x Impact
2. Current Cost of the Problem
3. Ripple Effects When I asked about the impact of their current problems:
Head of Product described organization-wide pain:
"Sales and marketing teams also feel the impact... Customer support faces increased pressure from dissatisfied users... It can slow down our ability to innovate and respond to market changes."
Solo Founder focused on existential risk:
"Every hour I spend coding a feature based on a guess is an hour not spent on something that could definitively add value... Guesswork leading to irrelevant features can tarnish my brand's reputation."
4. Business Value of Solving It This is the question that separates good PMs from great ones. When I asked what solving these problems would enable:
Head of Product:
"It would allow us to make data-driven decisions with greater speed and confidence... enable us to be more proactive rather than reactive."
Solo Founder:
"It would significantly boost my confidence... allow me to iterate faster and more effectively... enhance my ability to communicate with investors."
Notice how both personas naturally described the business value without being asked about ROI directly. When you identify real problems, the value becomes self-evident.
Here's what's fascinating: Both personas essentially face the same core problem - making product decisions without validated user insights. But the solution might look completely different:
This is why "Feature A vs Feature B" is the wrong question. The right question is: "Which user problem creates the most value for our business to solve?"
Ready to escape the feature factory? Here's exactly how to run problem discovery with your users (or personas):
"Walk me through your typical workday. Where do you get stuck?"
"What tasks take way longer than they should?"
"What workarounds have you created?"
"How much time/money/energy does this cost you?"
"What would solving this enable you to do?"
Once you have problem clarity, here's how to decide what to build:
Problem | Frequency | Impact | Current Cost | Business Value | Our Ability to Solve |
---|---|---|---|---|---|
Manual feedback synthesis | Daily | High | 5-10 hrs/week | Faster decisions, less churn | High |
Feature validation guessing | Every sprint | Critical | 30% dev time | Product-market fit | Medium |
Don't fill this matrix with features. Fill it with problems. Features come later.
Both my personas revealed something crucial: They're not asking for more features. They're drowning in the features they already have while struggling with fundamental workflow problems.
The Head of Product has tools - CRM, analytics, surveys. She's cobbled together workarounds. But she's still spending "several hours each week" on manual processes.
The Solo Founder admitted to considering switching tools but worried about "just adding another layer of complexity."
More features isn't the answer. Solving the right problem is.
The next time you're stuck between Feature A and Feature B, stop. Instead:
Remember: Users don't want features. They want their problems solved.
Create your first persona and start getting research-backed insights in under 60 seconds.
P.S. The conversations in this article came from real Rooost persona interviews. Instead of guessing what users might say about problems, I asked them directly. The insights were worth more than any feature comparison chart I could have made.