Skip to content
Quevell
Menu

Your next user might know something about your product that you don't

Feature requests and bug reports reveal what product teams cannot always see themselves. Lessons from Firefox, Spotify, and Buffer show why listening to users matters, and why genuine conversations are worth preserving.

2026-11-13 · Practice

logo

Imagine receiving a message from a SaaS customer:

"Why can't I do this? I have to repeat the same five steps every time."

To the developer, it might look like another feature request. To the customer, it describes a problem they encounter every working day.

And it reveals something the product team may never have known.

When building software, we work with assumptions about how people will use it. We design interfaces, define workflows, test hypotheses, and try to anticipate unusual situations.

But even an experienced engineering team cannot reproduce every customer's environment, habits, and constraints.

That is why a feature request or bug report deserves more attention than simply being added to a backlog. It is an opportunity to learn something about the product from someone who experiences it differently.

Sometimes a single message exposes a weakness in a decision the team had considered perfectly reasonable.

What happens when thousands of users ask for the same thing?

On April 29, 2025, Mozilla released tab groups in Firefox.

The story started in March 2022, just one day after Mozilla Connect launched. Someone submitted a request for tab grouping, and it gradually became the most popular idea on the community platform.

By the time Mozilla announced the release, approximately 4,500 people had supported the request. The team had reviewed more than 64 pages of feedback and over a thousand comments.

The interesting part is what Mozilla did with that information.

Firefox product manager Stefan Smagula described how the team looked for underlying needs rather than implementing every suggestion literally.

Users had different ideas about how tab groups should work, but many were trying to solve the same problem: keeping a large number of open tabs manageable.

The team studied those workflows, observed early testing, and refined the implementation.

Users influenced both the decision to build the feature and the details of how it worked.

Mozilla documented the process in You asked, we built it: Firefox tab groups are here.

Three years is a substantial time between an initial request and a release. But the discussion did not simply disappear into an internal backlog. Community members kept contributing feedback and participating in testing.

Their experience became part of the development process.

Listening doesn't mean building everything

If a thousand users request a feature, should the company automatically implement it?

Not necessarily.

Customers understand their problems, but they don't always know the system's architecture, technical constraints, or the consequences a proposed solution might have for everyone else.

A feature request deserves investigation, not an automatic commitment.

Spotify provides an interesting example.

Its Idea Exchange allows users to submit suggestions, discuss proposals, and vote on ideas. Community staff pass feedback to relevant teams, and suggestions receive statuses reflecting the company's decisions.

Spotify explicitly explains that votes are not the only factor. The company also considers research, usage data, feedback from other channels, and its broader product strategy.

And some requests do become real features. Spotify lists Chromecast support, editable display names, playlist customization, and other additions among the changes influenced by community suggestions.

The process is described in How does my feedback reach Spotify?.

There is an important distinction here.

Spotify does not promise to implement every popular proposal. It explains how suggestions are considered and why some are not currently planned.

Having influence over a product is not the same as having the authority to dictate its roadmap.

Still, there is a significant difference between explaining why something cannot be implemented and leaving a customer with no meaningful response.

A bug report is a contribution, too

Feature requests are relatively easy to recognize as useful input. Bug reports deserve the same treatment.

A good bug report contains the results of an investigation that a user has effectively conducted on the developer's behalf.

Someone encountered unexpected behavior in a real environment, possibly discovered how to reproduce it, took time to describe the issue, and decided to share what they found.

That effort is not part of their subscription fee. Nor is it their obligation.

Some defects only appear with particular data, configurations, browser and operating system combinations, or uncommon workflows. No test suite can cover every possible real-world situation.

The most valuable reports often explain what the user was trying to accomplish, not merely what went wrong.

Sometimes the software behaves exactly as implemented, yet the experience is so confusing that the user reasonably considers it broken.

That may not be a code defect. The problem could be the interface, the documentation, or the original product decision.

Developers do not have to agree with every diagnosis. But the circumstances behind the report still deserve attention.

After all, those are the circumstances in which the product is actually being used.

Buffer: turning feedback into a working process

In May 2022, Buffer shared what it had learned from launching Start Page, a simple tool for building a page of links.

Rather than attempting to build every conceivable feature before launch, the company released a basic version and encouraged customers to describe what they needed next.

One request appeared repeatedly: people wanted statistics showing page views and link clicks.

Before launch, analytics had been one possible feature among many. Once customers started asking for it, the team had better evidence of its importance.

Buffer prioritized the work and released the feature a few months later.

The company described the experience in What We've Learned Launching a New Product as an 11-Year-Old Company.

The feedback process continued to evolve.

In August 2023, Buffer introduced an updated public system called Buffer Suggestions. Its announcement openly examined the weaknesses of the previous process.

Customers could submit duplicate requests without knowing that others had raised the same issue. They could not easily see what happened after submitting a suggestion. Even the team sometimes struggled to associate requests with the people who had made them.

The new system let users browse suggestions, vote, comment, and receive notifications as requests changed status.

Buffer also acknowledged the operational cost. The company received hundreds of feature requests each month, while the team's capacity remained limited.

Listening takes work. Someone has to read feedback, identify recurring problems, separate symptoms from causes, and make decisions.

But those activities are part of understanding what customers need. Treating them as unwanted administrative overhead misses their value.

Why I don't want AI to replace every support conversation

Modern support teams can automate a great deal.

AI can find answers in documentation, categorize incoming requests, identify recurring problems, and help staff prepare responses. For routine questions, these capabilities can be genuinely useful.

The problem begins when automation becomes a barrier between users and the people responsible for the product.

Imagine someone discovering a defect that the documentation doesn't cover. They contact support and receive automated instructions to repeat steps they have already tried.

After several exchanges, they give up.

The support system might record the conversation as resolved. The underlying problem remains, and the engineering team never learns what happened.

Buffer offers an interesting contrast. On its official website, the company explicitly emphasizes support from real people, including opportunities to speak with team members in roles ranging from marketing to engineering.

Buffer also develops AI features and integrations with AI assistants. Its approach demonstrates that adopting new technology does not require abandoning direct human relationships.

I don't believe every support response must be written manually. Automation is valuable when it reduces waiting times and repetitive work.

But there should be a clear route to a person when automated answers fail, particularly when a customer is reporting something unexpected.

For a young SaaS company, this matters even more. Without years of accumulated customer experience, every substantive conversation provides an opportunity to discover how the product behaves outside the development team.

Customers don't owe us product feedback

In Quevell, we have separate feedback options for feature suggestions, bug reports, and support requests.

That is a small technical mechanism. Its existence alone proves nothing about whether a company actually listens.

As a founder, I want real conversations to remain possible behind that form. I don't want to promise that every suggestion will be implemented. I want us to understand the problem, explain our decisions, and acknowledge when we have overlooked something.

As the product grows, doing that consistently will require discipline. We will need to preserve context, recognize recurring feedback, and make time for conversations even when the engineering backlog is full.

Giving up those conversations for better-looking support metrics would be a poor trade.

Customers have already paid for a product with money or time. They are under no obligation to test it for us, suggest improvements, or write detailed defect reports.

When someone chooses to do so anyway, they are giving us information we might never discover on our own.

We can disagree with a proposed solution. We can postpone a feature or explain why it does not fit the product.

Losing the opportunity to hear the customer in the first place is a much greater risk.

Knowledge of how software works in the real world is every bit as valuable as knowledge of how it works internally.

#Product development #User feedback #SaaS #Founder’s Vision