Teresa Torres on Why Discovery Teams Throw Ideas Away, Not Build Faster

The Real Test: Are You Throwing Ideas Away?

Most teams say they do continuous discovery. They run usability tests. They sit in on sales calls. They A/B test everything. But Teresa Torres has a simple filter that exposes the truth.

"You could do all the right activities, but if you're not changing your mind about anything, those activities aren't having an impact," she says. The tell is whether you're killing ideas. If every discovery session confirms what you already planned to build, you're just doing theater.

Torres has spent 13 years coaching product teams at startups and enterprise companies. Her book Continuous Discovery Habits has sold 750,000 copies. The shift she's seeing everywhere: teams moving from outputs to outcomes. They're no longer just shipping features—they're being held accountable for retention, acquisition, and engagement metrics.

The problem? Most teams weren't trained to drive outcomes. They were trained to execute roadmaps. Discovery is how you close that gap.

A/B Tests Measure Impact, Not Ideas

Here's where most growth teams get it wrong: they use A/B testing to decide what to build. Torres is blunt about this.

I like to frame A/B testing as it's a measurement stick. It's not informing should we build this or not. It's we decided to build this, did it have the impact we thought it would?

— Teresa Torres

The trap is obvious once you see it. You build the thing, run the test, wait for significance, then learn you built the wrong solution. You did all the work before learning you were off track. Even worse, most product teams don't have enough traffic to run statistically valid tests in a reasonable timeframe. A billion users don't use every feature.

Torres pushes teams toward assumption testing instead. Break your idea into its underlying assumptions. Map out what a customer has to do to get value. Then test whether those steps are realistic before you write a line of code.

"I work with teams that test multiple assumptions in a single day because that's the goal," she explains. When your assumptions are framed small enough, you can compare three solutions in the time it takes to design one.

Compare and Contrast, Or You're Just Guessing

Torres hammers one principle harder than any other: compare and contrast everything. Outcomes, customer needs, solutions. Stop asking "is this good enough?" and start asking "is this better than the alternative?"

Decades of decision-making research backs this up. We make better choices when we evaluate options against each other, not in isolation. But product teams resist. Their backlogs are full. Testing multiple ideas sounds like more work.

We tend to think about, like, I think we're trained to think in absolutes. Like, we grew up going to school, there's a right answer. And I think what we see in business is there's lots of good answers. Some are better than others.

— Teresa Torres

This is where frameworks like ICE scoring (impact, confidence, effort) can help—but only if you actually test the assumptions behind your scores. If you're just putting numbers on a spreadsheet in a room full of stakeholders, you're still guessing. Use ICE to surface what you believe, then go validate it with customers.

Customer Value and Business Value Are the Same Fight

Torres pushes back hard on the UX community's obsession with customer value alone. Creating value for users without creating value for the business is a disservice to everyone.

"If I create customer value but I don't create business value, what happens? My product is going to get shut down. Companies aren't charities," she says. The goal is alignment: find the customer need that, when solved, drives retention or acquisition or monetization.

Her discovery framework starts with an outcome. You need to know what success looks like—retain more customers, grow average order value, reduce churn by 3x. Then you interview customers to uncover unmet needs, pain points, desires. The synthesis step is where most teams drown. You hear everything is a little different, or worse, everything sounds exactly the same (confirmation bias).

It's really about humans and getting humans to do things and helping humans understand things. And helping humans get more value out of things.

— Teresa Torres

The final step is evaluating solutions by how well they address customer needs and drive the business outcome. If you can't connect a solution to both, kill it and move on.

Start With One Habit, Not a Framework

Torres didn't set out to build a rigid process. She named her book Continuous Discovery Habits because she wants teams to pick one tool at a time. Interview weekly. Run small assumption tests. Document why you're building something and revisit it 30 days later to see if it had the expected impact.

That last habit is a forcing function. When teams write down their predictions, they realize most releases don't move the needle. That's not failure—that's a feedback loop. You learn what assumptions you got wrong and get better next time.

The shift from project-based research to continuous activities is hard. Teams fall into mini-waterfall. They do two-week sprints but still ship what was planned six months ago. Torres wants you asking: are you doing the right activities regularly, and is your outcome providing actual focus?

Growth isn't just optimization. Changing a call-to-action can 4x conversion, but discovery helps you find the next mountain, not just climb the local peak faster.