Friction worth protecting when shaping products

https://thoughtbot.com/blog/friction-worth-protecting-when-shaping-products

Today, with the rise of LLMs that can generate a product brief or full prototypes in minutes, friction between a product concept and a polished deliverable is getting smaller and smaller. The tools and processes for product discovery and shaping have changed rapidly. Founders and product teams are shaping ideas with AI and can move quickly but this loss of friction can lead to shipping the wrong thing if we’re not careful.

Categorizing friction

The debate around the question “is the way AI removes friction a good thing” is certainly not new but I recently heard Fabio Cortez speak at Webflow Conf and this quote stuck with me:

there are two kinds of friction, one we want to automate, the friction between me and my intent, the repetition, translation, reconstruction happening between stakeholders, we should automate that. But the friction between us and the problem we’re trying to solve, we should not automate, we should protect it always, that’s where the magic happens, the ambiguity, the critique, the exploration, that’s the friction to protect.

Fabio does a great job articulating two categories of friction:

  • Intent friction – not ideal
  • Problem friction – important

Intent friction is what slows down your ability to effectively articulate an idea or build a clearly defined feature. It assumes the direction is validated and focused. This can look like visualizing a concept you’re discussing as a team, designing a common UI pattern, writing code for a component you’ve built before, or documenting existing patterns.

Problem friction arises when the solution isn’t clear. It looks like slowing down to ask why, discussing different opinions as a team, critiquing your design work, or exploring a wide variety of ideas. It’s ambiguity and constraints, critique and conflicting ideas, messy exploration and sparking new ideas. The human side of shaping and building products.

Friction uncovers assumptions

Problem friction is useful because it allows us to uncover assumptions that add risk to the product. It’s an opportunity to ask bigger questions. To slow down and ensure what you’re planning to build is grounded and the whole concept doesn’t hinge on an ambiguous, untested theory.

A note on LLMs

LLMs hate ambiguity so they fill in the gaps with assumptions. Attempting to minimize friction by using LLMs can result in products that don’t actually align with user needs or serve the long term goals of the business. LLMs look for patterns and their output tends to look the same, meaning the more you rely on LLMs to shape or build your product, the more it will look like every other app and gradually lose its soul and distinct purpose.

Sure LLMs can be a helpful sounding board for your ideas and their output can be improved with skills but it’s still an imperfect solution that often gives sycophantic answers, meaning it will tell you what you want to hear. This is partly why we built ReadySetGo, to help shape ideas and surface hidden assumptions that need testing before you build anything.

Testing validates assumptions

If problem friction finds assumptions, how do you test and validate these assumptions? Nothing beats talking to users. A team or LLM can hypothesize about how to move forward, but the best way to build true confidence is to get in front of the real people who will use the product you’re envisioning. Listen to users articulate their problems before offering a solution, often this will reveal far more interesting information than showing them a solution that already has your biased perspective baked in.

What we hear from people during user interviews almost always challenges some of our pre-conceived assumptions. This causes us to change everything from the features we prioritize building to small details like the text in a button. Sometimes it causes us to re-think the entire approach. On the flip side, it also gives us confidence when we see users respond well to our ideas.

Sprints as a framework for problem friction

At thoughtbot we use shaping sprints to clarify what we plan to build. Shaping sprints allow us to ask questions that probe at the product’s core idea. While this can sometimes feel uncomfortable, this kind of problem friction is critical for uncovering assumptions.

A quick story. On a recent sprint for a new project, the team came in confident with a product plan and prototype built with an LLM. On the surface it seemed sound, and it looked pretty. The prototype worked and proved the idea was technically possible, but it was hiding an array of assumptions about how people would use the product and how scaling features would change the experience. Thankfully the shaping sprint caught these assumptions early and we were able to quickly pivot and have confidence that our new direction would actually resonate with users.

An important question

Not every decision benefits from problem friction. When the solution is clear and the team knows why, either from data or because the problem has been solved before by others, any friction between your idea and a working product is intent friction and is safe to minimize through using LLMs.

How do you distinguish intent friction from problem friction? Ask yourself:

Can you fully describe the result you want, explain why it’s shaped that way, and point to evidence that validates the assumptions behind it?

If there’s any hesitation or ambiguity, you’re looking at problem friction. Ignoring this friction risks baking assumptions into your product, which could result in product/market mismatch, failed customer acquisition, or extra development time and expense.

About thoughtbot

We've been helping engineering teams deliver exceptional products for over 20 years. Our designers, developers, and product managers work closely with teams to solve your toughest software challenges through collaborative design and development. Learn more about us.