Five Smooth Stones Ann Fairbairn

A practical walk-through of the Five Smooth Stones Ann Fairbairn approach

You have been handed a mess of research, a bunch of opinions, and not enough clarity. That is usually where I start looking at the Five Smooth Stones Ann Fairbairn method. It is not a rigid framework. It is more of a visual thinking tool that helps teams sort through messy qualitative data by assigning each piece to one of five categories, each represented by a different kind of stone. Before we get into the how-to, a quick note on what these stones actually are. Ann Fairbairn built this around a metaphor of stones you might find in a riverbed. Each stone type maps to a different quality of information or insight. Pebbles are small, quick observations. Cobbles are larger chunks of pattern or theme. Boulders are the heavy structural findings that change how you think about the problem. Glass stones are shiny moments — the kind of thing that catches light and draws attention, usually user quotes or emotional highlights. And the five smooth stones collectively represent the whole set of insights you have gathered, sorted into something a team can actually talk about instead of drowning in raw notes.

Five Smooth Stones Ann Fairbairn explained simply

The method works because it forces you to stop treating every data point the same way. Most teams dump interview transcripts into a document and then argue about what is important. The stones approach gives you a physical or visual sorting step first. You take raw data and categorize it by weight and texture rather than by topic alone. This changes the conversation immediately because people naturally understand that a boulder and a pebble demand different amounts of attention. Here is how I actually run this in practice. First, I collect all the raw material. Interview notes, survey responses, heuristic findings, whatever the project produced. I print them out on individual cards or sticky notes. One piece of insight per card. Then I lay out five zones on a large table or wall and label them: Pebbles, Cobbles, Boulders, Glass Stones, and the fifth category which varies slightly depending on which version of the method you follow — sometimes called Riverbed or Foundation, depending on the workshop facilitator. I do this with a team of four to six people. Anyone more than that and the sorting step turns into a meeting you will all forget. Each person gets the deck of cards and sorts them silently for about ten minutes. No talking yet. This part matters because if you start discussing immediately, the loudest person in the room will dominate the categories. After silent sorting, you go around and everyone explains their boulders first. Those are the heavy findings. Then cobbles. Then the glass stones — the emotional or quote-level moments that are worth preserving verbatim. Pebbles go last because they are the low-weight observations that may or may not hold up under scrutiny.

I once ran this with a product team that was stuck for three weeks on whether to build a new onboarding flow or fix the existing checkout experience. Their data was a wall of contradictory quotes. We did the stone sorting and the boulders immediately shifted. Two boulders came from completely opposite teams but said essentially the same thing in different words. That overlap is invisible in traditional affinity diagramming because you group by theme instead of by weight. The glass stones revealed that users were not confused by the onboarding. They were frustrated by a specific error state in checkout that nobody had written down clearly. We ended up fixing the error state first and the onboarding redesign became secondary. The stones made that hierarchy obvious in about forty-five minutes instead of three weeks of debate. There is a practical workaround I learned the hard way. When a card genuinely fits two categories, do not force it into one. Set it aside in a temporary holding zone and come back to it after the main sort is done. I used to split the difference and average the placement, which just created garbage data. The holding zone method keeps the signal clean and you usually find a consensus after the team sees all the boulders together. The counter-intuitive part most beginners miss is that the glass stones often end up being the most strategically useful item in the pile. People want to focus on the boulders because they sound important. But in actual design decisions, the glass stones — the direct emotional hits — are what move stakeholders. A boulder like "users find navigation unclear" does not make anyone act. A glass stone quote like "I felt stupid clicking that button three times" does. Keep the glass stones visible and accessible throughout the project, not buried at the bottom of a deck.

Another nuance that is worth mentioning: the stone categories are not fixed in volume. A healthy insight set for a moderate research project might have two boulders, eight cobbles, fifteen glass stones, and thirty pebbles. If you end up with zero boulders, your research did not go deep enough. If you have more than four boulders, you have not been ruthless enough in your analysis. This is a rough heuristic, not a rule, but it has saved me from presenting half-baked findings to clients more than once. The method is not without flaws. It does not scale to large datasets. If you have five hundred survey responses, you are not printing them on cards. It also relies heavily on the team having some shared context about the project. A brand new team member who was not in the research phase will struggle to place stones correctly, and their placements can throw off the group if you let them dominate the discussion. I recommend pairing this method with a brief research summary handout so everyone starts from the same baseline. It also does not handle quantitative data well on its own. You can layer numbers on top as annotations on the stones, but the method is fundamentally qualitative. If your project is primarily metrics-driven, you are better off using a different sorting approach or combining this with a separate statistical analysis step. For a concrete implementation, you can find examples and facilitator guides by searching for Five Smooth Stones Ann Fairbairn. There are downloadable card templates and workshop outlines available from design research communities and some UX conference archives. I have used both printed card versions and digital whiteboard versions depending on whether the team is co-located or remote. The digital version works fine but you lose the tactile benefit of physically picking up a card and feeling like it belongs in a certain pile. That physical sensation is not a gimmick. It is part of why the method reduces analysis paralysis. You have to let go of the card and place it somewhere.

If you are running this for the first time, budget at least ninety minutes for a small team with moderate data volume. Less than that and you are rushing the sorting and the whole exercise becomes a superficial categorization game. More than two hours and people get tired of looking at sticky notes and start putting things wherever just to be done with it. The sweet spot is roughly twenty minutes of silent sorting, thirty minutes of group discussion starting from the boulders, and then fifteen to twenty minutes of deciding what to do with the output. After that, you have a ranked set of insights that a product team can actually act on.