Fromstickynotetotheprocessworthautomating

When a company brings me in to automate something, the first meeting usually ends with a full board: retyping the order that arrives as a photo, checking the spreadsheet, approving by email, assembling the report by hand. They’re all real tasks. The risk is walking out with a list of automations and no criterion for saying which one comes first.

The method I use to avoid that I learned on design projects. The clearest example I have is from a project I delivered for Eduzz, a Brazilian platform for digital products, in 2024.

In October of that year, the Eduzz Year in Review got a workshop with stakeholders. The board ended up with dozens of sticky notes: transaction volume processed, satisfaction score, an AI-made song for each creator, a timeline of the platform’s ten years, a notification saying the year in review was available, a version in another language.

They were all reasonable ideas. They all fit into a sequence of stories. That’s the problem with a full board: it gives you the feeling that “we’ve got it,” and the next meeting turns into a list of screens.

What prevented that was the order of things. First, the ideas were grouped into four fronts (Product, Creators, Journey, and +Human). Then each one was placed on an impact vs. effort matrix. And behind both steps was a question that had already come out of discovery: how to make a year in review people want to show off. Grouping before prioritizing is what lets you compare things that are comparable. Having a clear outcome is what lets you say which ideas serve it.

This article covers the two tools in that stretch: affinity mapping, to organize what was gathered, and the opportunity solution tree, to decide what to do with it. At the end, I show the same tree applied to a manual process.

The starting point: one idea per sticky note

Before you can group, the notes have to be in a format that can be grouped. Three simple rules:

• One idea per sticky note. If the sentence has an “and” in it, it’s probably two sticky notes.
• The speaker’s words. “Leave them wanting more” says more than “post-review engagement.” The second one is already a conclusion.
• Traceable origin. On the Year in Review board, each sticky note carried the name of whoever wrote it. When it’s time to defend an idea, knowing where it came from carries weight.

In the workshop, the one-idea rule was the one broken most often, and that’s normal. One sticky note combined three things: the creator’s year-over-year growth, a timeline with the milestones of each moment, and a story about how far they were from the next tier. In the synthesis, it’s worth splitting them.

Affinity mapping: grouping from the bottom up

Affinity mapping is easy to explain: put similar sticky notes together until the groups reveal themes. The hard part is doing it without bringing ready-made categories from home.

The way that works best for me:

1. Spread everything out, with no categories. No pre-made columns based on the product’s org chart.
2. Group in silence. Each person moves sticky notes next to others that “talk to each other.” Discussing too early lets the most talkative person define the groups.
3. Accept small groups and orphans. A lone sticky note may be noise, or the most important finding of the round.
4. Name last. Only once the groups have settled.

In the workshop, the four fronts worked as a first level of organization. This is a selection of the actual sticky notes from the board, condensed:

Product 4

• Transaction volume processed in the year
• How many improvements came from customer requests (“you’re part of this too”)
• Images of the old Eduzz: the logo’s evolution over ten years
• Show features they haven’t used yet: “we’re still waiting for you”

Creators 4

• Creator growth compared with 2022 and 2023
• Timeline with the creator’s and affiliate’s milestones
• Story about how far they are from the next tier
• An AI-made song for each creator

Journey 3

• Introduce the new recognition that goes beyond the numbers
• Notification saying the year in review is available
• Recognition journey metrics

+Human 4

• Show the creator’s impact over these ten years
• How many new customers the creator reached this year
• Leave them wanting more: “you didn’t hit the goal, but we’ll be here next year”
• Make the year in review a narrated story

Notice that the fronts mix two logics. “Product” and “Journey” are about what the platform wants to say. “Creators” and “+Human” are about what the creator experiences. That’s where the next step makes a difference.

Name the group after the need, not the feature

This is the step that separates a useful synthesis from a list of requests. A group’s name should describe what the person needs or feels, not what the product should have.

Discovery for the Year in Review had already provided the clue. In the previous edition, the most shared screen (21%) was the one saying how many lives the creator had helped change, and the second (20%) was the opening screen with the person’s name. Reread through that lens, the sticky notes reorganize around what the creator wants to show:

Feature name (avoid)	Need name (prefer)
Personalized opening screen	“I want to see myself in this year in review”
Lives-changed screen	“I want to show the difference I made for others”
Next-tier story	“I want to know where I’m headed next year”
Platform transaction volume	(this is what the platform wants to say, not what the creator wants to show)

The last row is the most useful one. It doesn’t say the idea is bad. It says the idea answers a different question. A good test: if the group’s name fits in the user’s mouth, in quotes, you’re on the right track. If it only fits in the company’s mouth, it belongs in a different piece.

The matrix: impact vs. effort

With the groups formed, the workshop placed the ideas on an impact vs. effort matrix. The original board used labels like “do now” and “schedule it.” The positions below are approximate, read from the real board:

Impact →

High impact, low effort

High impact, high effort

Low impact, low effort

Low impact, high effort

Availability notificationTransaction volumeImprovements from customer requestsRevisit company milestonesAI-personalized messageNew recognition programVideo testimonialsGrowth and next tierAI song per tier

Effort →

Two readings of this matrix. First: it’s great for cutting. The AI-made song shows up as high effort and low impact, and the sticky note itself already proposed a cheaper alternative (one song per tier, not per person). Second: on its own, the matrix doesn’t say what for. The creator’s growth with the next tier landed on the high-effort side, and yet next year’s tier became the closing screen of the final version. What decided it wasn’t just its position on the board. It was the link to the outcome.

That link is what the opportunity solution tree makes explicit.

From synthesis to decision: the opportunity solution tree

To decide, I use the Opportunity Solution Tree, proposed by Teresa Torres in Continuous Discovery Habits. It has four levels:

• Desired outcome. A measurable change, not a deliverable.
• Opportunities. Users’ needs, pain points, and desires that, if addressed, move the outcome.
• Solutions. Ways to address a specific opportunity. Several for each one.
• Experiments. Quick tests of the assumptions behind each solution.

The tree’s strength is traceability. Every solution hanging on it answers: which opportunity does it address? And which outcome does that opportunity move?

The Year in Review narrative fits this structure. It starts with the creator (personalization), moves through their impact on people, and ends looking ahead, with the tier of the recognition program they can reach next year. Each stretch is an opportunity:

• A year in review creators want to show offdesired outcome
  • “I want to see myself in this year in review”opportunity · personalized opening got 20%
    • Opening that greets them by namesolution
    • The creator’s achievements this yearsolution
    • A song for each creatorworkshop idea
  • “I want to show the difference I made”opportunity · lives changed got 21%
    • The creator’s impact on peoplesolution
    • New customers they reachedworkshop idea
  • “I want to know where I’m headed”opportunity
    • The tier they can reach next yearsolution
    • How far they are from the next tierworkshop idea

Look at what the tree does with the ideas from the “Product” front. Transaction volume and new features shipped don’t hang from any of the three opportunities. It’s not that they’re bad. It’s that they answer what the platform wants to say, and the outcome for this piece is what the creator wants to show. A feature that doesn’t fit anywhere on the tree is, by definition, a guess, or a branch of another tree.

How to build the tree without turning it into bureaucracy

• One outcome per tree. If the team has three goals, that’s three trees. Mixing them makes everything look like a priority.
• Opportunities can have children. “I want to know where I’m headed” can break down into “know which tier I’m in” and “know what I’m missing.” Breaking down a big opportunity usually reveals where to start.
• Compare opportunities, not solutions. The first choice is which need to go after. Only then do you compare solution against solution, within the chosen opportunity.
• More than one solution per opportunity. With only one, there’s no comparison; there’s a plan in disguise.
• An experiment tests one assumption, not the whole solution. “Do creators share a screen about the impact they’ve had?” had already been answered by data from the previous edition. When the evidence already exists, use it before inventing a test.

Common mistakes going from the wall to the tree

• An outcome that’s a deliverable. “Launch the 2024 year in review” at the top isn’t an outcome; it’s the project. The top has to be a change you can measure, like sharing.
• An opportunity that’s a solution in disguise. “We need a revenue screen” is already an answer. The opportunity behind it might be “I want to show that I grew,” and that one allows for other solutions.
• Prioritizing without grouping. A matrix with thirty loose sticky notes compares a notification with an AI-made song. Grouping is what makes the comparison fair.
• A tree made only of opinion. Over time, each opportunity needs to point to evidence. In the Year in Review, two of them pointed to numbers from the previous edition. The third, “where I’m headed,” was a narrative bet. It’s worth marking the difference.

The same tree for a manual process

The tree isn’t just for communication pieces. When the subject is automating an operation, it does the same job: it keeps the list of manual tasks from turning, unfiltered, into the list of automations.

At a lubricant distributor, orders arrived as photos of paper notes on WhatsApp, and someone retyped everything. Put that way, the solution looks like a single thing: an AI that reads the photo. Looking at the process, there were four pain points. For the back office, every order was duplicated work. For inventory, the balance lived in the head of whoever picked the order. For the buyer, there was no history. For the customer company, the order belonged to whoever sent the message.

• An order that reaches the store correct, with nobody retyping itdesired outcome
  • “I don’t want to type again what was already written”opportunity · back office
    • AI reads the photo and proposes the items; the buyer confirms line by linesolution
    • Order built from the catalog, on the phonesolution
  • “I want to know why the balance is what it is”opportunity · picker
    • Inventory as a ledger of movements, not an editable fieldsolution
  • “I want to find what I’ve already ordered and know if it’s ready”opportunity · buyer
    • History and status in the portalsolution
    • Email and WhatsApp notification when it’s readysolution
  • “The order belongs to the company, not to whoever sent the message”opportunity · customer company
    • Invite-only access; every buyer sees the same historysolution

Notice where the AI shows up: in one solution, under one opportunity, out of four. It’s the one that draws the most attention and the only one that requires a person confirming each line. The other three are record-keeping and access. Without the tree, the list would have had a single automation, and three pain points would have stayed right where they were.

The tree is alive

The tree isn’t an end-of-discovery deliverable. It changes with every round of conversations and every experiment. An opportunity that seemed central can lose strength; one nobody saw shows up on an orphan sticky note.

That’s why I prefer to keep the tree somewhere visible and keep it updated, instead of producing a polished version for a presentation. The value is in the conversation it sparks: “why are we doing this?” now has a drawn answer.

Practical summary

• Each sticky note has a single idea, in the speaker’s words, with its origin marked.
• The groups emerged from the notes, not from categories brought in ready-made.
• Each group has a need name that would fit, in quotes, in the user’s mouth.
• The impact vs. effort matrix came after the grouping, not instead of it.
• The tree has a single measurable outcome at the top.
• Every solution hangs from an opportunity, and every opportunity moves the outcome.
• Each opportunity points to evidence, or is marked as a bet.

If the inventory ended in a list of automations, go back a step. The middle was missing: the manual work that justifies each one, and the outcome it moves.