I thought I’d share something a practical today.

In my past post on Sönke Ahrens’ How to Take Smart Notes I mentioned that the other big game-changer for my personal knowledge management practice was enrolling in Tiago Forte’s Building a Second Brain course.

One of the core techniques from Tiago's course is “progressive summarization.” Since learning about the technique, I’ve applied it to just about every kind of textual note I take, including reading notes, meeting notes, and, most recently, notes I’ve taken during user interviews.

It might be helpful for anyone else who engages in this kind of research, so I thought I’d share how I do it.

What is progressive summarization?

Progressive summarization is Tiago Forte’s technique of compressing context-rich notes into information that is easy to consume and discovery in the future. Tiago points out that the primary challenge when taking notes is not in their collection. That’s actually pretty easy: everyone has a method, whether it’s jotting things down on your arm, scratching them onto a notepad, or typing them into a favourite app. Nor is it portability difficult. With digital technology, it’s a simple task to capture a note and move it between different platforms.

No, the real challenge with taking notes is retrieval across time. The most important moment in note taking isn’t when you create the note, it’s when you need it again in the future. How can you take notes that are valuable for a future self, who might be living in a vastly different context?

This is a big issue in user research. Many a user researcher has bemoaned the insight that went to die in a deck. Many others have come up with solutions, some simple, some elaborate, to prevent that from happening. For instance, there’ s Tomer Sharon’s “research nuggets” database that looks to do for research what atomic design has done for UX. There are also software platforms focused on the effective archival and retrieval of user research, like Dovetail and Handrail. These are complex, expensive solutions to what seems like it should be a simple enough problem.

Balancing discoverability with detail

Even if your research is properly stored and archived, there’s another problem: context.  A key point that Tiago makes is that note-taking needs to balance discoverability and detail. When we take notes, we’re usually compressing data. That is, we’re taking something long and distilling it into something shorter and more accessible. That makes insight easier to discover later.

The trouble again is that time gap. The problems we have when we first make a note are not necessarily the problems we have later, when we actually need that insight.

Moreover, this is lossy compression. When we compress the original too much, we lose rich information. There’s always a risk of something getting lost in translation. I’m sure we’ve all encountered a situation in which we’ve come back to an old note and been baffled trying to work out what we were trying to say or why it mattered.

“The information you produce needs to also be consumable by others, while preserving as much of the original context as possible.”

If this is a problem in personal knowledge management, it’s worse when you’re working on a team.

Often you need to produce insights that are consumable not just to yourself but to others. Ideally, your entire team is getting plenty of hours of exposure to real users. But, it’s not always feasible for everyone to be present. The information you produce needs to understandable by others, preserving as much of the original context as possible.

That’s the trade-off: detail. It’s hard to maintain the deep, rich information you get out of something like a user interview while still producing something that’s also easily consumed by someone else who, if we’re being realistic, might not have time to read an entire transcript or watch or listen to a recording.

Layers of summarization

Progressive summarization looks to preserve as much detail as possible while also making notes easily skimmable.

It involves four “layers.”

The base layer (0) is the initial “text.” Tiago’s usually talking about books or articles, and so this would involve going through and underlining anything that is at all interesting. (Many people who practice Tiago’s methods work with e-texts. For me, I underline things in my Kindle and then use a service called Readwise to automatically import the highlights into Evernote. Readwise does the same job with pieces I’ve saved and highlighted in Instapaper. Those excerpted passages represent my level 0 text.)

Doing UX research, you might have a transcript, or written notes taken while watching or listening to a recorded interview. Whatever your method, identify anything you find interesting. You can be a bit liberal here, but you don’t need to go crazy. One of the benefits of this method is that it’s non-destructive: you’re never going to completely lose that original text, so you can be a bit selective. In fact, I think it’s a sign of maturity with the method that you’re able to let go of the compulsion to highlight every single little idea.

Example of progressive summarization

Example of progressive summarization of Robert Shiller’s Narrative Economics.

The next layer (1) involves going through your underlined passages and summarizing them further by bolding the best bits. You need to be more selective now. These are the most striking insights—the ideas that you really want to keep. Then, the next layer (2) means reviewing the bolded passages and highlighting the best bits of the bolded passages.

Already, you have a document that maintains rich detail but is much more scannable. Anyone could quickly skim through it and pause over the most important elements. Even long documents can be processed quickly.

There are two more layers to progressive summarization. The next (3) involves writing a summary of the original. Crucially, this requires reframing the original text in your own words, from your own perspective.

The final stage (4) is remixing: combining the insights you’ve gleaned from the original with concepts and ideas from other places, and producing new, original thought.

“By engaging in progressive summarization, you’ll be able to very quickly assess the worth of the notes you’ve taken just by quickly at how much it has been highlighted or summarized.”

This seems like a lot. But one other thing Tiago emphasizes is that you wouldn’t necessarily go through all the stages with every single note. In fact, he points out that the amount of summarization is correlative to the value that the document has. A quick, pithy blog post might produce only a few original notes, one of which gets bolded. You may never go back and highlight anything at all. It might not be worth returning to. Meanwhile, a meaty, resonant text could become the lynchpin of any number of new pieces of writing. You can easily tell that it has value by the visual cues from the bolded and highlighted text.

By engaging in progressive summarization, you’ll be able to very quickly assess the worth of the notes you’ve taken just by quickly at how much it has been highlighted or summarized.

Modifying for user research

User research is a bit different from the kind work that I suspect Tiago does most of the time. So, when I am engaging in progressive summarization of a research session, I make a few tweaks.

The primary change is that I use colour-coding into the highlighting to help me identify some of the things I’m particularly attuned to when doing research. For example, I might designate one highlight colour to represent pain points—moments where the interviewee was providing evidence of something that might need to be improved. Another colour might signify tasks that the interviewee is performing. Another might highlight points that reveal their mental model. These are just suggestions; you can, of course, adapt the categories to fit your needs.

Then, when I summarize, I usually home in on these highlights and summarize them at the top of the document. This wouldn’t be simply a list of everything I highlighted. That would be pointless. Summarization requires a level of interpretation. I am teasing out the things that stood out to me as the most significant aspects of the interview, using my expertise as an interviewer and experience as a researcher to distill the interview into the points most salient to whomever is going to need to use the research. I can add some colour-coding here, as well, to help any subsequent reviewers to see how I arrived at the condensed summary.

I may also add some notes to help clarify why I might emphasize something, especially if it’s not immediately obvious. When doing so, I make sure to signify that it’s my interpretation by using yet another colour, as well as a recognizable prefix. I don’t want anyone to think that my interpretation is an objective truth that came out of the interview, so I call out moments where I’ve added any spin.

From there, there’s likely going to be another level of translation after the various interviews or conversations are collectively summarized into some other kind of artifact, like a persona or a journey map. It’s nice, though, if you can link those artifacts back to the research. With progressive summarization, you never really discard much, so it makes it possible for anyone to move from that artifact back through time to an actual quotation or piece of testimony from the user.