Is Roam Research the next big thing "for networked thought" or just vaporware?

I’ve had a love-hate relationship with note-taking applications for years, at one point spending years on Evernote before abandoning it for a paper journal and the eloquently simple Apple Notes. Apple Notes is present on all my devices, it’s fast, and it’s simple. But organizing and finding notes becomes an impossible chore at some point, and notes in my paper journal go there to never been read again. I try to use notes for everything—self-therapy, articles or book summaries, insights, personal organization, research, and more—so having a good and easy system is important.

In the desire to find the “perfect” tool, I often end up using rather esoteric tools. When I do software development, I usually use Vim, a complicated but efficient text-based editor, despite my seeming inability to ever remember more than a dozen of its endless keyboard shortcuts and commands. Even though it’s not as flashy as many of the more graphical user interfaces, Vim allows me to efficiently program and edit, freeing my mind to focus on the big picture.

However, because no tool is truly perfect, it’s easy to wonder if the grass really is greener. So, it should not have been a surprise that when during the pandemic I found myself stuck at home in an NYC apartment with young kids that I would decide to reach for a new note-taking tool. I was trying to organize research notes I was taking about the pandemic and related topics and was very dissatisfied with what I was using.

I couldn’t help but continue to see tweets in my timeline about Roam Research, which describes itself as “a note-taking tool for networked thought.” At its core, this means that it is easy to link and connect not just different note pages, but paragraphs inside those note pages. Roam seemed to be much loved by its cult-like supporters on Twitter, who argue that connecting notes this way is more natural and may allow better organization than a top-down categorization system. Roam would make it easy to trace your thoughts on topics and enable serendipitous discovery of past insights and ideas. I found this idea of a relatively simple notes tool that promised to let me forget about top-down categorization and just start taking notes very appealing.

Having now used Roam for more than half a year, I can only say that I’m moderately happy with it. It’s good enough for me to use almost daily, which is certainly something, but not nearly great enough for me to fall in love with it—I’d switch to something better if it were offered to me. In a sense, Roam does remove some of the mental overhead of attempting to categorize everything, but it doesn’t eliminate it. My impression is that not having a plan or strategy for your notes on Roam will ultimately result in unfindable notes again, and I haven’t yet really experienced the serendipitous note discovery I was hoping for from the app. My experiences with the app and my nagging worry about the future direction of the application have kept me from being willing to drink the Roamcult Kool-Aid.

How I use Roam Research

Everyone seems to use Roam differently, which is both a blessing and a curse—it’s difficult to figure out what works best and to learn from others’ experiences. I’m sure that my usage of Roam is no exception, which makes it difficult to really recommend for or against Roam for a particular person.

When I first load my Roam “graph,” I am sent to the Daily Notes section, a thoughtful and somewhat unique way to prompt my note-writing and allow a natural organization to emerge. Some Roam experts recommend starting all notes through the Daily Notes section.

In the Daily Notes section, I immediately type ;; and select a Daily notes template that I have previously created in Roam. This automatically inserts several sections like “Daily Journal” and “Daily Todo List”, which I use to make sure that nothing is slipping through the cracks while avoiding any sort of strongly structured TODO system. Roam probably isn’t a great comprehensive TODO application replacement. On the other hand, I’ve never actually managed to use a TODO app or system long-term. They always become unmaintainable and taunt me with everything undone, while making me feel misleadingly productive because I’m checking off TODOs rather than focusing on the important big picture. Casually throwing TODOs into my notes and occasionally onto project pages seems to work better for me.

I also initially created a separate “Meal Planning” and “Week Planning” page, which I fill, making links to the upcoming days of the week. When I fill in details under the date for Wednesday on each of these pages, those details will show up when I get to Wednesday on the Daily Notes page as a “Linked Reference.” These references are a large part of the power of Roam. I put important events I want to make sure I don’t miss, even though they’re on my calendar, and I put any meals I am planning to make.

Most importantly, I use Roam to take notes about things I’m thinking about and notes about decisions I’m trying to make. Roam has replaced my paper journal thinking. I also use Roam for taking notes on articles, books, and research papers. Although I like taking notes in Roam, I haven’t yet come back to these notes in a way that makes me think “Aha! This is why I’m using Roam” — most notes sit there as notes do in most note apps: forgotten. I’m not sure that’s bad, part of the value of note-taking is the act of note-taking, but I haven’t yet figured out how to leverage my notes in a more useful way. I think (or imagine, anyway) that if I were doing deep research on a topic, it would be easier to make the notes useful, so this may just reflect where I am at the current moment: mostly jumping back and forth between unrelated ideas and topics.

Browsing around Twitter, YouTube, and Google, I have found that people have invented all sorts of amazing ways to use Roam, but unfortunately, many of them are complicated to implement, very rigid to use appropriately, and seem very individualized to the particular person showing off the technique. I sometimes wonder if these people actually get anything important done other than inventing the perfect productivity system.

What’s great about Roam Research

In the big picture, I find the ease and simplicity (if you want) of Roam, combined with the block/paragraph level linking and referencing, to be great. The fact that Roam also creates implicit “unlinked references” for words and phrases that match creates a lot of interesting possibilities. Roam seems great for certain types of research and thought and (maybe) good for those of us who crave organization but are bad at organizing. Roam’s approach seems to help solve some of the problems of perfectionists like myself who get stymied by figuring out categories, hierarchies, and tags. It’s clear that if you can master Roam’s linking functionality, the possibilities become great, perhaps even allowing Roam to become a mostly all in one solution to Build a Second Brain.

Roam is also extremely customizable. It’s relatively easy to add your own custom CSS and JavaScript and there are countless themes and extensions for Roam. I’ve only used these minimally; I added some custom CSS and JavaScript to strike out an entire line when a TODO is crossed out and also slightly tweak a couple of visual parts of Roam to make it more pleasant. The value so far has been minimal, but I could see it being useful to really make the app fit me in the future.

Finally, Roam has a great depth of tools that are not immediately obvious. It has Pomodoro timers, block-level version control, kanban boards, calculations, tables, word counts, diagrams, latex syntax, and many other (poorly documented and practically hidden) features. It’s easy to geek out on some of these, but I’ve found that most of them aren’t useful (or usable) enough.

What’s not so great

I started taking notes on everything that annoyed and frustrated me (using Roam, of course), and the list of things that are “not so great” could fill their own blog post. I think that the vast majority of these things are of minor consequence, but I sometimes feel that Roam is killing me with a thousand little cuts and annoyances. Everything feels unpolished and clunky and as a result, I can’t imagine asking a team of people to use it with me or using it for something like a company wiki.

Loading time

When you first load up a Roam page, the loading time can be infuriatingly long. Sometimes I click on someone else’s publicly available Roam document and I begin wondering if the page has broken as the seconds tick by with just the obnoxious Roam loading animation mocking me. Fans of Roam encourage users to “keep a tab open for Roam” because of how universally bad the loading time is, but I find it impossible to stick to this habit, let alone that leaving a tab open doesn’t help when I want to open a second Roam window or want to check something quickly on my phone.

Access on the go

Speaking of phones, there’s no phone app, just a web page. I find it to be difficult to capture anything quickly using this page and especially difficult to find something quickly. Multiple times I’ve significantly regretted putting something that I need to access on the go into Roam. Enough that if I know I might want to see something when I’m out, I avoid putting it into Roam. Because I find it annoying and frustrating to take notes on my phone, I end up just never taking notes if I don’t have my computer near me. Some people have workflows where they take notes in another app on their phone and then later transfer these into Roam, but I can never get around to that. This is a significant issue that continually makes me doubt my decision to keep using Roam.

Steep learning curve

Many features are confusing, poorly documented, and difficult to remember how to use. I’m not very enthusiastic about having to find a ten minute YouTube video to try to remind myself how to do something or to understand a new feature. Does everything have to be a ten minute YouTube video, often made by some third party?

While Roam sort of allows you to write in a “freeform” way, much of the benefit of Roam also comes about through carefully planned structures. People have made countless different systems to make Roam work for them, and it seems like those systems do work for them. But, these systems are complicated and once you commit to a data structure, it’s nearly impossible to ever change.

Data security

I don’t have high confidence yet about security or data backup, and it’s easy to accidentally mess up a page with no good way to revert to the way it was an hour ago. I try to remember to export and backup locally, but this is a bit cumbersome. Neither of these things is atypical for a startup, but part of why I rage quit Evernote years ago was that they permanently lost some of my data through a bad sync.

Is Roam Research the next big thing?

I have very mixed feelings about Roam long term. The team behind Roam is obviously smart, talented, and motivated. At the same time, I don’t have the feeling that the team behind Roam has a clearly defined product roadmap. While I don’t wish to ignorantly criticize from my armchair, I’m often surprised by some random new niche feature while there are so many things that need improvement. I’m worried that they’ve coded themselves into a tough spot based on how little fundamental improvement I’ve seen since I’ve started using Roam. I worry that the Roam team built the application so quickly that now it’s difficult to make further changes, a problem familiar to any software developer.

Roam also seems to be putting a lot of hope into third party customization to fix up Roam’s issues, with custom CSS, custom Javascript, chrome extensions, or the always coming soon API. It’s nice to have community customization of an application and this is a real strength of a web-based product compared to a desktop application. At the same time, I find that solutions that rely on heavy community customization ultimately become overwhelming and less valuable — maintainers stop maintaining a project, people’s HOWTOs and other online information begin to depend on particular customizations, and different customizations conflict or break when the software is updated. Customizations are part of why some software is incredible, but I want a solid core product too.

And I keep going back to asking, What am I really getting out of this? I do like the ease of creating new pages and the ease of linking information together, but I’m worried that the lack of categorization and hierarchy is going to haunt me as the number of my notes increase. I’m considering building my own simple hierarchy to help with this, but doing so feels daunting and non-obvious, and it wouldn’t be easy to change. This far into Roam, I haven’t really experienced the promised serendipity from “networked thought,” and I’m not confident that this will ever truly emerge naturally from Roam usage. The lack of structure and best practices may ultimately hinder Roam.

I expect I’ll keep paying for Roam and using it almost daily for now, so I clearly like the product. At the same time, I keep wondering about something better. Most people are probably better off with something different, although that doesn’t mean there’s not a place for Roam Research. There’s a temptation to imagine that the next note-taking app will revolutionize your thinking and organization, which is why so many people try to create new ones, but I always find that the promises aren’t delivered. I hope the Roam team can ultimately deliver on the promise of Roam.

Determining the Right MVP

The Minimum Viable Product is a key strategy of the Lean Startup methodology, but many entrepreneurs struggle to implement the MVP in practice. The purpose behind the MVP is to maximize learning about the market with the least effort to avoid spending money and time building something that nobody actually wants. Discussion boards are filled with entrepreneurs battling it out over the best MVP because there is no simple formula: different products require different MVPs.

The Scientist Entrepreneur

To figure out what is necessary for your MVP, try thinking about your product as if you were a scientist. Think about what premises you hold about the market and your potential customers that you may not actually have solid evidence for. What hypotheses form the basis of your idea? What must be true for your product to be successful?

At the core of these hypotheses is usually a value proposition or “leap-of-faith” question. Your product solves a problem for some group of people or in some way provides them value. If we can do X, will people use us? How can you test that people actually have this problem and that you can solve it? If you have trouble answering this question, you’re going to have a hard time determining what MVP you should build. If your leap-of-faith question seems too convuluted and complicated – If we do A, B, C, D, E, F, and G, then people will use us – then you might be trying to do too much.

Some things can be tested first with basic MVPs, such as a simple signup page. Dropbox’s MVP was simply a video demo and beta waiting list signup page. But, an MVP is not always the smallest product possible and you eventually need a real product. Is there any way that a simpler product would appeal to a smaller group of people? Maybe most people won’t use your product without Super Amazing Complicated Feature, but you should be able to get a small group of people using it anyway. When Pinterest launched, it started with a small core community of users, who enjoyed using it even at its early stages. Instagram and Foursquare both used up their 100 person provisioning profiles before launch to test their MVPs.

In all of these cases, however, the companies were essentially testing a hypothesis. For Dropbox, the hypothesis was simpler: “People are interested in a product that seamlessly allows sharing files across multiple platforms.” They didn’t know yet if people would pay for it, but they at least knew people were excited about the concept. If your product plan will take six months to find out if people are even interested in your product, you’re probably not building an MVP.

Even apps that partially rely on having a large network to be fully valuable can often find a small group of users who will use the app early. If nobody is actually willing to use your app, be skeptical of the thought that you just need to find more users, that what’s needed is a big marketing push and launch. Zero active users multiplied by a thousand is still zero. Reddit didn’t have as much value before it got more popular, but some people liked it. Facebook is more useful now that you can connect with the whole world, but it still had value to users on isolated college campuses.

You should be skeptical of the thought that everything has to be polished before adoption by users. If your product really provides a value, some segment of users will be OK with bugs and an unpolished interface, even if you think your design or quality will be a differentiator. If you can’t find anyone who will use your unpolished product, ask why you really think a better design will change the value of your product. Maybe your hypothesis about the value of your product was wrong.

The biggest risk an entrepreneur faces is that you’ll make something that nobody wants. Repeat that and remember it. You need to know that what you’re building is actually wanted. Surveying your friends won’t tell you if this is true. To avoid making this mistake and avoid wasting your time, you need to test your assumptions early and fast. You need to ignore your feelings and examine your market objectively. An appropriately scaled MVP is the easiest way and has little downside. The costs of not doing so could be years of your life building something that nobody cares about.

Yes, it's better to succeed than fail. So what?

Bruce Nussbaum wrote an interesting post arguing that entrepreneurs should stop fetishizing failure. While he brings up a number of interesting points, I feel that Bruce has missed a large part of the point of those who chant “Fail early, fail fast, fail often.”

Claim: You learn as much or more from success

Of course it’s better to succeed than fail, but Bruce seems to be taking success as a single separate event. I suspect that if you look at most big successes in the world, there is lots of failure involved. Did they really learn more from their success, or did their failures enable their success?

I don’t believe that the failure “fetishizers” intend to mean that failure is the end goal. The purpose is not failure, failure is the means. You should not be afraid to try things that might fail because if you think an action is the right action to take, you shouldn’t let the fear of failure get in your way. If you do fail, find out why the failure occurred and what was incorrect about your initial assumptions. Nobody holds all perfectly valid ideas in their minds and the only way to correct invalid premises is to learn that they are wrong. A failure is evidence of an error in our knowledge and the fastest way to invalidate an assumption can be to act on it and fail. Is the alternative not to act on our beliefs because we fear we may fail and therefore prove ourselves wrong?

Claim: Failure has consequences and feels bad

Of course failure has consequences, often negative ones. Is “failure” more serious for a kid in an “urban public school” than for a well-off kid attending Stanford? Probably (although the consequences of not taking a chance may also be more serious for that urban kid). This is part of what the “fail fast” crowd seeks to alleviate. All talented, smart kids will experience failure no matter how hard they try. Rather than labelling a person a “failure” for a couple instances of failure, we can tell them, “It’s OK, many successful people have failed, and you can use this failure to learn more.”

Failure does feel bad. Failing to achieve a value or goal feels bad practically by definition. But Bruce’s approach would seek to only reinforce that and punish risk takers. Instead of simply feeling bad because you didn’t achieve what you wanted, you can also celebrate what you’ve learned from the risk you took.

Fallacy of Failure

According to Bruce, the fallacy of failure is that failure is really only celebrated when you ultimately succeed. Aristotle said “we are what we repeatedly do,” so if we are always failing, don’t we become failures? But, Aristotle also believed that a person only found virtue through trial and error in the experiences of his own life. We have to have failures to improve ourselves and succeed. This alleged “fallacy” of failure is really just evidence that failure is not the intended purpose, but the necessary means to achieve sucess.

People who never risk failures are probably the people most commonly viewed as “Failures.” The rest of us are making mistakes and trying new things, sometimes with big consequences to ourselves. The best way to avoid becoming a “Failure” is to accept failures as a part of life and part of the path to success, and then to learn from our failures.

Redefining Failure?

At the end of his post, Bruce makes an interesting clarification about what he means. He argues for a “play” approach rather than a failure-focused one. There are different outcomes, different ways of winning, and when something doesn’t work, you try another. He says this is not failure, although the consequences to an “urban” kid who tries things that don’t work may be indistinguishable from the consequences to a kid who tries something and “fails” at it. Bruce asks, “Do kids who go to Montessori school think of themselves as Failures when their blocks don’t quite fit together?” The answer is no, because kids at Montessori school learn that little failures don’t make you a Failure.

Unlike traditional educational enviroments that emphasize achievement, performance, and results, the Montessori school of education believes in failure and mistakes. Children are not punished or reprimanded for failure, even failures as big as breaking classroom objects. Mistakes are necessary educational tools for children to develop in a Montessori classroom. Dr. Montessori is quoted as saying “it is well to cultivate a friendly feeling towards error, to treat it as a companion, inseparable from our lives, as something having a purpose, which it truly has.” In other words, as a Montessori educator puts it, “Nobody praises his achievement, but nobody needs to. The lessons learned from his mistakes, and his ensuing success, are his rewards.”

Yes, it’s better to succeed than fail, but to succeed, you’re going to have to fail along the way. Remove the possibility of failure and you most likely remove the possibility of success as well. Fail early (rather than later), fail fast (so that your failures are small and you learn from them quickly), and fail often (keep taking new chances and pushing yourself).