RIP materialsGIRL

I started materialsGIRL because I couldn’t find an online community to discuss materials science. This seemed like a huge ‘unmet need’ to me, but ultimately I failed to build a critical mass. Here is the welcome message on materialsGIRL for context:

Welcome to materialsGIRL - a community enthusiastic about making materials.

We encourage anyone interested in learning about materials to join, regardless of your education, background, discipline or specialty. GIRL stands for Group Interaction Reinforces Learning - we hope to develop tools to promote scientific understanding, collaboration & discourse. Materials science is an interdisciplinary field, we hope curious students, chemists, physicists, biologists, engineers, artists, designers & makers will engage in this experiment.

materialsGIRL logo and header

After about ten months of trying to make things click, I ultimately decided to throw in the towel. I learned a lot along the way, and I thought it would be fun to share some of that:

1) Building a community is hard

With the exponentially-increasing number of competing demands for our attention in this hyper-connected world, building a community is hard. Hosting a community has never been easier - just ‘one click’ on Digital Ocean to a new Discourse image - but the corollary is that now everyone is doing it. In the same way that you don’t need a record label to make an album these days, but having a record label really helps if you want anyone to listen to the album - content is easy and community is hard.

If you’re thinking about getting started, I would encourage you to check out Jeff Atwood’s advice on building a Discourse community. There are also a few questions I think you should ask yourself before you start, and periodically throughout the first year or so:

  • Are you passionate enough to persist? Can you overcome the dip?
  • Are you really providing valuable content? How is it unique?
  • Are you really encouraging participation? Believe it or not, you actually need to respond to the other people in your community (preferably in an encouraging way).

2) You need to get past the ‘awkward house party’

You get invited to a house party. You don’t know the host. You don’t know of anyone going other than the one friend that invited you. You park the car across the street. Before you get out of the car, you can see through the big dining room window a rather depressing scene: two people, sitting against separate walls, heads down, alternating their attention between the solo cups in their left hands, and the phones in their right hands. Do you even get out of the car?

This is my favorite ‘IRL’ analogy to one of the most frustrating aspects of the power law (Pareto) distributions of online attention. You have to have followers to get followers. There has to be a discussion to spark a discussion. It totally sucks when you are trying to launch something new, but hopefully the house party analogy can get you to appreciate the perspective of the other side.

One of the best stories I’ve heard about overcoming the ‘awkward house party’ phase is in the founding lore of Reddit. Basically, the founders created a ton of fake users and had conversations with themselves to make it seem like there was something going on. I have no idea how long they did this for, but overcoming the awkward house party was probably the single biggest risk for Reddit at the early stages of the company. I tried this on materialsGIRL briefly - I found it to be a very humbling exercise.

First, it is legitimately challenging to pretend to be two different sides of a conversation. I would recommend this as an interesting writing exercise. It reminds me of Demosthenes and Locke from Ender’s Game.

xkcd Locke & Demosthenes

In the end, however, I found this simply too depressing and too much work to justify continuing. Feeling like I was tricking people into coming to my super lame house party made me less excited about what I was building. If materialsGIRL was a business, rather than a fun side project, perhaps I would have had the grit to stick it out. If you’re the only one(s) at the party for a while, you’re probably doing something wrong. Change it up or quit before it gets even worse.

3) Social media is a bad place to start

Related to #2, let’s face it: for 99% of people trying to build something new, social media is a terrible place to start building their user base. The power laws are not in your favor. Paying for clicks is expensive, working for clicks is depressing, the conversion rates are low, and the actual amount of meaningful engagement you will get from social is effectively zero at this early stage. Social media works to let happy users/customers tell other potential users/customers. Social media might work if you’re selling a mobile app or a social media product, but if you’re trying to build a community around a niche topic - fuhgetaboutit.

I’m not saying don’t make social profiles. I’m not saying don’t share what you are working on. I’m saying don’t expect anyone to give a single shit and don’t expect to seed your user base by blasting out a ton of tweets. Don’t waste your life away on Twitter/Facebook.

You need to start with communities of mutual interest. The early users we did get on materialsGIRL came from adjacent communities:, my friends from graduate school, our LBL colleagues, people I was reaching out to directly for my work at polySpectra.

As a post-mortem, I would say that the experience was 100% worth it, and I hope that this post spares you some agony in your next endeavor. One really valuable thing about materialsGIRL was that it gave me a new channel to learn out loud. Thanks to the magic of Markdown, it took about 20 seconds to copy/paste my favorite posts from the site here. You can find them with the tag matGIRL. It also helped me build confidence in running Linux virtual machines and web apps, something which I still suck at, but have come to view as an invaluable skill.