User Adoption and the Decentralized Web

In order to build a protocol with broad and lasting impact, we need to understand people use it -- or not -- at every layer of the decentralized web stack.

My goal, 19 minutes from now, is that no matter who *your* users are, whether it's end users or developers, you leave with better frameworks to think about they choose decentralized technologies, and grow your usage.

Transcript

1. Michelle Lee Protocol Labs @mishmosh User Adoption and the Decentralized Web Hi everyone. My name is Michelle Lee and I work at Protocol Labs. We exist to improve upon how the internet works.
2. protocols services applications Decentralized Web Stack …and its users end-users individuals or orgs developers developers Our focus is on the very fundamental layers — the protocols. In order to build a protocol with broad and lasting impact, we need to understand people use it -- or not -- at every layer of the decentralized web stack. - How protocols are used by developers to build services and networks - And then how those services are used by other developers to build applications for end users - And then how end users seek and use applications My goal, 19 minutes from now, is that no matter who *your* users are, whether it's end users or developers, YOU leave with better frameworks to think about they choose decentralized technologies, and grow your usage.
3. Want a bike Go to bike shop Try 3 bikes Buy bike Ideal Path ✔ As it turns out, "how people choose decentralized technologies" is a subset of a more general problem: "how people choose things." Let's start with a simple example: buying a bike. [Read through steps.] But does it really work this way? Raise your hand if in the past 2 years you've thought about buying a bike, but haven't done it.
4. Wow, those bikes are all going faster than my subway Want a bike Test bikes Specs for 18 models by 3 brands Bikeshare instead of buy? Too heavy Too fancy Maybe I need TWO bikes Not this summer Hey, does friend bike? ✘ Actual Path Maybe your journey looks more like this. [Read through steps.] ...I'll try again next year. Sound familiar?
5. Want a decentralized chat app Search for app Download app Use the app Ideal Path ✔ Now, let's apply that to using decentralized tech. One very lucky person gets this nice ideal path. You’re inspired to try a decentralized chat app, search for that online, download the app of your choice, and get chatting.
6. Hears about libp2p at a meetup, sounds weird libp2p lightning talks Asks livepeer about their experience Helpful blog post Unexpected timeouts Try it in a branch New project can use it Wooooot! Try it locally ✔ Actual Path Now, let's look a decentralized web protocol this way. You first hear about libp2p at a meetup. It sounds interesting and kinda weird. A few months later, you’re at the Decentralized Web Summit, and you hear a bunch of talks about it, so you decide to try it out. First you try it locally, and that’s your Friday afternoon. You think it might be really great for this other project you’re working on, so you look up some blog posts, ask your friends who’ve used it, and then try implementing it in that project. After some stumbling blocks, 9 months later of stops and starts, you’re up and running! This looks insane. But it's real. These are the hurdles and stumbling blocks users come across before they can use your project.
7. Hears about libp2p at a meetup, sounds weird libp2p lightning talks Asks friend for their experience Helpful blog post Unexpected timeouts Try it locally New project at work can use it Wooooot! No compatible projects ✔ Adoption Funnel 1. Awareness 2. Interest 3. Evaluation 4. Adoption But don't worry. We can break this down into phases.
8. Hears about libp2p at a meetup, sounds weird libp2p lightning talks Asks friend for their experience Helpful blog post Unexpected timeouts Try it locally New project at work can use it Wooooot! No compatible projects ✔ Adoption Funnel 1. Awareness 2. Interest 3. Evaluation 4. Adoption Actually at each phase, you lose some people, so each box is a trapezoid and they get smaller as you move right. This is called the adoption funnel.
9. Hears about libp2p at a meetup, sounds weird libp2p lightning talks Asks friend for their experience Helpful blog post Unexpected timeouts Try it locally New project at work can use it Wooooot! No compatible projects ✔ Evaluation works differently. 1. Awareness 2. Interest 3. Evaluation 4. Adoption The first 2 stages are pretty similar for both apps I use in my personal life, but the 3rd stage is very different.
10. Awareness talks forums blogs friends The defining trait is that people haven’t gotten to your website yet. So you need to reach outside your website, github, & IRC Attack vectors: conferences, podcasts, other blogs, forums, word of mouth Metrics: website mentions, HN points, twitter followers Let's take a look at each stage.
11. Interest Now, people are intrigued and want to know a bit more. They’re checking out *your* stuff: website, github, app store Attack vectors: case studies, teach how the system works vs. command-line, app store listing Metrics: web traffic, docs traffic, github stars, app installs website app store github
12. Evaluation for applications Deciding if this works for me 1 stakeholder (me) Attack vectors: account creation, onboarding <3, app store listing Metrics: accounts created, app installs install try
13. Evaluation for services & protocols Hears about libp2p at a meetup, sounds weird libp2p lightning talks Asks livepeer about their experience New project could use it Unexpected timeouts Try it in branch Helpful blog post walkthrough Wooooot! Try it locally ✔ Okay, I said Evaluation was the most different stage for protocols & systems vs. apps. Remember that crazy diagram of steps?
14. Evaluation for services & protocols ✔ My decision Other stakeholders Maintenance + Compability (avoiding future worlds of pain) It's actually crazier. Imagine if, instead of a bike, you were getting a pet. Then the constellation of steps increases. You have to convince your partner. You have to convince your landlord. You've got *other stakeholders.* Imagine you convince them, and get this dog. You have to walk it every day. You need a dogsitter when you travel. You have *maintenance requirements*. So the original crazy path is just 1 of 3 parts in evaluating whether this dog, I mean this protocol, is the right choice for you. Alright, that's enough pets, let's go back to protocols.
15. Evaluation for services & protocols Okay, it’s interesting. But is it useful? Will it solve my problem? Attack vectors: docs, guides, overviews, forum Metrics: npm installs, github repo counts think abt stack Many, many stakeholders (current + new OSS contributors) read docs decide with others Maintenance & upgradeability
16. 1. A user’s journey to using your stuff is rarely linear. 2. But you can divide & conquer into 4 phases 3. Stage 3 (Evaluation) is *very* different. Services & protocols have more stakeholders, more maintenance considerations, etc. than apps. Three takeaways on user adoption Three takeaways from this.
17. After building, comes adoption HTTP created 1990, mainstreamed 1993 (4 yrs) HTTPS created 1994, mainstreamed 2018 (24 yrs!) IPv6 created 1996, still at 25% in 2018 (22 yrs) Signal created 2013, 100M users in 2016 (3 yrs) Airplanes invented 1903, popularized 1950s (50 yrs) Now, here’s the part that I don't know. How long will it take for all those user journeys to happen? How long will it take for the decentralized web to become the default web? A small sample of adoption times raises some interesting questions. [Airplanes, Signal, HTTP, HTTPS, IPv6]. They vary widely from 3 years to 50 years. The longest cycles seem to networks, where a critical mass is needed — see airplanes & airports, or IPv6. Applications, on the other hand, can be adopted really quickly.
18. Michelle Lee Protocol Labs @mishmosh User Adoption and the Decentralized Web If this type of problem solving & exploration interests you, I’d love to talk with you more, and maybe work with you on solving them. (We’re hiring!) I’m mishmosh on Twitter, and I’ll be here at Lab Day all day. Thanks!