The Tor Project

User Growth Strategy

2015-2016

isabela@torproject.org

User growth through user experience

The best way to assure user growth is to fix broken user experiences quickly, so more people can use your product.

User experience can break not only in the product UI, but prior to even knowing about the product (awareness/marketing), or when the user decides to give it a try during the download and onboarding process.

It’s also about the experience while using the product. Does the user obtain value from their time and effort? What are the factors that keep people using your product?

​

Tor’s approach

Most of the industry uses user behavioral data as the main source of information driving decisions on fixing broken user experience, from identifying problems to testing and validating hypotheses.

Tor’s approach is different because we refuse to collect this type of data. Our approach is built on respecting and safeguarding the privacy of our users.

We will do observations and hypothesis testing in the safest way possible, which in most cases will be in person with our users, not by collecting data on them.

Some thoughts on growth

Tor users are measured by the number of clients connecting directly to the network (~2M/day), and connecting through bridges and PTs (~25k/day).

This measurement does not capture how the user is accessing it (Tor Browser, Orfox, chat client, TAILS, etc). We assume Tor Browser is the most popular method.

Demand for privacy is growing quickly with the increase of data creation and rising understanding of the implications of not being able to control who can access such data.

​

Some thoughts on growth

The good news: Tor’s user facing products have some “low-hanging fruit” of broken user experience that would be comparatively easy to fix. Unfortunately, the major risk for these opportunities is the lack of funding.

“Discoverability” is an issue, but awareness can be easily increased, as Tor has a lot of good support and media attention. But users still have some trouble finding and using Tor, for the different use cases around the world, especially if you add mobile to the story.

Some thoughts on growth

User expectations: it’s very important that the user understand as much as possible about what is going on, why things work the way they do.

This is crucial not only for creating a good experience but also to guarantee the user’s security.

The more the user understands the context of secure communications, the better we can set the right expectations. This must be the fundamental approach to building a good relationship with our users.

Growth is a multidisciplinary effort

The Tor community is growing as well! The most recent Dev Meeting was over 100 people, and it still missed many who were unable to make it.

Growth involves collaboration between different teams, where everyone comes together to work for the user, including:

Designers, UX Researchers, Metrics team, application developers, UX developers, network developers, Localization, User Support, Trainers, PR, etc.

We need to build a culture of organizing all these efforts together, to better serve our users.

Current Tor Organization (Vegas Plan):

Plan and time frame

Re-org, recruit (community) and hire new roles as necessary to build the teams to success.

Build user-centric culture within the organization, recognize ongoing efforts by empowering them.

Execute low-hanging fruits opportunities - goal is to do them by Jan 2017.

Build a model that can be reproduced with other products/platforms.

Goal

In terms of user growth, our effort is to help our users. We are not motivated by fast growth per se, but by doing the right thing that will enable people who need our tools to obtain and use them.

But we do plan to document our wins and celebrate the growth that results.

In terms of execution - we hope to have a model that can be replicated with other teams, products, platforms, etc.

​

Building teams for success

2015-2016

Quick summary of initiatives per team

Network team: reorg and possible headcount additions.

Applications team: possible headcount additions and increase collaboration with Mobile.

Metrics team: possible headcount additions and increase collaboration across teams.

Usability team: create the team, mobilize different members of community, and possible headcount additions.

Network Team

Applications Team

Metrics Team

Usability Team

Building a user-centric culture

2015-2016

How others do usability

In a normal scenario, we would be talking about a company who owns the full user path to its product.

The in-house team (product, UX researcher, Design and Engineer) would analyze data about their users, build hypotheses for their problems, test them by using some experimentation framework. Based on results, the team would go back to zero or ship the best version.

How Tor does usability

The challenge: At The Tor Project, we have a very particular scenario that changes how all this is done.

Our user sometimes interacts with three or more different projects not directly maintained by The Tor Project, before using our product.

The projects that the user touches are maintained by Tor Project paid staff, volunteers, Ph.D. researchers, other organizations - people in different locations, with different capacities, priorities, funding, etc.

Embrace the culture

The solution: Fortunately, at Tor there is already a good awareness of the need to have a user-centric development process. It won’t be a problem to get all these teams, volunteers, etc. behind the vision.

Tor Project’s culture involves many things like free software, transparency, cooperation, etc. These should be taken as a benefit, not a challenge.

The best way to build a user-centric development process for Tor, is to do it following Tor’s culture.

Tor Project’s usability team model

The Usability team is not embedded to any developer team. It “sits between” the users (problem reporters) and the developers (solution builders).

It will operate like any other Free Software project: the source of their work (icons, logos, brand guidelines) should be shareable with other teams who might want to use it in their product.

Problems and solutions should be documented publicly, so the evolution of the product can be easily understood by anyone who wants to help.

​

Tor Project’s usability team model

Usability Team

Designers

Product

UX researcher

UX Dev

Localization

Copy Writer

The goal of this team is to help solve user problems.

• Works closely with users, helps identify broken UX
• Helps trainers learn about user research techniques to help report broken UX
• Works closely with Ph.D. students doing UX research
• Works closely with native speakers to improve localization

User side:

Developer side:

• Helps promote interaction between developer and user
• Works closely with developers to promote adoption of UI improvements
• Helps prioritize features according to user needs

UX open sourced

The Usability Team will operate as a Free Software project.

All resources produced by this team will be available for any developer to use in their project. Icons, brand guidelines, typography, and even suggestions of copy will be available in a repo for developers.

The goal is to bring consistency to the visual language of the products in the Tor ecosystem.

This will help the user understand similar functionalities across products.

How Tor does usability

We plan to have a page dedicated to usability team work.

This page will document case studies by explaining the problem, hypothesis tested, and why the chosen course was selected.

We will also have a list of “open problems” in order to reach out to the broader community for help on building solutions.

As an organization grows, is very important to build this institutional memory so the product evolution is understood by everyone.

Harvesting the low-hanging fruits

2015-2016

The product

To kick off this strategy, we decided to focus on Tor Browser (desktop).

We will establish the “happy path” - the first, most logical path the user can take without any problem to get Tor Browser and start using it. Then we will work all the variations of it and also work locally with users to understand their needs.

We will also look at retention factors, like knowing how to change settings to improve experience (e.g. how to try different obfuscation mechanisms or different circuits to increase connection speed).

The happy path

Fruit #1 - the website

Website needs to be simplified and focused on main message to user.

By simplifying, we will be able to easily localize the website.

Tor is a global product that aims to simplify a complicated concept.

The fact the site is complicated and not localized is a huge source of broken experience.

First concept discussed at Tor Dev Meeting

Fruit #2 - download flow

We will make it easier for users to find what they want by building a responsive download flow.

We plan to build our site with responsive design.

This will allow us to be smart and provide the right path for the user according to the platform they are using to access our site.

Fruit #3 - GetTor

GetTor is a service that provides alternative methods to download Tor Browser, especially for situations where access to Tor Project's official website and mirrors is blocked. For now, the only available method is SMTP (email).

​

We plan to expand the delivery methods and localize them.

​

We also plan on facilitating access by increasing its promotion.

​

For instance, by editing our information for search engines, to include a phrase at Tor Project’s site description explaining that in case the user can’t open Tor’s website, they can still get Tor using GetTor service.

Fruit #4 - Tor BridgeDB

Today, this use case takes at least 5 steps for the user. Some of them involve the user having another browser, or email option..or..

​

We plan to fix that by performing all of these steps in the backend. Just like the Transport behaviour, the user will select the bridge option and we will pick a set for them, automatically.

​

This is an area we can improve the user experience significantly. The work described above is just one of the implementations that can automate the process of bypassing this type of blocking of Tor.

More on backlog

Tor constantly receives feedback from users, trainers, researchers, etc. We have more opportunities on our backlog, which we plan to execute giving the team capacity (always looking to increase it).

There are a lot of efforts we can make on the “user awareness” front, such as building partnerships like .onion Facebook website or helping big websites support Tor users (e.g. Wikipedia, Twitter).

...

A reproducible execution model

2015-2016

Documentation and knowledge sharing

We plan on building a website to host documentation about our UX work, how to participate, list of open tickets, and projects “in-progress.”

We will also have materials to help others to learn how to execute some user research techniques such as needfinding.

Tor has a diverse community of trainers and translators. They are where our users are and know their culture and needs best. We will work together with this community to train them on UX to help us identify broken experiences.

Partnerships

We have done outreach to organizations around Tor’s ecosystem with a diverse set of skills (e.g. Simply Secure, SecondMuse, University of Berkeley, etc.) to reach out for partnership opportunities to help us execute this strategy.

The community has shown great interest, and we are looking forward to increasing and strengthening these relationships.

Building a network of collaboration will serve other projects as well.

Quality assurance

This is another important point for us. We are organizing with HackOne and OTF to have a Bug Bounty program for Core Tor and Tor Browser.

We hope that this program can help us guarantee another important side of user experience: the security side.

Tor already has a great cadre of researchers who test its security, and bugs reported by the free software community, so it will be great to add another point of quality assurance to our products.

Risks

2015-2016

What if

.. we don’t find funding for all these initiatives? We would probably still execute it, but it would be done much more slowly with insufficient resources.

.. Tor gets funded but one project of its ecosystem doesn’t, and it affects the user path enormously? We would probably pursue two strategies at once: 1) continue to look for funding (maybe crowdfund it); 2) Look into keeping up with any work with minimal resources.

Benefits

2015-2016

It could

… be a model used by multiple privacy tools, to building consistent UI across products/platforms, make it easier for users to onboard themselves on different apps in different phones or desktop operating systems.

… help millions of people around the world that are not capable of using Tor because of lack of localization support (current website is in English only).

… solve the problem: “Security/privacy is hard.”

… add more users, possibly more donors, possibly more independent revenue for Tor.

Questions?

2015-2016

isabela@torproject.org