I joined Instagram as their first product researcher in 2014. In addition to establishing a culture of research at the company, I was hired to build the product research team. Today Instagram has a team of 15 product researchers — 13 full-time and 2 contractors.
I often get asked how I went about building the team by people who are starting to build teams themselves. In this article, I describe how I built the product research team at Instagram. My hope is that some of the approaches I have taken and lessons learned will help others. While my experiences are building a product research team in the technology industry, I believe some of the lessons learned are generalizable to other disciplines and industries.
What type of team do you want to build?
Before building a team you need to figure out what type of team you want to build. What is its goal? What is its purpose? Peter Deng, my manager at the time and Director of Product Management at Instagram, and I discussed how to build product at Instagram, taking inspiration from the d.school and IDEO design processes. Instagram’s design, product management and research — me — teams even went to the d.school to experience their process first-hand.
Understanding is the foundation of the d.school and IDEO design processes. In order to build a product — a solution — you need to be solving a problem and that requires identifying and understanding the problem. Anything else is a solution in search of a problem that is destined to fail.
Maryam Tohidi, Bill Buxton, Ronald Baecker and Abigail Sellen argue the importance of getting the right design and the design right. Most product research teams focus on getting the design right through iterative testing e.g. usability testing. However, unless you have the right design, iterating on the wrong design is simply making a bad design better.
We wanted Instagram’s product research team to focus on understanding problems that would allow us to build products that were the right design.
There’s no I in team.
The product research team is one input to identifying and understanding problems. Peter once described to me that the people building a product were like musicians playing in an orchestra. There are different sections — the brass, percussion, strings and woodwinds — who all play a symphony.
Before building the product research team I aligned with, Adrien Friggeri, a data scientist working on Instagram, as well as, Raman Thirumalai and Rodrigo Schmidt, our heads of analytics and data engineering at the time. Together we created a strawman for how to identify and understand problems, as well as how we would work together to do so.
I also met with Andrew Tresolini Fiore, Christina Holsberry Janzer, Judd Antin, Pete Fleming, Pratiti Raychoudhury and Sarah Sled who at the time were the heads of research at Facebook — Instagram’s parent company — to learn how they built their teams. I also met with the researchers on their teams to understand what they did, what worked well and what could be improved.
These inputs defined the contributions of the Instagram product research team and hence the team we wanted to build.
As a result the Instagram product research team is closely aligned with our analytics and product management teams, as well as the larger Facebook research team. There is an opportunity for the team to more closely partner with design and engineering, which could be because I was not more closely align with those teams from the beginning—although back then Instagram did not have heads of design or engineering.
Actions speak louder than words.
One of the ways to convey the value a team will provide is to do the work yourself — set the standard.
My first project at Instagram was to understand how brands evaluate posts shared on social media platforms and get feedback on a few design explorations, Chris Marra, Jeff Kanter and Josh Williams, two product managers and a designer at the time, were considering for Instagram’s first insights product.
This project defined the way Instagram would do research and its value. While product had a vision for what the insights tool could be we took a step back to understand brand’s approach to evaluating posts shared on social media platforms, which would highlight what was most important to them and could inform design explorations we had not yet considered.
Rather than waiting for a high-fidelity prototype or build of the insights tool, we printed unfinished wireframes Josh had made of the insights tools, in addition to having multiple alternatives. This conveyed we were still very much in the design process and open to feedback. Josh also iterated on designs between talking with brands so we could further explore design alternatives to address issues we were identifying.
With this project I wanted to establish a culture of research understanding needs, being involved early in the process, partnering with product teams and getting the right design.
Learn to replace yourself.
While conducting research to establish a culture was important, I alone could not address all the research needs of the Instagram team.
In the early days of Instagram, Mike Krieger, Instagram’s CTO and co-founder, described how he would spend a lot of his time keeping the servers up and running. However, this prevented him from developing the Instagram app further. This problem was exacerbated as more people started using Instagram — it took more time to keep the servers up and running. Mike uses this as an example of why hiring is a vital part of scaling a company, requiring attention and time.
When I started at Instagram each month I would meet with the product managers of each product area, which at the time were called Engagement, Growth and Monetization. This allowed me to understand what their priorities and research needs were, as well as discuss how research could impact product. It was impossible for me to do everything that was being asked of me. As such, I scoped the research needs to one or two high priority projects where research could have the most impact on product. This allowed me to continue to develop the research culture at Instagram, impact product and have time to hire the product research team.
Peter Deng taught me that in order build a team you need to learn how to replace yourself. This can be a scary concept. If you replace yourself, are you not putting yourself out of a job? Peter explained that there is always something else to be done and replacing yourself allows you to identify this and add value in new ways. It took me time to accept this concept, but it is what has allowed research to scale at Instagram.
A lot of my work in the beginning was working on the experience for the core Instagram app — the Engagement product area. As this is what took up most of my time I hired someone to do this work, which allowed me to focus on other areas — Growth and Monetization for Instagram, as well as other apps Instagram was developing such as Bolt, Layout and Boomerang — and identify how to build the product research team for these areas.
I mentioned that when developing the Instagram design process we took inspiration from the d.school and IDEO. As such, my initial thinking was to hire a team of design researchers from companies such as IDEO.
However, this was rather limiting for a number of reasons. First, limiting yourself to such companies and skill sets limits the potential hiring pool. Second, companies such as IDEO are consultancies designing more blue sky concepts, which is different from in-house teams working on features or products to be launched within months.
Different teams need different skill sets; and different research methods are needed at each stage of the product development process. For example, Instagram’s Growth team are more incremental, iterative and metrics driven. While ethnographic techniques have had impact defining product roadmaps (i.e. what we are going to build), survey research which is more representative of a larger population has allowed the team to make decisions with confidence. Furthermore, while design research techniques are effective early in the process to get the right design, usability testing is more appropriate method to get the design right.
As Instagram’s product research team has grown, the team's diversity of backgrounds and skill sets has allowed the researchers on the team to suggest different approaches to research, as well as learn from each other expanding their research toolkits.
Completeness, not coverage.
When I started at Instagram I worked across all of Instagram’s product areas. Whenever I finished a project I moved on to another project with a different product team. This made it hard to establish trust and a working relationship with the team. As such, I was more of a service than a member of the product team.
My goal building the Instagram product research team was to have research be a core part of the product team in order to have a sense of accountability and ownership.
As I have hired more researchers the product research team has been able to move from supporting multiple product teams in a product area — coverage — to working on a dedicated product team — completeness.
At Instagram, research is now a core part of the product team along with analytics, design, engineering and product management, present in all product planning and reviews.
Experience is important.
When hiring the first researchers we hired people who had experience working with product teams. We needed people who could handle ambiguous situations and were self-driven. People who could work with multiple product teams to identify, prioritize, scope and then execute on research projects with little to no oversight.
Why would you not hire an entire team of people like this? More experienced people want to work on projects with larger scopes that are more complex and they also want to mentor people, eventually moving in to lead and management positions. This is not possible if you have a team of experienced people.
As such, once we established some minimum research coverage across the Instagram product areas, we hired our first intern and also started to begin hiring new grads. This was not just for the benefit of the more experienced researchers. It also allowed for new perspectives to be brought to the team, as well as the latest techniques being taught in grad schools.
Considerations building a team
To summarize, when building a team some things to consider are:
- Define the goals and purpose of the team;
- Align with other teams and discuss how the teams will work together;
- Establish the value of the team;
- Evaluate what you spend your time doing and hire people to do that;
- Hire people with different backgrounds, perspectives and skills;
- Don’t build a team as a service;
- Hire people with varying levels of experience.
These are some of the lessons I have learned building the Instagram product research team from 1 to 16 product researchers. As the team grows we will be facing new challenges, such as managing managers and growing the team beyond a single location.