Learnings from My "Tours of Duty" at LinkedIn

Inspired by LinkedIn’s vision to “create economic opportunity for everyone in the global workforce”, I joined LinkedIn 2,673 days ago [can you calculate my start date within 30 seconds? :)]. I have worked at LinkedIn for over seven years. It has been my most exciting and rewarding professional experience by far. LinkedIn has completely transformed my career; from being one of the earliest people working on business analytics for monetization to leading a world class analytics and data science team globally. I’m thrilled to see that data and insights have become essential for almost everything we do at LinkedIn. As I reflect on my tenure, I want to share a few of my key learnings from this journey.

Data can add value everywhere, in many different ways. Our team collaborates with almost every function in the company. Over time, we have come to realize that data can create value in many use cases, but the hard part is identifying the what, when, and how. This often requires data scientists to have strong domain knowledge and understanding of the most suitable technologies available, and come up with creative solutions to solve various problems. One example is the enterprise sales process, which was manual and nuanced. With a strong partnership and close engagement with our global sales organization, we were able to understand their pain points and create the analytics intelligence portal “Merlin”. This helps provide customized insights to our sales teams in finding the best timing to reach out to potential customers and with the right content at scale. This has been a game-changer in improving sales productivity and effectiveness for the last several years.  

There is a learning curve for people to understand how to leverage data-driven insights in the best way, so always seek ways to prove value quickly. In theory, almost everyone believes that data helps us make better decisions. In practice, however, data can become an afterthought, especially when the business is doing well and growing fast. There is also the reality that building a solid data foundation takes a long time, but we usually cannot wait for it to finish  before starting to leverage it. We have to identify opportunities to demonstrate how valuable data can be, with the existing problems and while having a vision for long-term solutions. For example, LinkedIn’s experimentation platform can support thousands of tests at any given time and is an indispensable component of all our product launch decision making. However, it was only six years ago that we were leveraging A/B tests for few products manually. The methodology of experimentation setup could be flawed at times, resulting in unreliable results on which product managers could not rely to make data-driven decisions. An important trigger that led us to start embracing the idea of experimentation systematically was identifying a key issue in the homepage design through A/B tests, which would cause a significant revenue drop if it were to be rolled out. We started to build momentum from there and now “make experimentation ubiquitous” has become the core strategy of our team.

A centralized Data organization is a strategic way to get the most value out of data for the company. Data organization is still a new concept for many companies, even in Silicon Valley. LinkedIn is one of the pioneers in having a data team that combines Analytics, Data science, Machine learning, Data engineering, and Data infrastructure. The best advantage of this structure is having end-to-end visibility, right from data creation all the way to data consumption. This puts the team in a great position to think strategically, balancing between quick wins and sustainable long-term solutions. It also makes it easier to execute what “has to be done”, such as legal compliance around data, data governance, fairness and transparency about using data, etc. In a world of exponentially growing data and increasingly stricter regulations on data, I can imagine a strong data organization becoming the number one choice for more and more companies.

Talent is our No. 1 priority. This is the aspect of the LinkedIn culture that I can relate to the most. I firmly believe that talent is the most valuable asset for a team; I personally spend a substantial portion of my time on recruiting, coaching, mentoring, and creating a culture and environment for people to grow and thrive. Hiring the best people and creating space for them to have the autonomy to contribute in a way they feel most comfortable is the best way to maximize the productivity and efficiency of a team. In addition, being alert about areas needing improvement, providing on-time coaching, and clearing roadblocks can help them be more productive. Building, developing, and managing a large team of super-talented people has been the most enjoyable part of working at LinkedIn. The process of scaling up a team in the ever-changing landscape of technology has also been very challenging and extremely rewarding.

What unites a team is the mission, values, and culture. As the team grows in size, it becomes increasingly important to create an inspiring team charter, provide clear guidance on how team members are evaluated and rewarded, and set great examples in practice. With extensive discussions and debates among the leadership team, two years ago we chose the mission for the team to “Drive understanding and impactful decisions through the rigorous use of Data”. This has been the true North Star, helping guide the team of 250+ people and growing. We have been evolving the team strategy and the operational mechanisms along this mission to rally the team in the same direction. Our mission also inspires the team to “amplify a data-driven culture for the company” as part of our responsibilities. We have launched initiatives such as Data Boot Camp, Data Hackathon, and Analytics University, which have helped create a much broader influence beyond just the team.

Executive support is critical for the team’s success. It goes without saying that this is an important consideration for a team that does a lot of behind-the-scenes work. It takes a village to extract value out of data. We need to perform various types of tasks, often in collaboration with other cross-functional teams, to achieve the results that we would like to see. Some of our work cannot be directly tied to business impacts and hence can be overlooked or even questioned, especially in comparison to priorities having more obvious business outcomes. Executive support, by trusting the team’s capability and judgment, is the best approach to enable and motivate the team to do what’s right for the company.

To put some of these learnings into perspective, I list my four “tours of duty” (the original concept by Reid Hoffman here) at LinkedIn:

• My first tour of duty started from the day I joined LinkedIn, working with our global sales organization to bring data and intelligence to the B2B sales and marketing process. From the very beginning, we have been focused on bringing insights at scale. To achieve this, we built an analytics intelligence platform called “Merlin”, which has transformed how data is used to increase sales productivity and customer success in the sale process. Some of this work is captured in this presentation.
• My second tour of duty started when we decided to align the Analytics team (earlier Business Analytics) with a line of business, to maximize the value that data can create for the overall business, from Talent Solutions to Sales Solutions and Premium Subscriptions. This role required me to think strategically about how to balance business needs across different functional teams, and prioritize what has the largest impact for the whole business across the various kinds of work we do. I came up with the EOI framework around that time to help us prioritize our valuable time and resources.
• My third tour of duty started when my responsibilities expanded to lead all Monetization Analytics teams along with Consumer Marketing and Market Research. This further broadened my scope and really challenged me to think about the teams’ contributions and the right areas to focus on, at the overall business level. Creating the right team structure and developing capable leaders has become really important in ensuring the sustainable growth of the team.
• My fourth tour of duty started two years ago, when I took the responsibility of leading a centralized Analytics team (merged Business Analytics and Consumer Analytics, previously known as Data Science). This is the most challenging of jobs I have ever done. Being the head of a function in a company requires a much higher caliber, a vision for the strategy for the team, and maximizing the value generated from the team. I have made many efforts in establishing the team strategy and operating mechanisms, including a standard and consistent hiring process, career tracks for individuals and management, and evaluation criteria specific to each track. My recent blog describes a high-level picture.

As my LinkedIn journey comes to an end, I hope to take this opportunity to express my gratitude to my managers—Simon Zhang, who hired me and inspired me to always challenge myself to become better; Laura Dholakia, who coached me on business strategic thinking and practicing compassionate management; and Igor Perisic, who guided me on unifying teams from different backgrounds and running a transformational Data team. I would also like to thank the LinkedIn executives—Dan Yoo, Steve Sordello, Dan Shapero, Mike Gamson, Shannon Brayton, Ryan Roslansky, and Mohak Shroff—who provided invaluable guidance and support to me and my team; and Jeff Weiner, the best CEO I have known throughout my career. Last but not the least, I am indebted to my dear colleagues and team members who have been my daily source of inspiration and energy to come to work over the last seven years. Thank you all for making my journey at LinkedIn forever memorable. If you are interested in knowing about my next  play, stay tuned to my LinkedIn profile : ).