Reflections on Embedding in a Product Team as a Data Scientist
Navigating Collaboration and Specialisation
As my first six months at Intercom have come to a close, I wanted to take the time to reflect on my experience and learnings thus far — particularly on my involvement being RAD’s first embedded data scientist in a product team. But first, a little context on who we are in RAD:
The RAD team is made up of researchers and data scientists with the goal to build, enable, and share a deep understanding of our customers, their needs, and their product behaviours. Typically, a RAD data scientist works in a hybrid way, spending about half their time working directly with product teams and the other half on RAD or company priorities.
Shortly before my start date, a new product team, Team Activation, was created with the mission to create a beautifully simple activation experience for new self-serve customers to set up Intercom. Being a new team, they had a high volume of data questions around setup and activation. They also wanted to ship and learn via AB tests. This meant that they required full time access to a data scientist. My role was the first time a data scientist within RAD was embedded within a singular team.
Being embedded as a data scientist within a product team has definitely been a net positive experience. In this blog post, I will reflect on the strengths and challenges of embedding as a data scientist and share some learnings and suggestions for other teams considering a similar approach.
The Power of Collaboration and Knowledge Sharing
Working closely with Team Activation has allowed me to build strong relationships with the product manager, engineers, and designers on my team. The open and frequent communication channels have created a culture of trust and fostered a deep sense of teamwork. Through daily stand-ups, 1:1 meetings, brainstorming exercises, State of Product meetings, and retrospectives, we maintain a constant loop of feedback, ensuring everyone is aligned and supported in achieving our goals. I am able to provide data insights early and often during these rituals. Centralised data scientists working across multiple teams may not have the time to attend all the rituals for each of the teams they are supporting. This level of collaboration and knowledge sharing simply would not be possible if I was not embedded in the team.
Gaining Deeper Domain Knowledge
The immersion within the activation domain has been immensely beneficial for me. By closely engaging with activation-related projects, I have developed a comprehensive understanding of the domain and the underlying data that drives it. This deep knowledge has enabled me to provide tailored analysis and insights based on the team’s past work, creating a seamless collaboration experience. Working in close proximity with the product manager has facilitated a better understanding of their needs, resulting in more impactful analysis and decision-making.
Fast Turnaround Time and Focus
As an embedded data scientist, I can dedicate the majority of my time to activation-related requests. This focused approach allows me to respond quickly to new queries, adapt priorities on the fly, and deliver timely insights without negatively impacting other teams. The ability to be agile and responsive enhances the overall efficiency of the team, ensuring that we can effectively tackle challenges and seize opportunities as they arise.
Challenges and Areas for Improvement
While embedding as a data scientist offers numerous benefits, there are a few challenges that should be acknowledged and addressed to optimise this approach:
- Difficulty in Drawing the Line: The deep embedding within a team can sometimes blur the boundaries of responsibilities. As a dedicated data scientist, there may be pressure to answer every question, even if it is a low-impact query. It is important to establish clear guidelines and priorities to ensure that the highest-value work receives proper attention and to avoid potential burnout or loss of focus.
- Specialist vs. Generalist Dilemma: Immersion within a specific team may limit exposure to insights and knowledge gained from collaborating with multiple teams. While specialisation has its advantages, it is crucial to strike a balance and seek opportunities to engage with other teams periodically. Diversity in projects ensures continuous growth and prevents siloed knowledge.
Learnings and Suggestions for Other Teams
Based on my experience as an embedded data scientist, here are a few suggestions for other teams considering this approach:
- Improve Data Science Onboarding through Embedding: Embedding data scientists within product teams can be an effective way to facilitate seamless onboarding. Providing new hires with a focused goal and a specific domain to explore helps them acclimate to the company culture and domain more smoothly, leading to quicker contributions and impact.
- Transitioning Embedded Data Scientists: To avoid siloed work and knowledge, consider having data scientists transition out of the embedded role after 5–7 months or after the appropriate onboarding time, into the centralised data science approach. The teammate can bring their expertise, best practices, and understanding of patterns and insights from their previous team across multiple teams, leading to more effective problem solving and a broader impact on the company.
It has been an amazing adventure here at Intercom. I have learned so much in a short period of time, growing exponentially in my technical skills and understanding of a product-led organisation. This is only the beginning — I can’t wait to see what the future holds for me and RAD team at Intercom.
We would love to hear more about the lessons and experiences of other data science teams — in particular the experiences of other data scientists who embed with product teams directly.
In Intercom, the Research, Analytics & Data Science (a.k.a. RAD) function exists to help drive effective, evidence-based decision making using Research and Data Science. We’re always hiring great folks for the team. If these learnings sound interesting to you and you want to help shape the future of a team like RAD at a fast-growing company that’s on a mission to make internet business personal, we’d love to hear from you.
