Last week, ElevateData hosted its first virtual event of 2021. Many of our previous ElevateData conversations have been deep dives into the role and impact that data has within organizations and how the Chief Data Officer can optimize those results. So, in a small departure from our typical round table discussion, we were lucky to have an interactive fireside chat with Manish Chandra (Founder & CEO, Poshmark) and Barkha Saxena (CDO, Poshmark & Co-Founder, ElevateData).
Data leaders joined us from Disney, Airtable, Bitly, AirBNB, Amazon, The Gap, Etsy, Frontdoor, Ethos, JLL, Monte Carlo, and SVB to get a CEO’s perspective on driving advocacy for the data organization. The result was a wonderfully rich and engaging conversation with many insights shared. We’ve summarized several of the key takeaways below, but there’s nothing like joining the conversation in person! Let us know if you’d like to join this amazing group of data leaders at the next ElevateData event!
Data is Foundational
The mission of Poshmark is to put people at the heart of commerce, empowering everyone to thrive. The team at Poshmark makes this happen by bringing the physical joy of shopping to the world of online marketplaces and commerce. We’d say Manish, Barkha, and the team has been executing well towards that mission, attracting more than 70 million users across the US, Canada, and Australia!
As with any success story, there are no silver bullets or panaceas to that success. Instead, you have to peel back the onion and understand all of the layers that enabled Poshmark to achieve success. Manish was very generous in sharing one of the elements of Poshmark’s success – data.
“So much of our business is unseen. When you have millions of people using the Poshmark app and millions of transactions on top of that, it’s impossible to look at what’s going on in the business qualitatively, so data has been foundational for us.”
Manish Chandra
Not only was data foundational to Poshmark, but it was also foundational to Manish. He started his career as a database engineer. So our first key takeaway, find a data-driven CEO!
Creating a Data Culture
So, how do you build out a data-driven culture? At Poshmark, a thoughtful vision for the data organization was created in partnership between Manish and Barkha, and the execution towards that vision was (and remains) a key priority for the business.
There are three key pillars to the success of Poshmark’s data-driven culture.
- Leadership & Organizational Buy-in
- Data Infrastructure and Technologies
- Team
We believe that this blueprint is applicable to many organizations and to data leaders and CEOs alike! We’ll dive deeper into each of these three areas below.
Leadership & Organizational Buy-in – It Starts at the Top, but it can’t end there
Manish is a self-proclaimed data addict, “I would say out of my nine-hour day, if I’m not in a meeting (and many of my meetings are focused on data), I spend at least three to four hours in the data. My morning starts with looking at various data dashboards and often the night ends with that.”
This laser focus of the CEO on utilizing data to measure the business and to ultimately leverage that information to drive decisions has permeated through the organization. One great example of how this has manifested itself at Poshmark is the “Core KPIs” meeting that Barkha organizes. This monthly meeting brings a cross-functional group of leaders together to evaluate the Key Performance Indicators (KPIs) in the business across all business functions. The data is a single source of truth and presented as facts for participants to synthesize and discuss. In Manish’s own words, “instead of data giving an opinion, the data is presented as fact, and everyone chimes in with their opinions.”
Manish has only missed this meeting once in the past five years.
Buy-in from leadership is critical, but in order to truly create a very data-driven organization, you have to enable data across every level of the organization. Leveraging tools like Looker and homegrown dashboarding, real-time insights and A/B testing tools have really helped Poshmark democratize its data. Barkha proudly can say, “despite being a large company, every single person at Poshmark uses data to make decisions.”
Data Infrastructure and Technologies – The Plumbing for the Single Source of Truth
In the early days of Poshmark, the Company had a very basic data stack – leveraging Google Analytics and RJ Metrics. These tools served their purpose but had limitations so, in partnership with Gautam Golwala (Poshmark’s co-founder and CTO), Barkha created an ambitious vision for turning data into an operating tool for all business functions at Poshmark and set out to build a comprehensive data infrastructure and technology stack to deliver on that vision powered by high-quality big data.
Today, that stack at Poshmark includes some great homegrown and some off-the-shelf solutions built on top of a massive data infrastructure that supports a wide range of use cases, including:
- Analytics data warehouse
- Real-time KPI monitoring
- Maximizing ROI from operational business initiatives
- Seller tools
- Machine Learning (ML) models
Given how foundational data is to Poshmark, it was critical to be incredibly disciplined with data quality and data consistency. Barkha and team have built processes to minimize data quality issues and surface any potential issues at the beginning of data pipelines.
With confidence in clean data pipelines and data that represents a single source of truth, data users from the C-suite to the marketing team to the product team can leverage Business Intelligence (BI) tools to inform their day-to-day decision-making. Poshmark leverages various third-party tools as well as various homegrown data tools and data platforms for dashboarding to A/B testing to ML modeling built on top of open source technology.
Team – Building and Sharing the Gospel of Data
As CDO, Barkha reports to Manish. Initially, she reported to Poshmark’s Chief Operating Officer (another very data-focused executive) further demonstrating the foundational and critical role of data at Poshmark.
However, building the team was a journey of many steps. In 2014, Barkha’s data team consisted of one – herself. Over time she built advocacy and demonstrated the ROI of data within the Company. Today, Barkha’s team is larger than the total number of employees who worked at Poshmark back in 2014 when she joined the Company.
Barkha’s team is organized into 5 operationally focused vertical data science teams and one horizontal Machine Learning team, each led by a strong leader. The vertical teams are each partnered with a business function within Poshmark (E.G. Marketing, Product, etc). These integrations allow the vertical data science teams to better understand specific business objectives for each org. They then bring their expertise in data science, data management, and data tools initiatives to help drive success. The horizontal ML team partners with ML Engineering team to focus on longer-term ML priorities which span across all business functions. Barkha’s team closely partners with the Data Engineering team in the CTO org to continue to scale data at Poshmark.
Barkha credits this highly collaborative environment to Poshmark’s culture. It’s uncommon to find this collaboration at large organizations and even rarer to find a culture that rallies around data in the same way. It was clear that Manish and Barkha are incredibly proud of the data organization they’ve built, and so they should be!
Cost and ROI of Data
Building out what we just described above is not cheap. As many know, the rising costs of data tools, infrastructure, and people have made it increasingly difficult for data leaders to get the appropriate resource allocation. Despite Manish’s affinity to data, he still has a fiduciary responsibility to the business like any other CEO. As a result, there were many uphill battles that Barkha had to fight. Here’s how she did it.
Start Small & Demonstrate ROI
Manish credits Barkha with being able to create a remarkable data infrastructure stack & organization by consistently demonstrating high ROI. In the early days, this started with building out dedicated teams to bring data to identified problems with clear and measurable parameters of success.
This is an important point to unpack. Rather than collecting all of the data and trying to surface ideas, problems, and/or strategies, Poshmark instead focused on first identifying the highest priority business problems and then worked to bring the right data and right data solution to solve them.
Barkha will be the first to acknowledge there are merits to collecting all of the data, but it was this sort of creative thinking (and budgeting) that enabled her to build advocacy with Manish and the rest of the C-suite at Poshmark.
Once this model and ROI were proven, it enabled Barkha to build out a bigger and bigger team. Despite the growing costs, Manish believes that “data is very very leverageable so I don’t think it’s expensive from an ROI perspective.”
———————
Special thanks to Manish for joining the ElevateData community to share your perspective as CEO and your thoughts on how to build a data-driven culture. Thank you to all of the ElevateData members for joining us and passionately exchanging your ideas. ElevateData would not be what it is without you!
Your ElevateData Founders,
elevate.data.blog 