Data-driven products for diverse audiences
Last night, we brought together experts from Women In Product and Mixpanel for a conversation about building data-driven products that serve diverse audiences.
The panel featured Naba Banjeree, Head of Product for Walmart’s SamsClub.com, and Yardley Ip, General Manager at Zillow Group and co-founder of Women In Product – an organization that brings together a strong community of women builders and leaders in product. It also included two of Mixpanel’s own experts: Design Manager Stephanie Hornung and Product Manager Veronica Pinchin.
The discussion centered on when and how data should be used to build better products. Based on a wide range of experiences, the group was able to cover a lot of ground: from using data to get buy-in on key decisions and develop thoughtful product roadmaps to understanding diverse user groups and tips for leading resilient product teams that take risks in service of innovation.
For those who couldn’t make it or want to a recap, we’ve shared the four key points we took away from the conversation. You can also watch the full panel here.
Tell the same story that the data does – even if it’s not the one everyone wanted to hear.
Naba told a story about a product manager who wanted to focus their efforts on post-purchase opportunities–but didn’t have the data to back up the plan. It sparked a conversation about the dangers of retrofitting data to the hypotheses that PMs want to be right. The panelists agreed that PMs should always check intuition against data, and accept when the data says “no” to an idea and proves their hypothesis wrong.
Make the product roadmap a customer problem roadmap.
According to the panel, customer pain points should inform what PMs prioritize on their roadmaps. Veronica talked about the 70/20/10 rule: have 70% of the roadmap focuses on customer problems, 20% on high risk/high reward projects, and 10% on crazy ideas that just might work.
That’s not to say that PMs shouldn’t take risks based on their intuition, but those changes should be made quickly and incrementally. Yardley chimed in on the 70/20/10 rule to advocate for setting up milestones along the way where the data can inform changes and tweaks. “I’m a big believer in beta programs,” Yardley said. That way, teams can “fail fast,” identify what isn’t working, and avoid developing design and engineering debt.
Remember: There’s a world outside of every product.
To understand those customer problems, PMs have to consider quantitative and qualitative data simultaneously. Looking at only surveys or only anonymized behavioral data can skew perceptions, particularly when a company’s customer base is large and diverse. Instead, the panel suggested complementing large-scale behavioral data with investments in user research, developing behavior models of high-value users, and shadowing sales and support to get a deeper sense of customer pain points, in order to get more context for larger trends.
To find out what else the panelists had to say, watch the video.
Many thanks to Hayley Leibson, Founder of Lady in Tech, for her photo.
If you use data to build better products, download our Product Benchmarks Report here. We captured data from 1.3 billion unique users performing 50 billion events to determine standard and elite benchmarks in four industries for product innovation: financial services, media & entertainment, SaaS, and eCommerce.