April 2021 Summaries

Phillip at Handshake, who manages business systems for the Employer (B2B side) sales team, shares his technique for building complex Salesforce objects with SQL to streamline analytics. The company was tracking Opportunities but wanted to track the full customer journey through their sales funnel, starting from the first interaction with the sales team or qualifying as an MQL. They built a new Customer Journey custom object to log important milestones in this lifecycle, using Census to construct the records nightly with a SQL query to their data warehouse. This approach allowed them to simplify the process and avoid complex Apex scripts, while still gaining valuable insights into prospect interactions and pipeline health indicators.

As companies grow, they face complex challenges in managing data, including scaling concerns related to both growing teams and increasing data volumes. To address these challenges, companies need to adopt a structured approach to data management, separating the people using data models from those running the platform, and implementing tools like GitOps and specialized storage engines to handle large datasets. This shift requires significant investment in new technologies and processes, as seen in the evolution of companies like Shopify, which has had to adapt its infrastructure to accommodate growing data volumes and teams. By normalizing on a Git-based process and embracing specialized tools, companies can overcome these challenges and continue to innovate and grow.

The Operational Analytics Conference brought together data pioneers and leading practitioners to discuss lessons learned, the evolution of data across teams and processes, company culture, and career advice. The panel discussion covered topics such as what being a data-driven company means, whether companies should build tools for their customers versus products they need, and where the data ecosystem is at right now. Industry experts shared their perspectives on the current state of the market, with some arguing that we are in an explosion phase, while others see signs of consolidation on the horizon. The conversation highlighted the importance of education and embracing chaos as key strategies for success in this rapidly evolving landscape.

The Operational Analytics Conference panel discussion featured an expert panel of product managers, ex-PMs, and product people who shared their perspectives on building data and machine learning products, up-leveling companies to be more data-centric, and when not to use data in making product decisions. Alignment and coordination between teams is crucial for successful project management, as a team can achieve high performance but hollow success if the OKRs or metrics don't meet the needs of its constituency. Co-defining specific metrics and OKRs with key customers can help identify what makes them productive, leading to thinking about the data stack as an end-to-end product. Product managers play an effective role in this by considering a use case from start to finish and ensuring that their goals are aligned with customer needs. The discussion provided valuable insights into the profession of product management, which is always changing and evolving, making it a must-listen for current and aspiring PMs.

The Operational Analytics Conference featured a panel discussion on creating and maintaining a data-driven culture inside a company and team, which is one of the two core pillars of operational analytics. The panelists, including heads from various companies such as Billie, Upkeep, Figma, Instacart, and Yelp, discussed how to become data-driven, what happens when they finally get there, and how to maintain that culture in a rapidly changing environment where new people, technologies, and approaches flow in and out. They shared their experiences on common problems such as analysis paralysis, the weaponization of data, and credibility exercises, which can hinder the effectiveness of data-driven decision making. The discussion highlighted the importance of having a license for leaders to make decisions based on data while acknowledging potential flaws, and the need to define key business metrics consistently across executives to avoid spiraling conversations. The panel was described as fascinating and insightful, with smart people sharing their expertise on building successful data-driven cultures.

The hiring process for data roles is extremely challenging due to a high demand for skilled professionals but limited supply of available talent. Experts in the field agree that soft skills such as communication and problem-solving are more valuable than technical skills, and that teaching these skills can be easier than acquiring them through formal education. Many panelists also emphasized the importance of understanding how data maps to business processes and value, rather than just focusing on technical skills like SQL. By removing academic prerequisites from job requirements and valuing diverse skill sets, companies can attract a wider range of talent and build more effective data teams.

Brooks Taylor's recent mistake involved building revenue tables in his company's self-service and invoice side of the business, where he tried to update old historical records with a simple timestamp change to facilitate analysis. However, this seemingly harmless action triggered a series of underlying calculations that relied on the same data, ultimately resulting in a loss of $5 million. The incident highlighted the interconnected nature of modern data ecosystems, emphasizing the importance of considering potential downstream effects when making changes, even if they appear minor. This experience serves as a cautionary tale for data professionals, illustrating the need for careful consideration and attention to detail in their work.

The modern data team is undergoing significant changes in response to new tools and technologies. The shift towards cloud-based architecture and automation has made it easier for teams to work with data, reducing the workload of both initial setup and daily responsibilities. Automation is causing a philosophical shift toward treating "data as code," allowing for more flexibility and collaboration. Remote work is changing day-to-day routines, providing individual team members with more control over their time, while also broadening the talent pool. Traditional methods are on their way out, replaced by new workflows and increased innovation incubated from a larger talent pool. Data team roles are changing, with old roles becoming obsolete and new ones gaining prominence. Specialized roles are emerging, allowing for more focus on specific areas of expertise without worrying about the intricacies of data infrastructure. The purpose of a data team will continue to shift, requiring an easily adaptable data stack to navigate these changes. A new formula for data stacks is starting to shake out, with tools like Fivetran, Snowflake, dbt, Census, and BI layers forming a resilient and adaptable infrastructure.

At Census, we aim to make analytics accessible across companies by integrating data from data warehouses into favorite applications, enabling teams to build better operations. We've launched our Front Connector, which allows support agents and customer success managers to access metrics and usage data directly, reducing time spent on information retrieval and improving ticket handling efficiency. This integration minimizes API usage while respecting API rate limits, and teams can start using it by checking out our documentation. With Census + Front, companies can operationalize their existing data and contact us for a demo.

Qasim Ashraf, an e-commerce expert with diverse experience, emphasizes that data teams need to adopt a people-centric approach to make companies more data-driven. He believes that data is inherently a support role, amplifying other departments' business needs rather than directly impacting the bottom line. To prove their value, lean e-commerce data teams should focus on tactical solutions that address specific pain points and prioritize balancing strategy with execution. Qasim also highlights the importance of operational analytics in making data actionable, reducing friction points for stakeholders, and increasing the organization's overall data-driven efficiency.