February 2021 Summaries

Jamie Quint, a seasoned analytics expert, has identified the best tools for an ideal analytics stack for founders. He uses Amplitude for easy data analysis, Mode for complex analysis with visualization capabilities, Segment as a meta analytics layer to federate data across tools, Snowflake as a central data warehouse due to its scalability and ease of use, dbt to transform and prepare data, Fivetran to send data into the warehouse, and Census to put data to use. Jamie's approach is based on identifying specific needs and functions, such as combining different types of data, sending data to a warehouse, transforming data, and getting it out for use in other platforms. He emphasizes that understanding how he landed on this stack can help others adapt as technology advances and changes.

The company Census has raised a $16 million Series A funding led by Sequoia Capital, with additional participation from Andreessen Horowitz and other notable operators. This round brings the total amount raised to over $20 million. To support startups, Census is launching a program that allows companies with fewer than 40 employees and less than $10MM in funding to use Census for a flat rate of $100 per month. The company has seen phenomenal growth since its launch in 2018, syncing analytics for over half a billion users every day, and has partnered with several notable organizations such as Canva, Figma, and Notion. Census is focused on enabling product-led companies to drive better marketing, sales, and customer success by leveraging data teams to drive operations. The company believes that the data warehouse should act as a new kind of Customer Data Platform, enabling hyper-scale businesses to deliver personalized experiences to customers. This vision is part of a larger movement in the data ecosystem, where software practices are washing over analytics, and Census is helping data teams build solutions like engineers. The company plans to expand its capabilities in areas such as code-based orchestration, deeper data validation, and visual query experience.

Fivetran is an essential tool for data integration in the cloud, allowing users to easily load data into their warehouses and unlock powerful use cases such as marketing attribution and customer journey analysis. The company's 150+ integrations catalog enables seamless ETL data transfer, while Census serves as a go-to-market tool for syncing insights back to various tools. A partnership between Census and Fivetran aims to expand operational analytics capabilities by integrating their products and making it easier to orchestrate across the stack.

Databricks has been a highly requested data source since its public launch, with its Delta Lake product widely available. The company is known for championing Apache Spark and making big data processing faster and easier. Census shares similar goals, aiming to turn warehouses into operational hubs that drive AI and ML outcomes. The integration of Databricks Delta Lake with Census enables direct connection of ML and analytics to business operations, expanding use cases such as lead scoring, customer journey tracking, and revenue optimization.

Operational analytics is a type of analytics that informs day-to-day decisions with the goal of improving efficiency and effectiveness in an organization's operations. It drives action by automatically delivering real-time data to where it'll be most useful, no matter the location within the organization. Operational analytics powers important daily decisions, both big and small, by providing insights that are not just about understanding business operations but also about driving business operations. At its core, operational analytics is about putting an organization's data to work so everyone can make smart decisions about their business. It introduces a set of fundamentals for leveraging data across the organization, making it a key tool for companies looking to achieve data-driven decision-making at scale. Operational analytics differs from traditional analytics in that it uses data to drive business operations rather than just providing an understanding of what's going on in the business to inform strategic decisions over time. It answers questions like "Which support ticket should I tackle first?" and is used by teams such as customer success, sales, and marketing to make specific activities more efficient. The modern data stack that supports operational analytics consists of four sections: data integration, data storage, data modeling, and data activation, which work together to provide a hub-and-spoke model for data flow. This data stack can be set up with tools such as Fivetran, Snowflake, dbt, and Census, making it accessible to businesses of all sizes. Operational analytics is necessary for companies with more than a handful of customers, as it allows them to overcome limitations in individual tools and enables their data teams to take a more proactive role in how the business uses data. With operational analytics, companies can automate daily brain-draining tasks, enable marketing ops people to create hyper-specific drip campaigns, and empower CS reps to prioritize support tickets without relying on spreadsheets or traditional analytics. The "X factor" making all this possible is the data team, which can diagnose workflow problems and create solutions using operational analytics. To start implementing operational analytics, businesses should build a data stack blueprint starting with an ETL tool, a data warehouse, dbt, and Census, and begin small with one use case before rolling out to other teams.

The key to getting a seat at the table in nominally "data-driven" companies is not technical know-how, but demonstrating and increasing the impact that your data team brings to your company. To achieve this, you need to establish accountability and trust with stakeholders, match their cadence, score C-level buy-in, and make your data team's role as an 'insights facilitator'. By narrowing down internal "customers", embracing proactive outreach, moving at the same speed as stakeholders, breaking bottlenecks, and demonstrating value, you can secure buy-in from C-level executives to product teams. Ultimately, it's about making data accessible, intuitive, and seamless for all stakeholders, and finding an executive champion to support your data team's growth.

Census has achieved SOC 2 Type 1 compliance for security, availability, and confidentiality of customer data through a set of internal controls, systems, policies, and procedures that meet industry best practices. An independent auditor verified the design of these controls and systems, ensuring high availability processing and protection of sensitive information. Existing customers can access their audit report upon request, while new trials will receive it under non-disclosure agreement. Census plans to undergo an additional SOC 2 Type 2 audit in the future to maintain compliance and continue demonstrating its commitment to protecting customer data.

The traditional Extract, Transform, Load (ETL) and Extract, Load, Transform (ELT) methods are being replaced by a new approach called dbt, which offers more flexibility and scalability for data transformation and analytics. The ELT method is better suited for large amounts of simple data transformations, while the ETL method is more suitable for real-time data with complex calculations. Modern tools like Fivetran, Airflow, Stitch, and cloud warehouses like BigQuery, Snowflake, and Redshift make it easier to use these pipelines. However, dbt allows users to create a flexible command-line data pipeline tool that can be quickly programmed, tested, and modified without huge waiting times. With dbt, users can aggregate, normalize, and sort the data as needed, without constantly updating their pipeline and resending data. The author encourages readers to try out dbt for themselves and explore its capabilities.