August 2020 Summaries

Idempotence is an important characteristic of many systems, including Fivetran, where if you execute an operation multiple times, the result will not change after the initial execution. In data integration, idempotence prevents duplicate records from being loaded and makes your data pipeline self-correcting. Without idempotence, duplication can lead to misleading insights, broken operations downstream, wasted storage space, and computational bandwidth.

Fivetran Women is an employee resource group that aims to connect, educate, empower, and advocate for female-identifying individuals and their allies within the company. The group holds monthly meetings to discuss experiences, share knowledge, and identify actionable steps to improve the workplace environment for women. They also host educational sessions featuring experts in various fields and engage with the community through mentorship programs and career talks at local schools. Fivetran Women has successfully encouraged leadership to focus on increasing the number of women in leadership positions throughout the organization.

Fivetran Women, an employee resource group launched by Fivetran in March 2019, aims to connect, educate, and empower female-identifying employees and their allies through social impact, professional development, intersectional support, and advocacy. The group holds monthly meetings to discuss topics such as salary negotiation and navigating male-dominated environments, fostering community and encouraging action to improve the workplace for women. Education sessions feature speakers who provide expertise on various subjects, and the group extends its efforts to the broader community by engaging with local schools and mentorship programs like Kode with Klossy. Fivetran Women has successfully influenced the company to increase female representation in leadership roles, demonstrating the collective strength of a supportive network in creating significant organizational change.

Fivetran, a company specializing in data integration, is currently seeking skilled Product Managers. The interview process at Fivetran involves several stages, including a recruiter screen, hiring manager interview, feature release presentation, and onsite interviews. Throughout the process, candidates are evaluated based on their communication skills, organization of ideas, alignment with company values, and ability to excel in the role. Interviewers focus on specific examples rather than generalities and encourage candidates to ask insightful questions. Successful candidates will be expected to bring curiosity, ownership, collaboration, kindness, and integrity to the team every day.

Drizly, the world's largest alcohol marketplace, uses analytics to enhance its customers' experience by connecting them with their preferred beverages. The company's data team leverages Fivetran, dbt, Snowflake, and Looker to analyze various factors such as coverage, availability, hours, capacity, selection, and price that contribute to the success of the Drizly network. By focusing on these metrics, Drizly has been able to improve its customer experience and retailer success within the marketplace.

GitHub data can provide valuable insights into an engineering team's productivity and performance. By using a connector like Fivetran, companies can gather information about issues, pull requests, and commits. Metrics such as the number of opened issues for bugs, code hygiene, and new features, work dedication to each category, and time taken to close pull requests can be analyzed. This data can help identify areas where improvements are needed and track the effectiveness of different strategies. Additionally, integrating GitHub data with other sources like Zendesk or customer support data can provide a more comprehensive understanding of product health.

Fivetran has introduced the Log Connector, which provides increased transparency and understanding of all data pipeline actions. The connector sends application events to track data pipeline actions, offering highlights such as managed replication actions, notifications for successful, failed or delayed syncs, and data usage metrics tied to Fivetran billing. Additionally, the company is participating in the dbt community with a new package that creates reporting tables for enhanced individual connector log tables, MAR history segmentation by table and destination, and greater visibility into post-load activities.

We developed the Fivetran Platform Connector to send application events for tracking data pipeline actions, providing managed replication action insights and sync notifications. The connector is free of charge as part of our commitment to product transparency. Our latest dbt package enhances individual connector log tables with sync status metrics and creates a history of measured monthly active rows and credit consumption.

Fivetran, a data infrastructure company, has shared its software engineering interview process and tips. The main principles of the interviews are to simulate real-life work situations and focus on technical communication and organization of ideas. The interview process consists of a recruiter screen, talking with the hiring manager, coding challenges, design interview, teamwork and culture fit evaluation, and a final interview with the hiring manager. Each interview is designed to assess different aspects of a candidate's skills and how they align with Fivetran's values and goals.

Fivetran introduces a new dbt package for LinkedIn Ads, designed to simplify ad analytics by aggregating commonly used metrics and creating base reporting tables. The package allows users to track clicks, conversions, and costs across campaigns and ad groups, as well as enrich creatives with associated campaigns and UTM parameters. It can be combined with other paid digital advertising campaign data for a comprehensive view of return on investment (ROI). Fivetran's native LinkedIn Ads integration simplifies the process by using OAuth to enter credentials and consistently delivers data in a pre-defined format, which is then transformed into overview tables containing shared ad metrics.

The article discusses five essential dashboards that businesses can leverage to answer critical business questions. These include tracking money flow in and out of the organization, understanding how people learn about the company, monitoring product usage, assessing customer support quality, and measuring work accomplishment. Each dashboard requires different metrics and sources of data, ranging from CRMs and ERPs to web analytics tools and project management software. The article emphasizes that these dashboards can help businesses make informed decisions and improve their overall performance.

Fivetran has introduced a new dbt package for Stripe, designed to help users make sense of Stripe's detailed API and improve their finance analytics. The package creates models that provide time-segmented overviews of balances and transactions, enrich subscriptions and invoices with details about associated objects, and tie balance transactions to relevant customer and payment details. These models can be used for immediate tracking or building more complex dashboards like MRR (monthly recurring revenue) for insights into month-over-month financial health. The Fivetran dbt package simplifies the integration process by delivering data in a consistent format, laying out object relations clearly, and running common joins across tables to determine holistic business impact.

Fivetran's journey from a modest beginning to becoming a key player in data pipelines involved significant development of their database connectors, marked by numerous trials and innovations. Initially, their connector efforts were basic, akin to simple data pipelines, lacking sophisticated data extraction methods and unable to capture deleted data or maintain historical data records. This approach, called "snapshot-style replication," faced limitations such as inefficiency and an inability to track changes or deletions comprehensively. To overcome these challenges, Fivetran transitioned to log-based replication, particularly utilizing Postgres's write-ahead log (WAL), which records all data manipulation operations. This method enables incremental updates, avoids the need for cumbersome workarounds like date columns, and allows for tracking deletions and full history of changes in tables, thus enhancing the performance and capability of the connectors significantly.