August 2022 Summaries

Stripe, a prominent fintech company, is widely used for its software and APIs that facilitate online payments and business management, with Stripe Payments and Stripe Billing being key products. While Stripe Billing is designed to manage subscriptions and calculate customer charges, replicating Stripe's own pricing model using it reveals several limitations, particularly for B2B and product-led companies. These limitations include difficulties in charging percentage fees, a non-event-based architecture requiring redundant usage records, and the need for managing multiple IDs for simple pricing setups. Additionally, rate limits pose scalability issues for high-volume transaction processors, and the requirement for uniform billing periods across products complicates usage-based pricing models with diverse billing needs. Despite these challenges, the experiment underscores the potential of alternative billing solutions like Lago, which are tailored for B2B contexts and usage-based pricing complexities.

Open source solutions are reshaping the traditional 'Build vs Buy' decision in the billing space, much like Airbyte has done for the ETL industry, by offering a compelling alternative to traditional closed-source billing solutions like Chargebee and Stripe Billing. Early-stage companies often opt for purchasing a billing system to prioritize growth, whereas more established companies, like those at Series C or D, face the challenge of upgrading their complex home-grown systems while managing engineering resources effectively. These larger companies often find that their intricate pricing models do not fit standard software solutions, leading to the need for substantial engineering teams focused on maintaining and evolving their billing systems, which can be a cumbersome and resource-intensive task. The complexity of billing systems, involving multi-currency, multi-tax, and multi-tenant considerations, often surprises companies and results in long-term projects, as illustrated by Kevin, former VP Engineering at Scaleway, who described his extensive experience in developing a billing system over several years. Open source frameworks, like Lago, provide a flexible and scalable alternative, allowing companies to build systems that can adapt to evolving pricing strategies without the increasing costs associated with traditional platforms, which typically take a share of revenue and may not accommodate the complexity of product-led growth models.

Heroku's flexible pricing model, which allows a single customer to have multiple plans with different prices and add-ons, serves as an inspiration for many API and cloud companies. This model is particularly useful for platforms that enable users to create multiple projects, workspaces, or repositories, as demonstrated by companies like Webflow, Shopify, Segment, Algolia, and Column. It simplifies billing for design and marketing agencies by providing a breakdown of costs per project, allowing them to invoice their clients accurately. The complexity of managing multiple subscriptions, especially with usage-based features, necessitates a robust billing system that generates unique subscription identifiers and incorporates display names for clarity. Consolidated invoicing is crucial to avoid overwhelming customers with multiple invoices, and aligning billing periods, as Heroku does by starting billing on the 1st of each month, can streamline this process. However, challenges remain in managing different billing intervals, calendar versus anniversary dates, and handling upgrades, downgrades, coupons, and prepaid credits, making it essential to have a comprehensive understanding of the billing system, as demonstrated by Lago's approach.

Lago's engineering stack is built with a focus on robustness and speed, primarily using Ruby on Rails due to the team's extensive experience and the language's wide community support. The company hosts its services on Amazon Web Services (AWS) and utilizes Kubernetes for container orchestration with an abstraction layer provided by Porter, a recommended choice by an advisor. For error alerting, Sentry is employed for its reliability, while New Relic is preferred for cloud monitoring due to its user-friendly onboarding and cost-effective entry-level plan. Internal VPN needs are met by Tailscale, which simplifies team member onboarding, and the back office is managed using Active Admin within the Ruby ecosystem. GitHub is used for code repositories, while documentation is handled by Docusaurus to maintain a consistent user experience and brand identity. Metabase is chosen for product analytics, offering flexibility for both technical and non-technical users, and Algolia is integrated into the documentation for search functionality. On the frontend, React and TailwindCSS are used, with PostgreSQL and Clickhouse for databases, Redis for in-memory caching, and Redpanda for streaming tasks.