How Better Legal hit $1M ARR by switching from custom code to no-code

Executive overview

Better Legal, a business-formation SaaS, spent five years building a custom-code platform only to migrate the entire product to the no-code tool Bubble — and crossed $1M in ARR shortly after. The move resolved chronic front-end instability and DevOps drag that was consuming the engineering team without adding competitive value. Within about a year, the team replicated five years of custom development, then spent the following year shipping entirely new AI-powered features that had previously been impossible to prioritise. The key insight is that no-code platforms are simply a new generation of WYSIWYG tooling, and fighting that trend wastes resources that should go toward product and customers.

Choosing the right infrastructure layer is a business decision, not an engineering purity test — if it is not your secret sauce, do not hand-code it.

The problem Better Legal solves

• Business formation involves three essentials: state filing, EIN, and legal documents for a bank account.
• Competitors (LegalZoom, ZenBusiness) bury customers in upsell menus and speed-tier tricks.
• Better Legal bundles everything at a single flat price with the same processing speed the state provides.
• The idea came from an attorney friend charging $1,500 for what was essentially manual data entry into a 50-page template.
• Automating that data entry cut a four-hour task to ten minutes; the co-founders validated demand through the attorney's own law firm before spending anything on ads.
• First paid ad test: $1,200 spent, two sales, $600 revenue — enough signal to continue.

The technology journey: four platform generations

• Gen 1 (no-code): Typeform + Zapier + Webmerge + Asana — cobbled together but functional.
• Gen 2: replaced Asana with Salesforce to allow data editing before document generation.
• Gen 3 (custom code): needed a customer portal; Salesforce/Heroku pricing became unpredictable above 10,000 rows, so the team rebuilt on custom code and spent five years there.
• Gen 4 (Bubble): chronic front-end developer conflicts over JavaScript framework churn and refactoring for marginal speed gains crowded out real features; moved to Bubble for a 10x speed advantage at the same developer cost.
• Migration ran old and new versions in parallel with nightly database syncs, letting the team revert when edge cases surfaced; full cutover completed in under a year.

Why no-code beat custom code for this business

• DevOps overhead (server crashes, broken packages, AWS tuning) consumed engineers without differentiating the product.
• Bubble is hosted on AWS anyway; using it just outsources infrastructure optimisation to a team whose entire job is that problem.
• No-code is not truly "no code" — bespoke logic (e.g., Stripe customer creation) is handled via small custom scripts called through APIs; the platform handles everything else.
• Investor concern about "not owning your code" is a false framing: every company is locked into something (Stripe, cloud providers, open-source libraries), so the only question is whether the lock-in is with a high-quality, responsible vendor.
• Salesforce — the original enterprise no-code tool — is used by LegalZoom, ZenBusiness, Walmart, and Coca-Cola; no institutional investor questions that choice.

Revenue and business model evolution

• Better Legal ran at $2–2.5M total revenue for several years, but most of it was one-time transactional revenue.
• The team deliberately shifted toward SaaS/recurring revenue; by December 2023 they hit $1M ARR.
• A new AI contract-analysis tool on usage and subscription pricing is expected to accelerate ARR growth further.
• The business is now roughly 50/50 between recurring and one-time revenue, compared to a small recurring fraction two years earlier.

Advice for founders building on no-code

• Do not worry about scaling until scaling is actually a problem; solving for 100,000 concurrent users before you have a single paying customer is wasted effort.
• If scalability becomes an issue at 10 or 100 customers, that is a good problem — pause signups, charge more, and fix it then.
• No-code platforms run on serious infrastructure (AWS, etc.) and their support teams can advise on optimisation; lean on them instead of reinventing the wheel.
• Build on no-code until you need something bespoke, then call a custom API — do not rebuild the entire stack.
• Unexpected large bills can happen on any usage-based system; the risk is not unique to no-code platforms.