Where the revenue leaked, and how it was sealed.

The leak

Customer data was spread across the product backend, billing, the CRM and the email tool.

• Sales couldn't see which team accounts were paying or growing.
• Lifecycle emails couldn't target people by real product usage or billing state.
• Ad platforms were optimizing on incomplete conversion data.

What was built

1. A CRM built from scratch. People and companies sync from the backend, team accounts are modeled as their own object, and automated workflows handle deal creation, deal stages driven by recurring revenue, won, lost and churn changes, and primary-contact assignment.
2. A lifecycle email data layer. 73 contact properties are unified from the backend, billing, CRM and database, covering subscription state, trial and renewal dates, onboarding progress, feature usage, team role and ad attribution. The email tool only accepts flat data, so the team hierarchy is flattened in backend code and every member record carries its role plus team-level state.
3. Edge cases are handled automatically: admin or payer changes, member removal, team churn and members on different plans.
4. Revenue analytics. Billing data, website events and ad spend are brought together, and conversion events are sent back to the ad platforms from the server. Dashboards cover return on ad spend, cost per click, revenue per user, monthly recurring revenue, lifetime value, payback period, churn, revenue by plan, time to first paid invoice and the trial to first invoice to second invoice funnel.

The leak

Two websites, one for marketing and one for documentation, gave no view of how visitors moved toward buying.

• Marketing and sales tools didn't share data.
• Signing up on a phone took too much effort.

What was built

1. Event tracking across both websites, and a visitor lifecycle model that keeps the full history of stage changes.
2. A two-way sync between the CRM and product analytics. Every CRM contact links to their behavior, and the reverse.
3. Bot and human traffic separated, a daily growth summary delivered to Slack, and an executive marketing report.
4. A before and after analysis of a website platform migration.
5. A mobile signup test. The control was a 7-field form. The variant asked for an email only, then sent an automated email so the visitor could finish on desktop. Each person's variant was stored on their CRM record.

The leak

Marketing spend, product adoption and billing lived in separate systems. Nobody could tell which campaigns produced paying customers, or whether developers using the open-source tooling ever converted.

What was built

1. Event tracking on the marketing website and the developer tooling.
2. Data pipelines bringing database and billing data into product analytics.
3. Ad spend imported from two ad platforms, with conversion events sent back to them.
4. Campaign-level financial dashboards with full filtering: return on ad spend, cost per click, revenue per user, monthly recurring revenue, active subscribers, lifetime value and payback period, trials started, first and second invoices paid, time to first paid invoice, churn, revenue by plan, and funnel drop-off.

The leak

Payment data was separate from product usage. The team couldn't see where users dropped off between signup, onboarding and paying.

What was built

1. Real-time payment events captured from billing webhooks in the backend and sent to product analytics.
2. A historical backfill of payment and user data from the production database.
3. Dashboards for conversion funnels, onboarding drop-off, retention, stickiness, lifecycle and user paths.
4. A lifecycle messaging tool wired into the codebase.
5. An analytics consent flow built to meet GDPR.

The leak

Project work and accounting didn't talk to each other.

• Estimates were built by hand.
• Invoices in the accounting system weren't linked to project milestones.
• Kickoffs were delayed because the project team couldn't see when a client had paid.

What was built

1. Lead intake and customer records. Several web forms feed one intake flow. New leads are checked against existing records, created in the project tool and mirrored in the accounting system.
2. A quoting tool inside the project tool that pulls from the firm's product and material lists. Once a quote is approved, the official estimate and invoice are generated from it automatically.
3. A two-way finance sync. Estimates or invoices created directly by the finance team are matched and attached to the right project board.
4. Payment-triggered kickoff. When an invoice is marked paid, project deliverables are created, the design team is assigned and the project moves to Active.
5. Purchase orders created from project boards and synced to accounting, so vendor costs show up in project profitability.

The leak

Intake forms, meeting notes, project setup, contracts and billing all lived in separate tools.

• Leadership drafted agreements by hand.
• Invoices had to be chased manually.
• Clients were updated by hand.
• What was said on sales and strategy calls got lost.

What was built

1. Client intake. Form inquiries and calendar bookings create the client record with contact details and meeting times, and notify the assigned closer in Slack.
2. Meeting notes sync. AI-generated transcripts, summaries and recordings attach to the right client record after each call.
3. Contract generation. When a deal reaches the commitment stage, retainer contracts are generated with the client's scope, pricing and entity details, then sent for e-signature.
4. A two-way finance sync. Once the contract is signed and the deal is marked won, the customer is created in the accounting system, the first invoice is generated and the invoice PDF is attached back to the client record.
5. Recurring task scheduling. Weekly, monthly and quarterly deliverables for each signed retainer are created and assigned to the right department.
6. Client and lead messages. Automated emails and text messages for lead nurturing, payment reminders and milestone updates.

The leak

Every step from an inbound web lead to vehicle handover was handled manually across disconnected tools.

• Sales and technicians had no shared source of truth.
• Contracts were drafted by hand.
• Loaner and rental cars were tracked separately from the repairs they were tied to.
• Keeping up with more leads meant hiring more admin staff.

What was built

1. An operations hub that tracks each vehicle from arrival to key handover.
2. Lead intake and routing. Web leads and sales-tool conversations go straight into the CRM and are assigned to an available sales rep.
3. Automated contracts and payouts. When a deal reaches the closing stage, contracts and payout forms are generated and pre-filled with deal data.
4. Loaner and rental fleet tracking in real time, matched to customers whose own vehicles are in for long repairs.
5. Separate workspaces for salespeople and technicians, with tasks, vehicles and client details routed to the right person by pipeline status.