Project Overview
| Client Industry | Marketing Technology / Digital Agency Services |
| Business Type | Digital marketing agency transitioning internal reporting tools into a licensable SaaS product |
| Project Duration | 20 weeks |
| AI Service Provided | SaaS Web Development |
| Technologies Used | React, Next.js, TypeScript, Node.js, PostgreSQL, Redis, AWS (ECS, RDS, ElastiCache), Stripe Billing, Meta/TikTok/Google Ads APIs |
The Client Challenge
The client, a mid-size digital marketing agency, had built an internal social media analytics dashboard over several years to consolidate ad performance data across Meta, TikTok, and Google Ads for their own client reporting. Other smaller agencies had seen it during pitches and repeatedly asked to license it directly.
The business opportunity and technical gap:
- The internal tool was built as a single-tenant application serving only the agency’s own client reports it had no concept of separate paying customers, billing, or self-service account management.
- Data connections to each ad platform’s API were hardcoded for the agency’s specific client accounts, with no generalized way for a new SaaS customer to connect their own ad accounts.
- The agency’s engineering team had deep marketing domain expertise but no experience building or operating a true multi-tenant SaaS product, subscription billing, or the security practices required to handle other companies’ advertising account credentials.
- Performance was a real concern the internal tool was slow with the agency’s own client volume, and a real SaaS product would need to handle far more concurrent tenants and data volume without degrading.
- The agency needed to move quickly to capture the market interest they’d already generated through word of mouth, but couldn’t take on a large internal engineering hire without validating the SaaS opportunity first.
Our Solution
Air Brite Labs rebuilt the agency’s internal analytics tool into a true multi-tenant SaaS platform preserving the reporting logic and ad platform integrations that already worked well, while adding the multi-tenancy, self-service onboarding, billing, and security practices required to sell it as a standalone product.
Key capabilities delivered:
- Self-service ad account connection — new SaaS customers can connect their own Meta, TikTok, and Google Ads accounts directly through OAuth, generalizing what had been hardcoded, agency-specific integrations.
- Multi-tenant architecture — each subscribing agency operates in an isolated account with its own clients, campaigns, and reporting data, built to serve many customers securely on shared infrastructure.
- Unified cross-platform reporting — the core value proposition that had generated organic interest consolidated ad performance reporting across three major platforms was preserved and refined rather than rebuilt from scratch.
- White-label client reporting — subscribing agencies can generate branded reports for their own clients, a feature added specifically because it addressed a gap competitors in the space didn’t offer well.
- Subscription billing — tiered pricing based on number of connected ad accounts and managed clients, with self-service plan management replacing what would otherwise require manual sales-assisted billing.
- Scalable data refresh pipeline — ad performance data refreshes on a reliable schedule across potentially hundreds of tenant accounts without the performance degradation that affected the original internal tool.
Technical Approach
Multi-Tenancy Architecture: Implemented tenant isolation in PostgreSQL using a shared-database, row-level security model, allowing efficient infrastructure use while maintaining strict data separation between subscribing agencies and their respective client data.
Ad Platform Integrations: Rebuilt the Meta, TikTok, and Google Ads API integrations to support OAuth-based, per-tenant account connection rather than the hardcoded credentials the internal tool used a substantial rework given each platform’s different authentication and rate-limiting requirements.
Data Pipeline: Built a scheduled data refresh pipeline using Node.js background workers that pull performance data from each connected ad platform per tenant, with intelligent rate-limit management to stay within each platform’s API constraints even as tenant count grows.
Performance: Redis (via AWS ElastiCache) caches frequently accessed dashboard data and report queries, addressing the performance issues that had affected the original single-tenant tool and ensuring dashboards stay responsive as tenant and data volume scales.
Frontend: Built with Next.js and TypeScript, reusing and refining much of the original tool’s reporting visualization logic the agency’s domain expertise in what made reports genuinely useful for marketers was preserved rather than discarded in the rebuild.
Billing: Stripe Billing integrated for subscription management, supporting tiered plans and self-service upgrades, replacing what would have otherwise required manual invoicing incompatible with a scalable SaaS business model.
Security: Given the platform handles other companies’ advertising account access, implemented encrypted credential storage for OAuth tokens, strict tenant data isolation, and audit logging for account access security practices the original internal tool never needed to consider.
Infrastructure: Deployed on AWS ECS with auto-scaling to handle variable load across tenants, with RDS for the managed multi-tenant database.
Implementation Process
- Discovery & Requirement Analysis (Weeks 1–3): Audited the existing internal tool’s codebase and reporting logic to determine what could be preserved versus what needed to be rebuilt for multi-tenancy, and interviewed the agencies who had expressed interest to validate must-have features.
- Prototype / PoC (Weeks 4–6): Built a working multi-tenant prototype with OAuth-based ad account connection for one platform (Meta), validated with two beta agency partners before expanding to the other platforms.
- Development (Weeks 7–15): Built out full multi-tenancy, all three ad platform integrations, the white-label reporting feature, and subscription billing.
- Integration (Weeks 16–17): Connected Stripe Billing and finalized the OAuth flows and tenant data isolation implementation across all platforms.
- Testing (Weeks 18–19): Ran a closed beta with 8 external agencies, validating real ad account connections, report accuracy, and billing flows before public launch.
- Deployment (Week 19): Public self-service signup launched, with the founding agency’s own client reporting migrated to run on the new multi-tenant platform as its first tenant.
- Optimization (Week 20 and ongoing): Tuned data refresh scheduling and caching based on real multi-tenant usage patterns from the beta cohort.
Key Features Delivered
- Multi-tenant SaaS architecture with secure, isolated tenant data
- Self-service OAuth-based ad account connection for Meta, TikTok, and Google Ads
- Unified cross-platform performance reporting and dashboards
- White-label client reporting for subscribing agencies
- Stripe-based subscription billing with tiered, usage-based pricing
- Scalable, rate-limit-aware data refresh pipeline across all connected platforms
- Encrypted credential storage and audit logging for connected ad accounts
Business Results
- The platform launched with 8 paying beta agencies converting to full subscriptions, validating market demand beyond the original word-of-mouth interest
- New agency onboarding, from signup to first connected ad account, averaged under 15 minutes, fully self-service with no sales involvement required
- The founding agency’s own internal reporting performance improved significantly after migrating onto the new, properly architected platform
- Dashboard load times remained consistent as tenant count grew, validating the caching and multi-tenant architecture under real, growing load
- The agency began fielding inbound interest from additional agencies beyond their existing network, indicating the product could grow independent of the founding team’s personal relationships
- Recurring subscription revenue became a genuine second business line for the agency, distinct from and complementary to their core services revenue
Technology Stack
| Layer | Technology |
|---|---|
| Frontend | React, Next.js, TypeScript |
| Backend | Node.js |
| Database | PostgreSQL (multi-tenant, row-level security) |
| Caching | Redis (AWS ElastiCache) |
| Billing | Stripe Billing |
| Ad Platform Integrations | Meta Ads API, TikTok Ads API, Google Ads API |
| Cloud Infrastructure | AWS ECS, RDS |
Why the Solution Worked
The platform succeeded because it preserved what already worked the agency’s genuine domain expertise in what made ad performance reporting actually useful to marketers while rebuilding the parts that needed real SaaS engineering discipline: multi-tenancy, security, and billing. A ground-up rewrite that discarded the original reporting logic would have thrown away the exact insight that had generated organic market interest in the first place.
Validating with real beta agencies before public launch, rather than assuming the internal tool’s value would translate directly to external customers, also mattered it surfaced real onboarding friction and feature gaps (like white-label reporting) that weren’t obvious from inside the founding agency’s own perspective.
Future Scalability
The SaaS platform is built to support continued product growth, including:
- Adding additional ad platform integrations (LinkedIn, Pinterest) as customer demand indicates
- Building an API for agencies to pull report data into their own client-facing tools rather than only using the built-in reporting
- Adding predictive budget optimization recommendations using the now-centralized cross-platform performance data
- Expanding white-label capabilities to support more extensive custom branding for larger agency customers
Because tenant isolation, ad platform integrations, and billing are built as distinct, well-defined services, new platform integrations and features extend the existing architecture rather than requiring structural rework.
Final Outcome
An internal tool built to solve one agency’s own reporting problem became a genuine multi-tenant SaaS product serving multiple paying agency customers without losing the domain expertise that made the original tool valuable in the first place. The agency now has a second, scalable revenue line built on real product-market validation rather than assumption.
