Arc theory.est. 2023

Program & Incentive Platform

Technical Lead · Solutions Architecture (sole technical resource)

A global consumer-technology company runs an internal program that gives employees and trusted communities early access to new offerings — and issues incentives to drive participation. It began as a feedback channel and grew into a strategic program, but the incentive side carried real cost and compliance exposure. Danielle was engaged as the sole technical resource on the platform.

The team needed to rapidly expand its reach across more programs and more user groups — without growing the team to match. That called for a platform that could support the dynamic, ever-changing nature of the program while keeping staff lean.

The platform is built to be configured, not rebuilt — and underneath it, a set of strategic moves and reusable modules carry the work.

The idea
One configurable platform, many programs

A new program spins up through an interface — its survey, budgets, monthly limits, and incentive structure all set with minimal custom configuration — and every module below flexes to that program's rules.

Key activities
Rules & controls
Systematized the program's rules — budgets, compliance, eligibility — into automated controls that reduce spend and keep programs compliant by default.
Data normalization
Rebuilt a fragmented, citizen-built data model into a clean, normalized schema with real referential integrity.
Pipeline consolidation → less surface area
Six separate data pipelines collapsed into one — far less of the system to secure, debug, and maintain.
Key modules
Content management
Create and manage standardized program content in one place — no duplication, no copy-paste setup.
Dynamic survey & sign-up
Surveys and sign-up flows that adapt per program, configured through the interface.
Automated controls
Budget shut-offs and sign-up / beta limits enforced automatically — the guardrails run themselves.
Incentive distribution
Automated distribution of incentives to user groups at their different price points.
User & group management
Assign users to special groups, then target programs, incentives, and access accordingly.
Technical evolution

From fragmented to normalized

Before: many pipelines feeding duplicated tables across separate bases. After: fewer sources into one normalized database — and the groundwork for moving onto the internal stack.

Before
6 data pipelines
Base 1SurveysSurveysBookingsBookings
Base 2SurveysBookings
After
1 pipeline
Normalized databaseSurveysSign-upsIncentivesProgramsBookings
Integration layerCompliance
AvailabilityStats
General platform maturity →
Normalized dataAccess modelsSchema designComplianceSecurityPRDs

The rebuilt platform, a new incentive model, and a set of compliance changes rolled out together in March 2026. The numbers below compare the months before to the months after — same program, same team.

Avg monthly program spend · after the redesign
−69%

≈$110K → $34K in average monthly incentive spend

Compliance flags / month
−82%

≈152 → 28 flagged reservations per month

Strategy, not just system

That reduction wasn't purely system-driven — it was a strategy play we pushed while making the platform simpler. The core problem was slippage: a fixed-value model (issuing a $150 coupon against a $50 item) left room for misuse and extra spend, while maintaining complicated per-category rules was impractical. We drove the program toward a percentage-based, two-tier incentive model — fewer rules to manage and far less leakage. It's a clear example of aligning business and technical strategy: the simpler model was both cheaper to run and easier to build.

Time saved across admin pillars

The redesign cut hands-on time across every administrative task — including a self-serve data dashboard that replaced join-heavy manual exports. The bar shows how much of the original time remains.

Page Creation
~45 min ~2 min
Survey & Sign-Up Set-Up
~40 min ~5 min
Coupon Upload
2–3 hrs ~10 min
Report Creation
~2 hrs ~5 min
Data Export
~55 min ~3 min

The rebuild reduced technical debt, paved the path toward moving onto the company's internal stack, and multiplied the team's capacity — supporting far more programs and user groups without adding headcount.

Source: program analytics, fully anonymized — client, program, and internal system names removed. Reservations are bucketed by claim date; one-off enforcement-sweep detections are excluded.