Market View
I joined NAIS to build a product that didn't exist. Seven years later, it had shipped AI, been adopted into a top graduate program, and become the most advanced enrollment tool in the independent school market. This is how I built Market View.
Executive Summary
Schools were closing because they couldn't see the enrollment cliff coming. Admission directors had data but no tools. I built Market View—an enrollment platform with drive-time mapping, demographic forecasting, and AI-powered benchmarking. My team developed the core GIS algorithms and published them as open source on Hex.pm.
The Problem
Enrollment is everything for independent schools. The 2008-2010 recession made that clear—dozens of schools closed because they saw the cliff too late. When I joined NAIS in January 2018, admission directors were making six-figure decisions without geographic intelligence: no market visualization, no demographic forecasting, no way to see NAIS' 30+ years of data on a map. The data existed, but the tools did not.
Options Considered
Enhance DASL
DASL excels as a data collection and benchmarking platform—that's its core mission. Adding geographic visualization would have expanded DASL beyond its purpose. Better to build a complementary tool that leverages DASL's data strengths.
License Existing Software
Competitors charge ~$30K/year and none had what we needed: geographic mapping + demographic analysis + NAIS's proprietary 30-year dataset. Licensing would lock schools into expensive contracts for inferior tools.
Build from Scratch (Chosen)
Control the roadmap. Keep it free for members. Own the technology. Build toward AI capabilities on our timeline. Selected for strategic control and member value.
The Solution
Q1 2018: Built an HTML/CSS prototype in my first weeks—proved the concept and got leadership buy-in. Q2-Q3 2018: Hired Dockyard, one of the best Elixir shops in the world, to build the MVP. September 2018: Shipped production in 8 months from hire. Late 2018 to present: Transitioned off Dockyard, built an internal team, and shipped 7 major versions culminating in AI-powered Comparisons in Q4 2025.
Technical Bets That Paid Off
Elixir/Phoenix: Functional paradigm ideal for data pipelines; real-time processing at scale.
Conrex Algorithm: Drive-time isochrones using real traffic data. Straight-line radius circles lie—a family 20 miles on a highway is closer than 10 miles through city traffic. I implemented Paul Bourke’s CONREC algorithm and published it on Hex.pm.
Deep DASL Integration: 30 years of proprietary independent school data. No competitor can replicate this.
AI-Powered Comparisons: Shipped Q4 2025. Algorithmically generates peer cohorts and delivers AI analysis with recommended actions.
Open Source Contributions
Conrex (Hex.pm): CONREC algorithm implementation for drive-time isochrones.
ember-media-query (GitHub): Responsive component rendering in Ember.js.
The Competitive Moat
- Us: Free. Competitors: $30K/year.
- Us: AI-powered (shipped Q4 2025). Competitors: Limited AI capabilities.
- Us: 30 years of DASL data. Competitors: Public data only.
- Us: Drive-time isochrones (real traffic). Competitors: Straight-line radius circles.
- Us: Direct integrations (Blackbaud, Veracross). Competitors: Manual spreadsheet uploads.
Implementation
Q1 2018: Prototype
Built HTML/CSS prototype in first weeks—proved concept, secured leadership buy-in
Q2-Q3 2018: MVP Build
Engaged Dockyard (premier Elixir consultancy) for 6-month MVP development
July 2018: First External Presentation
Institute for New Heads—external validation at just 6 months
September 2018: Production Launch
MVP shipped; 8 months from hire to production
February 2019: Vanderbilt Adoption
Peabody College (#2 Educational Administration) built curriculum around Market View; invited back July 2019
Q4 2020-Q2 2022: Versions 4-6
Financial gap calculations, demographic forecasting, trends dashboard, Blackbaud & Veracross integrations
Q4 2025: AI Launch
Market View 7 with AI-powered Comparisons for peer benchmarking
Impact & ROI
Risks & Mitigation
| Risk | Mitigation |
|---|---|
| Data quality skepticism | Integrated with DASL (30+ years validated); documented methodology clearly |
| Adoption friction | Made it free; built intuitive UI; wrote 35+ documentation articles |
| Geographic inaccuracy | Implemented drive-time isochrones with real traffic data (not radius circles) |
| Integration complexity | Built direct connections to Blackbaud, Veracross, Finalsite—killed spreadsheets |
Stakeholders
Sponsor:
NAIS leadership (funded as strategic member benefit)Users:
Admission directors, heads of school, board members, Vanderbilt graduate studentsDevelopment Partners:
Dockyard (MVP), Phillips Exeter & Peddie School (AI Symposium case studies)Key Features
Drive-Time Isochrones
Shows 15/30/45-minute commute zones using real traffic data. Straight-line circles lie—highways vs. city streets change everything. Built on my open-source Conrex implementation.
Trends Dashboard
Tracks enrollment, demographics, and affordability over time. Became the default landing page—users voted with their clicks.
Affordability Analysis
Breaks down tuition into gap vs. contribution by zip code. Schools finally see who can actually pay.
Your Data + Integrations
Syncs with Blackbaud, Veracross; accepts CSV uploads. Killed manual spreadsheet formatting forever.
AI Comparisons
Benchmarks against algorithmically generated peer cohorts. AI provides analysis and actionable recommendations.
Demographic Mapping
Overlays census, income, diversity data on maps. Find communities matching target enrollment profile.
Lessons Learned
Built HTML/CSS prototype myself—validated assumptions cheaply and got leadership buy-in fast. Hired Dockyard for MVP—quality Elixir codebase shipped in 6 months. Transitioned to individual devs for more control and lower costs. Free member benefit removed cost barriers. Deep DASL integration created an unreplicable competitive moat. Academic adoption at Vanderbilt provided third-party validation that convinced skeptical administrators. Hundreds of user conversations drove better product decisions than any formal research process.
The validation most product leaders never get:
- Vanderbilt’s Peabody College (#2-ranked Educational Administration program) built a curriculum unit around Market View and invited me to present—twice (February 2019 and July 2019)
- Inbound demand: Schools and consortiums request presentations; I don’t pitch them
- External speaking within 6 months: Institute for New Heads (July 2018), People of Color Conference (Nov 2018), NAIS Annual Conference (Feb 2019)
- Hundreds of user conversations shaped every major feature decision
What I personally built:
- HTML/CSS prototype that sold leadership on the project
- Hired Dockyard (premier Elixir consultancy) for MVP, then built the internal team
- Two open-source projects: Conrex algorithm (Hex.pm), ember-media-query (GitHub)
- 35+ documentation articles, written by hand
What I’d do differently:
- Invested in data integrations earlier—spreadsheet uploads created friction
- Advocated for AI sooner—we shipped 18 months after board mandate, but I’d seen the opportunity years earlier
- Documented user success stories systematically from day one
