Every pilot became a contract
A PropTech platform rebuilt from the assets of a collapsed incumbent went from 6 to 184 government contracts in fifteen months, with a replatform that took NPS from negative 5 to positive 40 and onboarding from three months to two weeks.
Context
Hera sells foreclosure and vacant-property registration to municipal governments. In June 2023 the incumbent administrator for the sector entered a bankruptcy proceeding, and the municipalities it served needed somewhere to go. Hera raised $300K in debt, acquired the incumbent's assets out of the proceeding, and set out to rebuild the platform and carry those communities across. When I joined at founding the company had six contracts, a platform with an NPS of negative 5, and onboarding that took three months. The expectation that came with the raise was not a better demo. It was production.
The buyers were governments, which meant the bar was higher than feature parity. Annual client audits, including Jacksonville FL, CJIS compliance, independently audited financials, and procurement processes that punish vendors who overpromise. Trust had to be rebuilt while the existing product kept running.
My role
I owned the technical vision and the delivery system.
I delivered the new platform MVP within five months of funding, built the engineering and support team from zero to six plus people across the US, Portugal, Brazil, and Australia, and implemented EOS across the organization so scorecards, meeting rhythms, and ownership kept pace with growth. The platform earned SOC 2, and the operating side passed every government audit it faced. At that stage a CTO title understates the job. I ran forecasting, budgeting, renewals, the hiring plan, vendor governance and contract negotiations including platform licensing and external data purchases, on an annual budget of roughly $1M. I also sold, because an operator at that stage has to.
What changed
The acquired assets arrived fragmented and incomplete, with no credentials and no documentation. All of it had been lost in the bankruptcy. Months went to regaining access, reconstructing what had changed, and rolling the platform back to an earlier version clients trusted. A new development team then had to learn the business and technical context of a large regulated platform with no documentation to hand them, while data integration reconciled three incompatible sources. The rebuild served live government clients the entire time.
A proprietary AWS and OutSystems platform replaced the system users had already voted against. NPS moved from negative 5 to positive 40 on SurveyMonkey data. Releases went from every other month to every three weeks, sprint velocity rose 85% on Jira reporting within 60 days of launch, and technical debt fell 30% using OutSystems Mentor code quality audits. The GIS dashboard, roughly 2,500 OutSystems AO, combined foreclosure, ordinance, and compliance data into real-time municipal insight, with Authorize.net, MapBox, ATTOM, and HubSpot integrated underneath.
Government onboarding was redesigned as part of the platform rather than a services afterthought. Ramp time fell from three months to two weeks, every pilot converted to a contract, and the client list grew to include Jacksonville FL, the City of Miami, Cicero IL, Dover NJ, Islip NY, Erie PA, and Santa Ana CA.
I sold alongside building the platform and closed about $1.3M in net-new ARR in year one, roughly $6.5M in total contract value on five-year terms. The logos included Jacksonville FL at $300K ARR, Miami FL at $150K, Atlantic City NJ at $150K, Camden County NJ at $150K, Palm Beach County FL at $100K, Passaic County NJ at $100K, Oyster Bay NY at $80K, and Islip NY at $75K. Average cycle was about three months in structured municipal procurement, with the Jacksonville RFP won in two. The method was unglamorous: weekly analysis of foreclosure filings to find the communities carrying the most risk, ten or more stakeholder meetings a month, five to eight live pursuits at a time, and a localized proof-of-concept demo in the first meeting whenever a buying committee told me the data could not exist.
Operations staff grew only from two to four while the customer base grew roughly thirty times. The remittance cycle, which had taken up to two people two weeks and sometimes a month, was cut sharply in both time and staffing even as the customer base grew. Proactive registration and enforcement activity rose 270%, and total registrations rose 400% for the communities the platform served, across roughly 325,000 site visits a month. Security and vendor work took another 20% off insurance premiums and operating costs, and the platform was featured at the Florida League of Cities.
Government contracts during my tenure.
Revenue growth during my tenure of fifteen months.
NPS across the replatform.
Pilot-to-contract conversion, with 100% client retention during my tenure.
Government onboarding, redesigned as part of the platform rather than a services afterthought.
I led the technical vision and the delivery system. The team built the platform. The municipalities achieved the compliance outcomes. Every result on this page describes my tenure.
What it changed in my thinking
A replatform is a trust decision before it is a technical one.
An NPS of negative 5 means the users have already voted. More features do not change that verdict. Reliability, speed, and an onboarding experience that respects the client's time are what change it, which is why the replatform was designed around uptime, response time, and ramp time rather than a feature list.
Carrying the number changed how I built. When the person answering the security questionnaire is also the person who has to close the deal, the roadmap stops being a wish list and starts being an argument you have to win in a room with five to ten stakeholders who have every reason to say no.
The pilot-to-contract number is the one I watch. A pilot that converts every time is not luck. It means the demo made a promise the production system could keep, and that the operating model, from onboarding to remittance, was designed as part of the product rather than around it.
Have a similar operating constraint?
Describe the stage, the system, and what is not moving. I will review the message directly.