A regional brokerage group · $8K → $52K platform MRR.

This archetype is a pattern we ship into reliably in residential real estate: a regional brokerage group with nine offices and roughly 140 agents across a mixed sales-and-lettings book. The second-generation managing broker running it knows exactly what is capping the group. The technology.

The pre-engagement state: every listing hand-entered twice, once into the regional MLS and once into a decade-old WordPress site. Viewing requests arrive by phone and inbox. Agent follow-up lives in spreadsheets. And a thin internal "platform", half-licensed to two friendly partner brokerages for $8K MRR combined, is held together by one overworked in-house developer.

The group's ambition was specific: stop being a brokerage with a website, and become the brokerage whose platform other regional offices license. The raw material was there. The MLS membership gave them clean, authoritative listing data they were ignoring in favour of manual re-entry. The nine offices gave them a built-in first tenant cohort for anything multi-office.

And the two partner brokerages already paying for the old platform proved something useful: regional operators would pay for shared listing infrastructure, once it stopped embarrassing the agents using it.

What they didn't have was software. The old platform pre-dated the RESO standards era, so listing data flowed in by hand. Search was a paginated filter list with no map. Viewing scheduling was a phone number. And there was no CRM at all: a viewing request that arrived on a Friday evening had a coin-flip chance of ever being answered.

The in-house developer knew all of this. What the group lacked wasn't diagnosis. It was delivery capacity, which is the honest shape of most brokerage-platform engagements we take on.

Three structural problems compounded the ceiling. First, the double-entry tax. Getting a property live took an average of 45 minutes of manual work per listing: re-typing MLS fields, resizing photos, fixing whatever the WordPress editor mangled. Multiply that across roughly 90 new listings a month.

Agents did the maths and simply stopped bothering. A third of the group's inventory never made it to their own site at all. The group was paying to send its own buyers to national portals.

Second, the demand side leaked. Listing pages converted views to viewing requests at 2.1 percent. Buyers and tenants couldn't search by area on a map, and couldn't see what was near a school or a station. When they did want a viewing, the call-to-action was a phone number answered during office hours.

Every viewing that did get booked cost two to four phone calls of back-and-forth between applicant, agent, and, for tenanted properties, the current occupier.

Third, the licensing story had no product under it. The two partner brokerages paying into the $8K MRR were buying goodwill, not software. Churn was a renewal-dinner conversation away. Charging real money to real third-party offices meant multi-tenancy, per-office billing, and an onboarding path that didn't require the in-house developer to hand-configure every new office. None of that existed.

The engagement brief we scoped against was blunt: kill the re-keying, fix the funnel, and make the platform licensable. In that order, because the feed is what makes everything downstream trustworthy.

Workstream 1 · RESO Web API feed sync. Manual entry went first. We replaced it with a scheduled replication service against the regional MLS's RESO Web API: Data Dictionary field mapping, delta polling on modification timestamps every 15 minutes, media replication with responsive image derivatives, and a nightly reconciliation job that flags drift between MLS truth and platform state.

A listing now lands on the platform pre-populated. The agent's remaining 6 minutes are pure enrichment: photo ordering, a description pass in the group's voice, and viewing-slot rules. Average time-to-list fell from 45 minutes to 6. The third of inventory that never used to reach the site now ships automatically.

Workstream 2 · Map-first search UX. Search was rebuilt around the map instead of the filter list, on Google Maps Platform. Clustered markers resolve to price pins on zoom. Buyers can draw their own search boundary, switch on school and transit layers, and save alerts keyed to map areas rather than postcode strings.

Listing pages went server-rendered and Core Web Vitals-clean: LCP under 2 seconds on mid-tier mobile, layout shift pinned at zero with explicit media dimensions. A search experience that stutters on a phone loses the applicant before the viewing button even exists.

Workstream 3 · Automated viewing scheduling. The phone number was replaced with self-serve booking. Viewing slots derive from agent calendars and per-property rules: tenant-notice windows for occupied lets, open-house blocks for launches. Applicants get instant confirmation with calendar invites, reminders at 24 hours and 2 hours, and a no-show recovery flow that offers the next available slot.

Combined with the map-first rebuild, listing-view to viewing-request conversion moved from 2.1 percent to 6.8 percent. And each booked viewing stopped costing the front desk three phone calls.

Workstream 4 · Lightweight brokerage CRM. Every viewing request now auto-creates a lead, with source, listing, applicant details, and stage, in a purpose-built brokerage CRM deliberately smaller than the enterprise suites agents refuse to open. Pipeline per office. Follow-up nudges when a lead sits untouched for 48 hours. Viewing-feedback capture on the agent's phone, in the car outside the property. A weekly digest to each managing broker.

Friday-evening enquiries stopped dying in inboxes. The group's own measure of "leads with a first response inside 4 hours" went from roughly half to 96 percent.

Workstream 5 · Multi-tenant licensing. The internal build became a licensable product: tenant-isolated data with per-office theming, seat-and-office billing on Stripe, a self-serve onboarding path that connects a new office's MLS credentials and branding without engineering time, and an admin plane for the group to manage tenants. This is the SaaS-development half of the engagement.

The two goodwill licensees were migrated first. Then the platform sold on its own terms. Platform MRR climbed from $8K to $52K across 18 months (6.5x), with month-12 logo retention at 88 percent.

This archetype is one of the most repeatable engagement shapes in our real-estate software development practice. It fits a regional brokerage group with 5 to 15 offices and mixed sales and lettings. Usually there is an MLS membership the group under-uses, a listings workflow held together by re-keying, and an ambition to license the platform to peer offices.

If that sounds like your group, the numbers travel. The 22-week timeline holds steady. The workstreams compress or expand in the same proportions. The metrics typically land within plus or minus 20 percent of the archetype numbers.

Five capabilities transfer directly. RESO Web API feed sync with reconciliation, because the feed is the foundation and nothing downstream is trustworthy without it. Map-first search UX with Core Web Vitals discipline. Automated viewing scheduling with occupier-aware slot rules. A lightweight brokerage CRM that agents actually open. And multi-tenant licensing infrastructure with per-office billing.

The sequence transfers too: feed first, funnel second, licensing last. It works just as well for lettings-only operators, commercial brokerages on RESO-adjacent data sources, and new-build developers syndicating inventory.

The five headline numbers on this page are the archetype's canonical pillars: $8K to $52K platform MRR, 6.5x, viewing-request conversion 2.1% to 6.8%, month-12 logo retention 88%, and time-to-list cut from 45 minutes to 6. You'll find the same five on our real-estate industry page, deliberately identical. A case study and its industry page quoting different numbers is how trust dies.

Every proptech engagement starts with a 30-minute discovery call. Scope, timeline, and budget come back in writing within 48 hours.