Choosing a real estate software development company is usually a legacy decision wrapped in a software decision. Most production real estate platforms still run on frameworks that reached end of life years ago, and the vendor you pick will either migrate that code without taking your business offline or sell you a rewrite that stalls. The checks that separate a useful vendor from an expensive one are specific: whether they can read your existing codebase before they propose anything, whether their migration plan keeps the live platform running, whether they know MLS, IDX, RESO and USPAP, whether they build the back office as seriously as the listing page, and whether the engineers who learn your system are still on it in year three. This guide lists nine checks, the red flags and a 30-minute call script, drawn from embedding engineers inside an appraisal management company's platform and from the legacy migrations and property platforms EltexSoft builds.
Check 1. They can read your legacy code before they propose a rewrite
The typical real estate platform that comes to a vendor runs on PHP 5, Kohana, CodeIgniter or CakePHP, with a jQuery-heavy frontend and a monolith deployed by hand with no CI/CD. It still runs the business. A vendor who proposes a rewrite from scratch on the first call has not read the code; a vendor who asks for repository access and a week to map the architecture has.
When EltexSoft engineers joined a Southern California appraisal management company in March 2025, the first job was learning a PHP system, its cloud infrastructure and a mix of proprietary and third-party tools before shipping anything. The client's review on Clutch notes that the team learned the architecture quickly and then handled both planned features and ad hoc requests. Ask any vendor for a codebase they inherited, what they kept, what they replaced and in what order.
Check 2. Incremental migration, not a big-bang rewrite
A big-bang rewrite of a platform that lenders, agents or tenants use every day fails for a predictable reason: the old system keeps changing while the new one is being built, and the cutover never comes. The alternative is incremental migration: move one module at a time to a modern stack, run old and new side by side, keep the data in sync, deploy without downtime, and retire the legacy code piece by piece.
Ask for the migration plan in writing: which module goes first and why, how the database migration is staged, how APIs are restructured, how integrations are rewired, and what the rollback is if a release goes wrong. EltexSoft's standard path is legacy PHP to Laravel with React or Vue on the frontend, PostgreSQL or MySQL, Elasticsearch and Redis, on AWS or GCP; the published range for a Kohana or CodeIgniter to Laravel migration is $50K to $150K over 4 to 8 months, and the real estate software development page has the details.
Check 3. They know the data standards: MLS, IDX, RETS, RESO, USPAP
Real estate software lives on other people's data. Listing portals depend on MLS feeds through IDX providers, the legacy RETS protocol or the RESO Web API; appraisal platforms live under USPAP and state licensing rules; mortgage and e-signature integrations sit on top. A vendor who has not integrated these will learn on your budget and your deadline.
Ask which standards the vendor has shipped against, how listing data is refreshed and deduplicated, what happens when a feed changes its schema, and how compliance rules (appraiser licensing by state, USPAP reporting requirements) are enforced in code rather than in a manual.
Check 4. The back office is built as seriously as the front end
An appraisal management system is order management, appraiser assignment, compliance tracking, status dashboards and invoicing; a property management platform is tenant portals, leases, rent collection, maintenance requests, vendor management, financial reporting and multi-property dashboards. The listing page is the easy part. The hard engineering sits in assignment rules, turnaround SLAs, exception handling and the reports that an operations team reads every morning.
A vendor who has done this will ask about your order volume, your turnaround commitments and who the users are (appraisers in the field, lenders, AMC staff, property managers, tenants) before talking about screens. Ask to see the operations screens of a platform they built and how an exception (an appraiser misses a deadline, a rent payment fails) flows through them.
Check 5. Search and marketplace engineering at real listing volume
Marketplaces and listing portals are search products: map search, geo queries, filters that stay fast at hundreds of thousands of listings, lead capture that routes to the right agent, scheduling and documents. Elasticsearch handles the search; the vendor's job is the indexing strategy, the relevance tuning and the cost at your volume.
Ask for numbers from a live portal: listing count, search latency, how leads are routed, and what the two-sided logic looks like when supply and demand are both users. Published ranges at EltexSoft: a listing portal with MLS integration $60K to $120K over 4 to 6 months; a marketplace with web, mobile, payments and search $120K to $300K over 6 to 12 months.
Check 6. Staff augmentation that becomes part of your team
Many real estate companies do not need a vendor to build a product; they need two to five senior engineers inside an existing team, on an existing codebase, now. That is a different test: how fast the engineers learn the system, whether they follow your development practices, and whether velocity goes up without overhead going up with it.
That client's Clutch review reports the results so far: feature delivery 25 to 35% faster, fewer production issues and faster fixes, a shrinking backlog, and added engineering capacity without a matching rise in overhead. The reviewer's headline was that the team operated as a true extension of the internal team. The same review names the thing to improve: real-time coordination across time zones and more proactive status updates. Ask every vendor for that level of specificity about their own weak spot, and ask how many overlap hours with your time zone are in the contract and what the written status cadence is.
Published rates: $50 to $99 per hour for a dedicated senior real estate engineer, roughly $8K to $16K a month full time, with engineers typically in your standup within 14 to 21 days.
Check 7. AI and valuation features with a data boundary
Automated valuation models, comparable sales analysis, predictive pricing, lead scoring, tenant analytics, document extraction from appraisal reports and leases, natural language search over listings: these are now standard requests, and JLL's data cited on our real estate page shows AI pilots in commercial real estate going from 5% to 92% of companies in three years. Most pilots never reach production.
Ask what "working" means before you ask which model: accuracy against a held-out set of closed sales for an AVM, extraction accuracy on your own documents, and what the system does when confidence is low. Then ask where your data goes: which models, under which agreement, and whether raw records leave your cloud. The AI vendor guide covers the evaluation harness, cost per request and the human-in-the-loop boundary in detail.
Check 8. Team continuity, code ownership and the first call
A real estate platform is a ten-year asset, and the engineer who knows why the assignment rules work the way they do is part of it. Ask for the average tenure of engineers on client accounts, the written replacement commitment, and whether the people on the first call will write the code. At EltexSoft the average client engagement is 3.4 years, replacement within two weeks is in the contract, the first call is with an engineer, and the how we work page has the terms.
Then check ownership in the contract: full work-for-hire assignment of source code, configurations, documentation and deployment artifacts, with MLS feed accounts, cloud accounts, domains and e-signature providers registered in your name, not the vendor's.
Check 9. Pricing you can compare, and the cost after launch
Published numbers to compare against. A property management MVP runs $40K to $80K over 3 to 5 months; a listing portal with MLS integration $60K to $120K over 4 to 6 months; an appraisal management platform $100K to $250K over 6 to 10 months; a real estate marketplace $120K to $300K over 6 to 12 months; a framework migration $50K to $150K over 4 to 8 months. Plan on 15 to 20% of the build cost per year for maintenance. US PropTech agencies commonly charge $100 to $250 per hour for comparable seniority; nearshore and Eastern European senior engineers run $50 to $99.
Ask what the rate includes (QA, DevOps, a technical lead, project management), insist on month-to-month terms, and get the maintenance number in writing before you sign the build. For the general version of the MVP math, see what an MVP really costs.
Red flags
- A rewrite from scratch proposed before anyone has read your code.
- No written migration plan with module order, parallel running and rollback.
- Blank looks at RESO, RETS, IDX or USPAP.
- A demo of the listing page and nothing for operations, assignment or reporting.
- No numbers from a live portal: listing count, search latency, lead routing.
- Staff augmentation sold as "resources" with no plan for learning your codebase.
- No answer to "how many overlap hours with our time zone".
- MLS feed, cloud or domain accounts registered in the vendor's name.
A 30-minute call script
- Here is our stack and our repository. What would you migrate first, and what would you keep?
- Walk me through a legacy platform you migrated without downtime: module order, data sync, rollback.
- Which of MLS, IDX, RETS, RESO Web API and USPAP have you shipped against?
- Show me the operations screens of a real estate platform you built.
- What are the listing count and search latency on a live portal of yours?
- If we only need engineers inside our team, how fast do they learn our system, and what did velocity do for your last augmentation client?
- How many overlap hours with our time zone, and what is the written status cadence?
- What would make an AVM or document extraction feature "working", and where does our data go?
- What is your rate, what does it include, what is the minimum term, and what will maintenance cost in year two?
Where EltexSoft fits
We are a Los Angeles software engineering studio founded in 2015, with 35 to 50 senior engineers and no juniors on client projects. Our real estate work includes engineering inside a US appraisal management company's platform (staff augmentation since 2025, reviewed on Clutch); the real estate software development page lists what we build (property management platforms, listing portals with MLS integration, real estate marketplaces, legacy PHP to Laravel migrations) and the published prices above. Rates are $50 to $99 per hour on a month-to-month retainer, and the first call is with an engineer. If you are shortlisting, we are one of the teams worth 30 minutes. For the general version of this framework, see how to choose a software development partner.