Mykhailo Ralduhin

Written by: Senior QA Engineer

Mykhailo Ralduhin

Posted: 18.11.2026

22 min read

A listing is a data record that passes through eight systems and at least three regulators before anyone gets keys. Each hand-off is a place for it to go stale, show to the wrong person or violate a rule written into law.

This article is a defect catalogue organised by that lifecycle: for each stage, the defects we see most often, how they show up, what they cost and the test that catches them. We have tested real estate products since 2010 for brokerages, rental platforms and property operators; what we offer is on our real estate software testing services page. This is the catalogue behind it.

How to read the tables. Every table has the same four columns:

1.

Defect (what is wrong in the software)

2.

Shows up as (what a user or auditor sees)

3.

Impact (what it costs)

4.

Test that catches it (the automated or manual check that would have stopped it).

The life of a listing: Feed → Listing → Lead → Showing → Offer → Transaction → Close → Operate.

Ready to take your software quality to the next level?

Why real estate software fails differently

Real estate software fails differently because the most important data arrives from systems you do not own, the rules live inside the search and pricing logic, and each transaction is too valuable to retry.

1.

Data you do not own. Multiple listing service (MLS) feeds, syndication portals, public records, tenant screening and valuation data each have their own schema, refresh cadence and licence terms – and they change without asking.

2.

Regulation inside the logic. Fair Housing constraints apply to search filters, ranking and generated copy; RESPA/TILA disclosure timing is a workflow timer; state-specific forms are a version matrix. Compliance is not a module; it is scattered through the code.

3.

High-value, low-frequency transactions. A payment gateway can retry a $40 order. A closing that fails at signing is a lost client and a reputational event.

4.

Many parties, many views. Buyer, seller, two agents, lender, title company, inspector – each sees a different slice of the same transaction, and a permission error is a privacy incident.

5.

Local spikes. A well-priced listing in a hot market gets thousands of views in its first hour, and buyers are on their phones: in NAR's 2024 profile, 43% of buyers started their search online and 69% used a mobile device.

Stage 1 — MLS and IDX feeds: the data arrives

MLS feed testing is contract and freshness testing: the RESO Web API (which replaced the RETS standard) delivers listings under display rules that are part of the data licence, and a silent failure here makes every downstream screen wrong at once.

Context: brokerages and portals ingest listings through RESO Web API or Internet Data Exchange (IDX) vendor feeds, then syndicate to Zillow, Realtor and others. Each MLS has its own resource definitions, status vocabulary, refresh cadence and attribution requirements.

Defect
Shows up as
Impact
Test that catches it

Schema change in a RESO resource (renamed or retyped field)

Listings silently stop updating; prices and photos freeze

Stale inventory shown to buyers; MLS compliance exposure

Contract tests per MLS against recorded payloads; schema-drift alert

Refresh cadence violated under load

Homes shown as Active that went under contract hours ago

Wasted showings; agent credibility

Freshness monitor: time since last successful sync per MLS, with SLA threshold

Status mapping error (e.g., Active Under Contract mapped to Active)

Portal displays a status the MLS display rules forbid

Licence violation; feed suspension

Status-mapping fixture per MLS with every source status asserted

Photo ordering lost on ingest

Wrong hero image on the listing card

Lower click-through; seller complaints

Ordering assertion on media arrays in the ingest test

Duplicate listings across overlapping MLSs

Same property twice in search with different data

Confused buyers; attribution disputes

De-duplication fixtures with known cross-MLS duplicates

Missing attribution or disclaimer text

Listing shown without the required broker credit

Licence violation

Display-rule assertions on the rendered listing per MLS

Time zone on ListingContractDate or ModificationTimestamp

Days-on-market off by one; "new today" filter wrong

Misleading market data

Time-zone fixtures around midnight boundaries

Throttling or partial pages ignored

Partial updates; some listings never refreshed

Inconsistent inventory

Pagination and 429-handling tests; count reconciliation feed vs. database

The contract-testing approach – recording real payloads, asserting on schema and behaviour, running on a schedule against the live source – is laid out step by step in our API testing automation tutorial. For MLS feeds the schedule matters more than the deploy trigger: feeds change between your releases.

Stage 2 — Listings and search: the data is shown

Listing and search testing has one check that outranks all others: no filter, ranking factor or generated line of copy may discriminate on a protected class. Fair housing complaints reached a record 34,150 in 2023, and search and advertising logic is now part of the enforcement conversation.

Context: search and map, filters, saved searches and alerts, automated valuation (AVM) widgets, listing detail pages, agent attribution.

Defect
Shows up as
Impact
Test that catches it

Ranking factor or filter correlated with a protected class (school ratings as a boost, "family-friendly" in generated copy)

Different results for equivalent searches; flagged copy

Fair Housing Act exposure; enforcement action

Blocking rule tests on search inputs, ranking features and generated text against a prohibited-terms and proxy-feature list

Geo-search boundary error (polygon vs. radius, parcel on a boundary)

Listing missing from the neighbourhood it is in

Lost exposure; agent complaints

Geo fixtures on boundary parcels and irregular polygons

Saved-search alerts late or duplicated

Buyer told about a home after it is pending, or told twice

Churn; unsubscribes

Alert latency and de-duplication tests on the notification queue

Price-history rendering error

A "price drop" that never happened

Seller complaint; misleading advertising

Reconciliation of displayed history against feed events

AVM shown without the required disclaimer

Valuation presented as an appraisal

Regulatory and consumer-protection exposure

Rendering assertion for disclaimer presence per AVM widget

Map clustering hides listings at a zoom level

Homes vanish until the user zooms in

Lost exposure

Visual and count tests at every zoom level on fixture data

Accessibility failures on map, filters and photo galleries

Screen-reader users cannot search

ADA/WCAG exposure; excluded buyers

Automated scans plus manual assistive-technology run on the search path

Image-heavy pages slow on mobile

High bounce on the listing page

Lost leads from the 69% on mobile

Mobile performance budgets on listing pages across the device matrix

Accessibility on a search product is a legal and a commercial requirement; the framework we use – standards, risks, real flows – is described in our accessibility testing guide, and it transfers from finance to property without change. Mobile performance budgets and how to measure them on real devices are in our guide to performance testing for mobile applications.

Stage 3 — Leads, CRM and showings: the people arrive

Lead and CRM testing is mostly a question of who sees what: leads, contacts and showing data flow between portals, the brokerage CRM, agent apps and scheduling tools, and the access model is where defects turn into disputes.

Context: lead capture and routing, CRM, agent mobile apps, showing scheduling and lockbox integrations, SMS and email with consent rules (TCPA in the US, GDPR where applicable).

Defect
Shows up as
Impact
Test that catches it

Lead routed to an agent who left the brokerage

Unanswered enquiry

Lost client; commission dispute

Routing matrix with departed-agent, team-change and territory cases

Duplicate lead from two portals

Two agents call the same buyer

Buyer annoyance; attribution conflict

De-duplication fixtures across lead sources

SMS sent without a consent record

Message to a number that never opted in

TCPA liability per message

Consent-state assertion before every outbound message

Showing slot double-booked

Two buyers at the door

Seller and agent complaints

Concurrency tests on slot booking

Lockbox code issued for the wrong time window

Access when the seller is home — or no access at all

Safety and trust incident

Window-boundary tests on code issuance

Agent sees another team's leads or contacts

Data visible outside the permitted scope

Privacy incident; internal dispute

Role × data access matrix: agent, team, office, brokerage

Lead-source attribution wrong

Referral fee paid to the wrong source

Financial dispute

Attribution reconciliation against raw capture events

In practice — Compass

Compass provides a digital platform for buying, selling and renting real estate, with a web CRM for communication between agents and clients and companion mobile apps. Since 2016 our team – eight full-stack and one manual QA engineer – has taken full responsibility for testing it: 2,000+ test cases, 85% test coverage, 1,600+ automation scripts covering 80% of regression, regression 2.5× faster than before, 800+ bugs reported, 400+ of them caught by automated tests, and 4,000+ feature tickets verified. The stack runs on Cypress, Supertest and Locust in CircleCI. The lesson from brokerage scale: the CRM is where every integration converges, and the data-access matrix – agent, team, office, brokerage – is the suite that protects the business. Read the Compass case study.

Stage 4 — Offers and transactions: the money and the paperwork

Transaction testing is where real estate software meets the law most directly: form versions by state and date, disclosure windows measured in business days, signatures in a required order, and money that must move exactly once.

Context: offer management, transaction management, e-signature (DocuSign, Dotloop and similar), disclosures, earnest money and payments, lender and title integrations, broker compliance review. Under TRID, the Loan Estimate must be delivered within three business days of application and the Closing Disclosure received at least three business days before consummation – timers that the software enforces.

Defect
Shows up as
Impact
Test that catches it

Wrong disclosure form version for state and date

Outdated form signed

Unenforceable document; compliance finding

Form-version matrix: state × transaction type × effective date

E-signature envelope missing a signer or in the wrong order

Document completes without a required party

Invalid contract; re-execution

Contract tests on envelope creation asserting signer list and order

Disclosure window timer miscalculated (calendar vs. business days, holidays)

Closing Disclosure delivered two business days before closing

TRID violation; delayed closing

Timer tests with holiday calendars and business-day fixtures

Earnest-money payment duplicated on retry

Buyer charged twice

Refund, trust damage

Idempotency tests on payment submission

Party sees a document outside their role (buyer views seller's net sheet)

Confidential data exposed

Privacy incident; deal leverage lost

Party × document access matrix, every pair asserted

Commission split wrong on referral or team deals

Incorrect disbursement

Financial dispute

Golden commission calculations per split type

Closing-date arithmetic across time zones or holidays

Wrong deadline displayed to parties

Missed contingency

Date fixtures with holidays and multi-zone parties

Audit-trail gap in compliance review

Reviewer cannot see who changed a term

Failed audit

Audit-event assertions on every transaction mutation

Payment flows in a transaction product carry the same requirements as any checkout – authorisation, capture, refund, idempotent retries – and our payment gateway testing guide covers what to assert at each step. The transaction itself should be tested as one journey from accepted offer to recorded deed, which is the subject of our end-to-end testing guide.

Stage 5 — Rentals and residents: the other half of real estate

Rental platform testing adds consumer-protection rules to the catalogue – tenant screening under the Fair Credit Reporting Act, lease arithmetic, rent ledgers – and, for resident communities, a privacy problem: conversations between strangers.

Context: rental marketplaces, applications and screening, lease signing, rent payments and ledgers, maintenance requests, resident portals and referral programmes.

Defect
Shows up as
Impact
Test that catches it

Adverse-action notice not issued after a screening decline

Applicant rejected with no notice

FCRA violation

Screening workflow tests asserting notice generation and content

Application fee charged twice on retry

Duplicate charge

Refund; complaint

Idempotency tests on fee payment

Pro-rata rent on mid-month move-in miscalculated

Wrong first invoice

Dispute; manual correction

Lease-calculation matrix by move-in date and billing method

Late fee applied before the grace period ends

Fee on day 3 of a 5-day grace period

Local-law violation; disputes

Ledger rule tests with jurisdiction fixtures

Maintenance request lost between portal and vendor system

Resident waits; nothing dispatched

Satisfaction, safety

Workflow state tests across the integration

Referral reward paid to the wrong resident

Wrong account credited

Financial loss; trust

Attribution tests on referral chains

Prospect–resident messaging leaks personal data

Phone number or unit visible

Privacy incident

PII masking assertions on message rendering and logs

In practice — Rentgrata

Rentgrata lets prospective renters message current residents of a building before they move in, and rewards residents for referrals. Since 2020 one manual and one automation engineer have built 9,000+ test cases to 90% coverage and logged 2,500+ bugs, 300+ of them critical, with WebdriverIO, Appium and JMeter in the pipeline. The lesson: a marketplace of conversations between strangers is a privacy product first, so identity, messaging and reward flows get the same rigour as payments. Read the Rentgrata case study.

Stage 6 — Operating the asset: hotels, facilities and storage

After the sale or the lease, the building runs on software, and property operations testing has its own catalogue – shorter here, because each system is a discipline of its own.

Hotel property management systems

Hotel quality assurance software testing centres on one failure: the property management system (PMS), the channel manager and the online travel agencies disagreeing about how many rooms are left.

Defect
Shows up as
Impact
Test that catches it

Channel-manager sync lag

Room sold twice across two OTAs

Overbooking; walked guest

Sync-loop latency test with SLA threshold

Rate-plan restriction ignored by one channel

Minimum-stay rule bypassed

Revenue leakage

Restriction matrix asserted per channel

Night audit miscalculates city or occupancy tax

Wrong folio totals

Tax exposure

Golden folio calculations per jurisdiction

Folio split error on group bookings

Charges on the wrong guest

Disputes at checkout

Split and transfer scenarios in regression

Check-in date time zone for multi-property chains

Arrival on the wrong day in reports

Operational confusion

Time-zone fixtures per property

Overbooking simulations, night audit and multi-property setups are the core of our hotel property management software testing work.

Facility management and building systems

A facility management software test has to include the building: sensor data, maintenance state machines and technicians working offline.

Defect
Shows up as
Impact
Test that catches it

Gaps in BMS/IoT sensor data

False alarms, or missed faults

Downtime; safety

Sensor-data simulation with gaps, noise and drift

Preventive-maintenance schedule drifts after a state-machine change

Tasks skipped or duplicated

Compliance; asset life

Golden schedules diffed per release

Work order lost between dispatch and the offline technician app

Job never done, system says dispatched

Tenant complaint

Offline sync and conflict scenarios on real devices

Multi-tenant permission leak across buildings

One client sees another's assets

Privacy incident

Tenant isolation matrix

Energy report wrong after tariff change

Mis-billed utilities

Financial dispute

Reference calculations per tariff version

Integration with building systems, asset hierarchies and multi-tenant isolation is described on our facility management software testing page.

Self-storage platforms

Defect
Shows up as
Impact
Test that catches it

Gate code not revoked after move-out or non-payment

Former tenant retains access

Security; liability

Access-revocation tests on every lifecycle event

Auto-pay retry double-charges

Duplicate rent

Refund; churn

Idempotency on payment retries

Unit availability oversold online vs. kiosk

Two tenants, one unit

Operational failure

Concurrency tests across channels

Lien and auction timeline computed wrongly against state law

Notice sent too early

Legal exposure

Timeline fixtures per state

Our scope for storage operators' booking, access and billing systems is on the self-storage software testing page.

Security testing for real estate software, automated

Security testing for real estate software is automated the same way as for any web platform, with one extra priority: the wire instructions at closing are the most targeted document in the industry. Business email compromise – the category that includes real estate wire fraud – cost victims $2.77 billion in 2024 according to the FBI's IC3 report, second only to investment fraud in losses.

What to automate:

  • Static analysis and dependency scanning in the build pipeline for your own code and its libraries.

  • Scheduled dynamic scanning of every externally reachable surface: consumer portals, agent apps, the API, and the webhook receivers for e-signature and payment providers.

  • Authorisation tests for every party × document and role × data pair – the matrices from Stages 3 and 4, run on every release.

  • Webhook signature verification for e-signature and payment callbacks, with tests that reject unsigned or replayed events.

  • Session and authentication tests on agent and client logins: MFA enforcement, session expiry, device changes.

  • PII detection in logs and masking in every non-production environment.

  • Wire-instruction handling as a product requirement: instructions are never sent by email, are displayed only inside authenticated sessions, and any change triggers out-of-band verification – and those rules are asserted by tests, not policy.

The web-application fundamentals – injection, broken authentication, misconfiguration – and how to put them in a pipeline are covered in our guide to security testing for web applications.

Automation for real estate platforms: what the experts automate first

Real estate software automation experts start with the suites that protect data licences, legal compliance and money – not with the screens.

1.

MLS contract tests and freshness monitors, per MLS, on a schedule against the live feed.

2.

Fair Housing rule tests on search inputs, ranking features and generated copy, as blocking checks.

3.

The access matrices: role × data for leads and contacts; party × document for transactions.

4.

The disclosure-form version matrix by state, transaction type and date, with the TRID timers.

5.

Idempotency on payments and e-signature, including webhook replay.

6.

Ledger reconciliation for rentals; channel-sync latency for hotels.

7.

UI smoke last, as a thin layer across web and mobile.

What stays manual: onboarding a new MLS (its rules need a human read), reviewing generated listing copy, and the agent's mobile workflow on a showing.

Tooling by layer: Postman or REST Assured for feed and partner contracts; pytest or TestNG parameterised from the form-version and commission matrices; Cypress or Playwright for web, Appium for mobile; Locust or k6 for the listing-launch and open-house spikes. Compass runs on Cypress, Supertest and Locust in CircleCI; Rentgrata on WebdriverIO, Appium and JMeter – the choice follows the team's stack, the priorities above do not change.

Real estate software testing checklist

Feeds

  • Every MLS has contract tests on recorded payloads and a freshness monitor with an SLA.

  • Status mapping, attribution and display rules are asserted per MLS.

  • Feed-to-database counts and statuses reconcile daily.

Listings and search

  • Fair Housing rule tests block any discriminatory filter, ranking factor or generated copy.

  • Geo boundaries, alert latency and AVM disclaimers are tested.

  • The search path passes WCAG 2.2 AA and mobile performance budgets.

Leads and CRM

  • The role × data access matrix is asserted for agent, team, office and brokerage.

  • Lead routing handles departed agents and territory changes; leads are de-duplicated across sources.

  • Consent is asserted before every SMS and email.

Transactions

  • Form versions are tested by state × type × date; TRID timers use business-day and holiday fixtures.

  • E-signature envelopes are asserted for signer list and order; webhooks verify signatures.

  • Payments are idempotent; commissions match golden calculations.

  • Party × document access is asserted for every pair.

Rentals

  • Screening flows issue adverse-action notices; application fees are idempotent.

  • Lease pro-rata, grace periods and late fees are tested per jurisdiction.

  • Messaging masks PII; referral attribution reconciles.

Operations

  • Hotel channel sync latency, rate restrictions and night audit are tested.

  • Facility sensor gaps, maintenance schedules and offline work orders are covered.

  • Storage access revocation and lien timelines are tested per state.

Security

  • SAST, dependency scanning and scheduled DAST cover all surfaces.

  • Wire-instruction handling rules are asserted by tests.

Conclusion

Three defects in this catalogue cost more than all the others combined: a stale MLS status shown as Active, a party seeing a document they should not, and a disclosure window missed by a timer.

Each is invisible on a screen that looks correct, and each is caught by a test that runs on data and rules rather than on pixels – a contract and freshness monitor, an access matrix, a business-day fixture. Keep the catalogue as a living document: every production incident is a new row.

If you would like to see it applied to your platform, talk to us about a free proof of concept on one stage – an MLS feed reconciliation against what your site displays, or a party-access matrix on your transaction module.

Book a strategic QA consultation

FAQ

What is real estate software testing?

Real estate software testing verifies the systems that carry a property from listing to closing and beyond — MLS and IDX feeds, search and listing display, CRM and lead routing, showing and transaction management, e-signature and payments, rental and resident platforms, and property operations — with emphasis on data freshness, Fair Housing and disclosure compliance, multi-party access control and payment integrity.

How do you test MLS and IDX integrations?

With contract tests per MLS built from recorded RESO Web API payloads, run on a schedule against the live feed; freshness monitors with an SLA per source; status-mapping and display-rule fixtures that assert compliance with each MLS's licence; de-duplication fixtures for overlapping markets; and a daily reconciliation of counts, statuses and prices between the feed and what is displayed.

How do you test for Fair Housing compliance in search and advertising features?

Maintain a list of prohibited terms and proxy features — attributes that correlate with protected classes — and run blocking tests on every search input, ranking factor and piece of generated copy against it; test that equivalent searches return equivalent results regardless of user attributes; and review any new ranking or personalisation feature against the list before release.

How to automate security testing for real estate?

Put static analysis and dependency scanning in the build; schedule dynamic scans of portals, APIs and webhook receivers; run the party × document and role × data authorisation matrices on every release; verify webhook signatures and reject replays; test MFA and session handling; detect PII in logs; and assert the wire-instruction handling rules — never by email, only in authenticated sessions, out-of-band verification on change — as product tests.

Can you suggest partners who provide quality assurance for high-stakes real estate applications like underwriting and fraud detection?

Look for a partner that tests the regulated logic rather than the screens — Fair Housing constraints, TRID timers, FCRA screening flows, party-based access, payment idempotency — and can show real estate clients with numbers. DeviQA has tested real estate products since 2016 for Compass, Century 21, Rentgrata, Realforce, Sellers Shield and DiversyFund, and offers a free proof of concept on one flow before any commitment.

How do you test a hotel PMS for overbooking?

Automate the loop from OTA booking to PMS to availability update on every other channel and measure its latency against the SLA; fire parallel bookings across channels at the last room and assert exactly one succeeds; cover group bookings, rate-plan restrictions and multi-property setups; and run the suite at peak booking volume, not at idle.

Mykhailo Ralduhin

About the author

Mykhailo Ralduhin

Senior QA engineer

Mykhailo Ralduhin is a Senior QA Engineer at DeviQA, specializing in building stable, well-structured testing processes for complex software products.