Turn Portal Listing Discrepancies Into Controlled Corrections

A property can be correct in the estate agency's own system and still be misleading on a portal. The asking price may have changed but the old figure remains in a cached search card; a new floor plan may have been accepted by the feed but the public page still displays the previous image. An agency needs a reliable way to identify these discrepancies, decide which ones matter most and prove that a correction has reached the consumer-facing destination. A portal discrepancy log provides that chain of evidence.

The need for care is not just technical. Section 227 of the Digital Markets, Competition and Consumers Act 2024 addresses the omission of material information a consumer needs for an informed transactional decision. The relevant consumer-protection regime came into force in April 2025. What counts as material depends on the property and circumstances. A discrepancy register should help the agency find and resolve inaccurate or confusing claims without presenting a software feed's status as a substitute for responsible review.

Establish which property information is actually approved

A discrepancy cannot be determined solely by comparing two public pages; both may be obsolete. Identify the agency's authoritative current property record, including approved asking price, source-verified material details, approved photographs, floor plans and instruction status. Each item should be associated with its effective version and approving person. That baseline should be available before staff attempt to correct a portal.

Different fields require different sources. The seller's current instruction may establish the asking price; the tenure description may require an appropriate property document; a photograph needs reliable property identity and usage rights. A source register should preserve those distinctions. If the internal record itself is disputed, fixing a portal to match it may spread the error. Put the information into verification first.

Include the exact portal and listing reference. An agency may market several properties with similar addresses, or one development with multiple units. A vague note saying 'wrong image on portal' is insufficient for someone else to reproduce the problem.

Capture the observed public mismatch with evidence

Record the destination URL or reference, date and local time of observation, device or view where useful, the affected field and what was actually visible. Preserve an appropriate screenshot or other evidence where the discrepancy is consequential, subject to rights and privacy. Distinguish the observed public content from the approved internal value. The register should enable a colleague to see the difference even if the portal changes after the issue is reported.

Do not speculate prematurely about cause. A stale photograph could result from a delayed feed, image caching, incorrect asset mapping or a change that was never submitted. The first step is to document the actual mismatch. Root-cause analysis can follow, using logs and vendor responses where available. This prevents staff from blaming a portal for an error that originated in their own system.

Pay attention to differences across contexts. A portal's listing detail may be current whilst a search-result thumbnail is stale. Mobile and desktop layouts may expose different fields. A successful internal preview does not demonstrate that every consumer-facing representation has been corrected.

Prioritise material claims and the customer's decision

Not every difference demands the same response. An incorrect price, tenure description, availability state or right of access may affect a person's decision to pursue a home. A minor typographical inconsistency may be less urgent, although still worth correcting. Define a practical priority rule that considers the significance of the claim, number of channels affected, whether people are actively relying on it and whether a correction can be made promptly.

The government's 2025–26 consultation on property listing material information makes clear that relevance depends on the individual circumstances rather than a single exhaustive checklist. The agency must decide which facts matter here, on evidence. A portal field being technically optional does not establish that omission of the corresponding information is harmless.

If a discrepancy could mislead an active buyer, consider whether a separate clarification is needed while the public correction is pending. Record that decision and its basis. A change made to the website should not be presented as completed across all channels without verification.

Assign one owner and identify the correction route

For every issue, name a person responsible for resolution. That person may coordinate an internal editor, feed administrator or portal support team, but should retain ownership until the public outcome is checked. Give the issue a state such as new, investigating, correction submitted, awaiting public verification, resolved or blocked. The wording matters because colleagues and managers need to distinguish work in progress from confirmed closure.

Record the action taken. Did staff change the agency's master record, resend an existing correct value, remove a superseded asset or request intervention from a portal? These are different routes with different risks. The correction record should point to evidence of submission and, if applicable, the vendor's acknowledgement.

Do not allow responsibility to become diffuse. An issue assigned jointly to 'marketing and IT' can remain unaddressed if neither team owns the final check. Handover between staff should identify the current owner and the last verified state, rather than restarting the investigation from scratch.

Verify the actual customer-facing result

Submission is not completion. After the ordinary publishing interval, revisit the public destination and compare it against the authoritative property record. Check that the intended value appears and that the fix did not create a different problem, such as a duplicated photograph or vanished floor plan. Where the result is not yet visible, keep the discrepancy open.

Timing may differ between feed acceptance and cache refresh. Record when the correction was requested and when its public effect was observed. A technical delay may be understandable, but the agency should not describe a stale material claim as corrected before it can verify that result. If the change is urgent, determine whether public marketing must be temporarily restricted or supplemented with direct information to affected customers.

Close only when there is adequate evidence. The final record should show the old public content, the approved current content, corrective action, actual verified destination and closure time. This avoids reports that overstate performance because staff count tickets submitted instead of discrepancies resolved.

Protect data and consider attached media rights

Portal workflows frequently carry photographs, seller-related details, floor plans and communication histories. Review which information is appropriate for public distribution and which belongs in private internal records. Avoid placing seller contact details or sensitive instructions in public fields by accident. The ICO's data-protection guidance provides relevant principles for personal information, while the agency needs separate checks for property accuracy and image permissions.

Image errors require their own evidence. An old or wrong-property photograph can misrepresent the offer even when all text fields are correct. Keep the source asset, photographer's licence, property identity and approval state linked together. A newly supplied photograph should not be assumed reusable across every portal or promotional format if rights limit the permitted use.

Version control also prevents leakage of drafts. A pending seller change, private price discussion or internal note may exist in an agency system without being authorised for public release. A feed integration should use approved publication fields, not every string available in the working database.

Learn from recurring discrepancies without inventing causes

A backlog of resolved issues can identify patterns. Are thumbnails commonly stale after image changes? Do particular structured fields disappear on one destination? Does a staff workflow repeatedly omit the final public check? Group issues by observed symptom first, then investigate technical causes. The same symptom may have more than one cause, so do not treat frequency as proof of a specific vendor defect.

Compare relevant periods and volumes. Ten discrepancies on a thousand changed listings are not directly comparable with ten on twenty, and a large portal inventory may naturally produce more tickets than a small one. Define denominators carefully, including what was actually checked. Review unresolved ageing and high-consequence issues separately from minor presentation tasks.

Use findings to improve checklists, field mapping, source-data control or supplier escalation. A better process may eliminate repeated corrections without purchasing an additional AI product. Staff should still be able to stop a publication when the controlling fact itself is uncertain.

Assess automation on verified outcomes and total cost

A conventional discrepancy dashboard can compare internal and external fields where reliable integrations exist. An AI assistant may assist with detecting semantic differences in longer descriptions, but it can confuse harmless wording changes with material contradictions or fail to notice an important omitted qualification. Human reviewers should inspect examples and determine whether the tool identifies the right property and claim.

Test five awkward cases: an outdated asking price, a correct master value with stale cache, wrong-property imagery, an unverified tenure claim and a corrected page with an old search thumbnail. Count actual correct closures and reviewer time, not merely alerts generated. A system that produces thousands of vague warnings without reliable prioritisation may create more work than it saves.

For illustration, assume a branch spends 40 hours monthly on finding and resolving portal mismatches. Better checks reduce initial detection and preparation by 14 hours, but add five hours of review, leaving nine potential hours of capacity before licensing and support. These numbers are hypothetical, not observed agency savings. The more important result may be fewer misleading public listings and a faster correction of material errors.

Research sources: Digital Markets, Competition and Consumers Act 2024, section 227; UK government consultation and response on material information in property listings; The Property Ombudsman Codes of Practice; ICO data-protection guidance. The discrepancy controls and example arithmetic are editorial, not statutory certification or comparative trial evidence.