AboutInsightsToolsSign in
All insightsProperty search
Oct 14, 202621 min

Why the Same Home Looks Different on Zillow, Redfin and Realtor.com

The address is unchanged. The search that reveals it, the clock beside it and the person behind the contact button may not be.

Begin the story
One ivory house viewed through three transparent glass planes, a visual metaphor for different information lenses.

Before you press the button

Imagine a buyer with a house open on three browser tabs.

The kitchen is the same. So is the stretch of lawn beside the driveway, the upstairs window and the price. But one screen makes the listing look new. Another gives it a longer history. On a third, the buyer cannot find it in the neighborhood search at all.

There is a practical decision waiting behind this small discrepancy. Should they rearrange tomorrow to see the house? Is there time to compare it with another? Would an offer below the asking price be worth discussing, or is the home already drawing offers?

The buyer moves toward a contact button. What they want to ask is simple: Is this home still available?

What the button will do is less obvious. Will it reach the person selling this house, a buyer's agent who might help them buy it, or a service that connects them to an agent? And what does available mean if the seller has already accepted an offer but the transaction has not closed?

The house has not become three houses. The screens are exposing different pieces of a transaction, sometimes under labels that appear to answer the same question.

That distinction is worth understanding before it costs someone an afternoon, a negotiating position or a misplaced confidence in the person on the other end of the call.

The Consumer Policy Center's October 13, 2026 comparison of Zillow, Realtor.com, Redfin and Homes.com brings these questions into focus. The consequential differences are often less obvious than a missing photograph or a different price.

The exception is not the whole market

Stephen Brobeck's study examined 500 agent-marketed properties in 21 communities across ten metropolitan areas during August and September 2026. This was a selected case study, not a representative national sample; the main comparison excluded builders, auctions, for-sale-by-owner homes and properties already contingent or pending.

Most listings agreed. About three quarters appeared on all four portals. The revealing exception: almost every property absent from a portal's community search still had its own page there. The full study also found disagreements in listing ages, including one Volente, Texas, example shown as 27 hours on Redfin and 77 days elsewhere. It does not provide the record history needed to explain that case.

Treat this as a closer look at selected communities, not a national portal ranking. Before judging a discrepant clock, we also need to establish what it counts.

We can ask a more useful question: What is the buyer actually comparing?

A home is a physical place. A listing is a record about an attempt to sell it. A search result is the answer to a particular query. A contact button is a route to a person or service.

Each layer can change without the others changing. A new listing record does not rebuild the house. A change to the search boundary does not remove it from the neighborhood. A new agent connection does not establish that the buyer has acquired representation.

The interesting discrepancies occur where we mistake one layer for another.

One address.
Four different objects.

The discrepancies become clearer when we separate what changed.

The physical place. Changing a record or a search boundary does not rebuild it.

Conceptual layers, not separate properties or a ranking of websites.

Consider what a buyer means when they search for a neighborhood.

They might mean the streets near a school, the walkable area around a station, or the place residents recognize by a familiar name. A search system needs something it can calculate: perhaps a boundary, a geographic label or a set of coordinates. Those definitions need not coincide.

This creates a peculiar experience. A home can be close enough to count in a person's idea of the neighborhood, yet outside the shape used by a website. It can belong to a property category the buyer did not intend to exclude. It can fall outside a price limit entered before the buyer understood the local market. These are possible mechanisms, not diagnoses of the study's individual omissions.

The missing result can feel more definitive than it is. The homes that passed the search are visible; the reasons others did not are harder to see. A tidy map gives the resulting selection the appearance of a complete market.

Finding a direct address page settles one narrow question: the website has a page for this property. It does not settle whether the home is actively for sale. Property pages can outlive sale attempts. The current status and history still need checking.

There are separate questions to resolve: inclusion in this search, presence on this website and availability for purchase. An empty result settles the first. A direct property page helps with the second. The third requires current transaction information.

For a buyer who knows the address, the practical response is modest. Search the address directly. Recheck the map edge, the property type and the status filters. If the discrepancy survives, ask someone with access to the relevant local listing record to explain it.

Once the omission has an explanation, the buyer can get on with evaluating the house.

Four screens are not necessarily four witnesses

The local multiple listing service, or MLS, matters to this story because it sits upstream of much public home-search information. NAR's consumer guidance on listing distribution describes MLSs as systems through which participating brokers share listings, with information also distributed to public websites.

Suppose several websites receive the same asking price from a common listing source. If all display it, their agreement is useful. It suggests that the price has traveled consistently through those routes.

But the websites have not independently measured what the house is worth. They may be repeating the same seller's asking price. Four appearances do not turn a request into four valuations.

The distinction applies to less obvious fields, too. If an upstream record contains an error, agreement downstream can reproduce it. If one route updates sooner, disagreement can briefly reflect timing rather than a different underlying home.

That shared-source problem explains why counting the agreeing websites can be a poor way to resolve a disputed fact. It does not establish an error in any particular listing.

Instead, trace the fact back to what it purports to describe. Is the number a seller's current request, an assessor's record, a past transaction, an estimated value or an elapsed-time calculation? When was the underlying event? When did the page last acquire the information?

The Real Estate Standards Organization's data dictionary makes a distinction that is easy to lose on a consumer screen: a date supplied for a transaction event need not be the timestamp when a system recorded the change. Its definitions also leave days-on-market calculations to MLS business rules. A standardized field name is not a guarantee of one nationwide counting method.

For the buyer, the consequence is simple. Agreement can establish consistency without establishing independence. A discrepancy can reveal a useful question without identifying which website is at fault.

Four screens.
Not always four witnesses.

Agreement may repeat a common source. It does not turn an asking price into four valuations.

Listing record

One possible shared origin

  • Zillow
  • Redfin
  • Realtor.com
  • Homes.com

An event is not the same as an observation.

The event date and the time a website receives the change can differ.

These equal-weight links illustrate shared provenance, not a measured feed architecture or completeness score.

Possible audiences.
Not a universal sequence.

Brokerage audienceIn this scenario
MLS participantsIn this scenario
Public sites via shared feedsIn this scenario

Public websites can receive the listing through distribution arrangements. This diagram does not promise every portal receives every listing at the same instant.

Explanatory schematic, not a live feed or a guarantee of universal coverage. NAR's distribution guidance distinguishes office exclusives and delayed public marketing. Local rules vary.

A house can have three ages

Build an illustrative listing history using our own counting convention, rather than an official MLS rule.

A house spends 42 days actively offered at $640,000. The seller takes it off the market for 21 days. It returns under a new listing record at $610,000. Nine active days later, a buyer finds it.

How old is the listing?

Under the new record, it is nine days old. Add the two active periods, and the house has accumulated 51 days of active marketing. Count the whole interval from the beginning of the first attempt, including the interruption, and 72 days have elapsed.

All three numbers answer a defensible question. They do not answer the same one.

The nine-day figure describes the current record. The 51-day figure describes our sum of active periods. The 72-day figure describes elapsed time since the initial attempt. A local MLS's cumulative measure might apply different rules about the interruption. None of these figures, on its own, tells us how urgently the seller wants to move.

The $30,000 price reduction adds another detail, approximately 4.7% of the initial asking price. A buyer who sees only the new record could miss the relationship between the two prices. But even the complete sequence cannot tell us why the property left the market. Repairs, a family decision, a failed plan or a change in strategy would imply different things. The timeline does not supply the motive.

Now return to the buyer deciding whether to rush.

Nine days might suggest a fresh contest. Seventy-two days might suggest patience. Either impression could mislead if it becomes a substitute for asking about current interest, offers and the seller's position.

Instead of asking only, "How many days?", ask: "Since which event, under which record, and with which interruptions?"

That question removes some drama from the badge. It adds something much more valuable to the decision.

One home.
Three clocks.

Change the interruption.
Watch the counts separate.

First listing42 active daysOff market21 off-market daysNew record9 active daysDay 0Observed now: day 72
  1. First listing42 active days
  2. Off-market pause21 off-market days
  3. New record9 active days
Current record9 days

The current active interval.

Active-day sum51 days

42 + 9. Counts only active intervals.

Elapsed interval72 days

42 + 21 + 9. Includes the pause.

Change the history

Only elapsed time changes when you change the pause alone.

Adjust the active intervals

Illustrative asking price$640,000 to $610,000Approximately 4.7% lower. Not a completed sale.

CSV
Illustrative history. These are our counting conventions, not figures assigned to specific portals. Our active-day sum is not a universal MLS cumulative-DOM formula. Mobile event positions are not to scale. RESO's definitions leave local DOM calculations to MLS business rules.

What the portal clocks actually promise

The portals' own published explanations show why a single label can be a poor shortcut.

Zillow's help page on Time on Zillow, last updated in December 2023, distinguishes feed types. For listings received through an IDX feed, a form of broker listing-data distribution, its Time on Zillow calculation uses the MLS's on-market timestamp. Its separate rules for non-IDX listings include circumstances in which a clock resets, such as an extended off-market interval. Those non-IDX reset rules should not be applied to every Zillow listing.

Redfin's help page on Time on Redfin, last updated in September 2022, describes the property's advertising under the current MLS number. It distinguishes that from a cumulative days-on-market measure, which can incorporate earlier activity under local rules.

It would therefore be a mistake to decide that the largest displayed age must always be the true one. A longer count might include a period another count intentionally excludes. A shorter one might refer to a new record rather than deny an earlier sale attempt.

A clock can also be wrong or stale. Different definitions are not an excuse to accept every number. They are a reason to establish the intended measure before testing its accuracy.

Think of the work involved as reconciliation, not voting. Identify the record. Look at the history. Separate active periods from elapsed time. Ask what the local system counts when a listing is withdrawn, expires or returns.

For negotiations, even a perfectly reconciled clock remains only one input. A house can wait because it is overpriced, difficult to finance or difficult to inspect. It can wait because the owner is in no hurry. It can return after a change that genuinely makes the offering different.

Time on a website is information about marketing. It is not a meter attached to the seller's willingness to compromise.

The label needs
a definition.

Time on Zillow

For IDX listings, the published explanation uses the MLS's on-market timestamp. Separate non-IDX rules must not be generalized to every listing.

Published help, Dec 2023

Time on Redfin

The published explanation describes advertising under the current MLS number and distinguishes cumulative market time.

Published help, Sep 2022
Definitions checked October 14, 2026. These explain published measures, not an independent test of every live implementation. The largest number is not automatically the correct one.

New to the public is not necessarily new to the market

In its May 18, 2026 launch announcement, Redfin describes Early Access as a route for sellers to test interest before a broader launch, including pre-market homes from Compass-affiliated brokerages. The announcement says that Early Access listings do not accrue days on market or publicly display price-drop history.

A home might appear newly available to a broad audience after an earlier, more limited period of exposure. If the public clock begins at the broad launch, it measures that stage. It does not automatically describe everything that happened before it.

The same principle applies to a price history whose visible window begins late. No displayed reduction during that window does not establish that the seller never tested another figure elsewhere.

A careful buyer can ask when marketing began, to whom it was available and whether the price changed during earlier exposure. The answer may be entirely ordinary. The point is to understand the beginning of the story, rather than assign a motive to information the screen does not contain.

There is a seller's perspective worth keeping in view. Wide exposure can bring opportunities, but publicity can also carry costs for someone managing security, privacy, tenants or a sensitive move. Testing a price before a larger launch may be an intentional choice, made with trade-offs.

The buyer, meanwhile, needs to evaluate the offering rather than the promise of being early. The label alone supplies no evidence of a bargain or of a seller's motive.

For buyers, the useful curiosity is specific: Is this the first opportunity for anyone to consider the home, or the first opportunity for people using this search?

New to this audience.
Not necessarily untried.

The visible history may begin after an earlier period of exposure.

Earlier exposureAsk what happened before the window.
Broad public visibilityA public clock may describe this stage.
Conceptual exposure windows, not measured durations. Redfin's May 18, 2026 announcement states the published Early Access treatment of DOM and public price-drop history. This is not evidence of any seller's motive.

The audience changes before the house changes

A listing can pass through several audiences: one brokerage's clients, other participating agents, a public portal, then a wider network of websites. Those stages are not interchangeable.

NAR's guidance on alternative listing options distinguishes an office-exclusive listing from delayed public marketing. In the former, distribution is restricted rather than shared with other MLS participants or publicly advertised. In the latter, other MLS participants can have access while distribution through IDX and syndication to public websites is delayed. The listing broker and seller can still advertise during that period. Delayed distribution is not a ban on all public exposure, and local implementation matters.

The organization's May 2025 policy FAQ explicitly leaves decisions about days-on-market treatment during delayed marketing to local MLS rules. This is industry policy with local implementation, not a universal federal rule governing every sale.

The relationships between portals can change as well. Zillow and Realtor.com's September 17, 2026 joint announcement said Preview listings were live on both sites. A comparison made during an earlier period cannot be treated as a permanent map of which brand has which early inventory.

This is why the geography of a search and the distribution of a listing should be kept separate. A home missing because a map boundary excludes it is a different problem from a home not distributed to that audience yet. Both may produce an empty search. They call for different explanations.

For a seller, the central question is not how many logos appear on a marketing plan. It is which potential buyers can actually encounter the home, when they can encounter it and what the seller gives up in exchange for the chosen sequence.

For a buyer, early visibility becomes useful through an actual opportunity. Can the property be shown? Will the seller entertain an offer? What information is available to evaluate it? Learning the address is the beginning; inspecting and securing the home require further steps.

This difference between awareness and action is where the next label can become misleading.

Under contract is not sold. It is not freely available, either

The word pending sits awkwardly between two everyday ideas: a house for sale and a house already sold.

Zillow's explanation of contingent and pending listings distinguishes the stages and describes variants that can allow continued showings or backup offers. The precise status and opportunities depend on the transaction and local practices. Pending generally describes a transaction moving toward completion, with an existing contractual position to respect.

From the buyer's side, three questions are often squeezed into one word. Can I see the house? Can I submit an offer? Can this seller accept my offer in the position I imagine?

The answers can differ.

A showing might be possible without a realistic immediate path to purchase. A backup offer might create a position if the existing transaction fails, not a right to displace it. A property visible on a website might be there because its transaction is still in progress rather than because the seller is seeking another primary buyer.

National statistics about contracts falling through cannot answer whether this particular transaction will fail. That requires information about a specific contract, its conditions and the parties involved, much of which may not be public.

The sensible question to an agent is therefore not simply whether the listing can be viewed online. Ask its current contractual status and what, if any, opportunity exists for a new buyer. Does the seller permit a showing? Are backup offers invited? Is the page's status current?

It is possible to be curious about a house without mistaking curiosity for access. That makes the buyer's next action more precise, and it gives the person answering a better question than a badge can.

Available for what?

Does the website have a page? A page may persist after a sale attempt.

Separate questions, not guaranteed stages or contract-failure probabilities. Status codes, seller instructions and the actual contract determine what a buyer can do.

The person behind the button

A contact button often appears beside a photograph of a particular house. The placement encourages a natural assumption: this is how to reach the person responsible for selling that house.

Sometimes it is. Sometimes the service is designed to connect a shopper with a buyer's agent.

Zillow's Premier Agent materials describe advertising and connection products for participating agents. Realtor.com's Connections Plus delivers buyer leads, while ReadyConnect Concierge offers live referral connections.

Homes.com's published agent products emphasize routing inquiries to the listing agent. The company also sells membership and promotion that increase listing visibility. Direct routing and paid exposure can coexist.

Even within a portal, the route can vary by product. The September Preview announcement describes a choice between contacting a listing agent and connecting with a buyer's agent. The useful thing to inspect is the particular offer beside the button, not just the brand above it.

None of these routes is inherently the right one for every buyer.

Someone who needs confirmation of a showing instruction may want the listing agent. Someone who needs advice on the offer, inspections and financing may want an agent representing their own interests. A connection to a buyer's agent can be useful, even if it is not what the shopper initially expected.

The problem is ambiguity about the role, not the existence of a connection business.

NAR's guidance on agency and non-agency relationships explains that seller and buyer representation are different relationships, with the available arrangements governed by state law. Speaking to the seller's agent does not by itself mean that person represents the buyer.

Before sharing a confidential negotiating limit, it is worth establishing whom the person represents. Are they the listing agent? An agent offering to work for the buyer? A referral service? What agreements and fees would apply before they provide the requested services?

The answer is not supplied by the fact that a button sat beside the kitchen photograph.

The photograph is the same.
The person answering may not be.

A particular home
Contact
A useful question

Are you the listing agent, and what showing or offer opportunities are available?

A direct route to the seller's agent can help establish current instructions. It does not by itself give the buyer representation.

Possible routes, not an audit of every button on any portal. Published products include different choices. State law and actual agreements control representation, as NAR's agency guidance explains.

The order of the homes is part of the experience

A buyer rarely gives every listing equal attention. The first few results receive a different kind of opportunity from a home several screens down. An appealing badge can prompt a click that an ordinary entry would not receive.

Paid promotion makes this distinction concrete. Homes.com's published agent page offers increased visibility and placement. A business model can route inquiries to the listing agent while also charging for the chance to be seen sooner. The two features address different parts of the experience.

This does not establish that any particular search order is improperly influenced. Sort settings, location, property details, personalization and commercial placement need to be considered in the actual interface. The relevant point is narrower: being easy to find and being the best fit are different properties.

A search system cannot know a buyer's complete future. It may know the price filter. It does not necessarily know the childcare arrangement, the partner's commute, the aging parent who might need a first-floor room or the savings the household must retain after closing.

An early result can be useful. It should not quietly become a recommendation merely through repetition.

For the seller, placement is a way to attract attention. For the buyer, attention must eventually become evaluation. That requires leaving the sequence of attractive cards and returning to the household's constraints.

Is the property within the budget once recurring ownership costs are included? Does it satisfy the requirements that cannot be repaired later, such as location? What information is still missing before an offer would be responsible?

The portal can help discover a candidate. It cannot make the candidate suitable by displaying it often.

Easy to find.
Still yours to evaluate.

Attention

Placement, presentation, promotion.

A route to discovery.

Suitability

Budget, location, condition, household.

A separate decision about the home.

Qualitative distinction, not an estimate of promotion's impact. Published agent products describe paid visibility; visibility is not a measure of suitability.

A better conversation about one address

The remedy for imperfect information is not to make every buyer a forensic data analyst. It is to ask a few better questions at the points where a misunderstanding could change the decision.

Start with the identity. Are the pages describing the same unit or parcel, and the same sale attempt? An address is a useful starting point, but units and repeated records deserve attention. Similar photographs should not substitute for establishing the property.

Then establish the present. What is the current asking price and contractual status? When was that information confirmed? A detailed historical page is useful, but an old fact does not become current because it is easy to find.

Reconcile the history only as far as the decision requires. When did active marketing begin? Did the property leave and return? Does the displayed age refer to the current listing number, active days or a longer interval? What changed between attempts?

Finally, establish the action and the relationship. Is a showing available? Is a new offer invited, or only a backup? Who is answering, whose interests do they represent and what commitment would be required to proceed?

NAR's guidance on open houses and written agreements distinguishes an initial conversation or an independently attended open house from working with a covered agent on a tour, where its rules require a written agreement. State law and the actual service arrangement matter. Clicking for information should not be treated as equivalent to entering representation.

These are small questions, but they interrupt large assumptions. Their purpose is to turn the page's information into a clear next step, including a decision to leave the house alone.

They also allow the buyer to stop researching when the uncertainty that matters has been resolved. More tabs are not always more understanding.

A better conversation
about one address.

01The property

Are the pages describing the same unit or parcel, and the same sale attempt?

02The present

What is the current asking price and contractual status? When was that information confirmed?

03The history

When did active marketing begin? Did the home leave and return? Which interval does the displayed age describe?

04The relationship

Who is answering, whose interests do they represent, and what agreement or fees would apply?

An open navy door leading into a quiet sunlit ivory interior.

What an intelligent property system must remember

The issue extends beyond shoppers. Any system that uses listing information to describe a market, estimate liquidity or help a person decide needs to preserve these distinctions.

Suppose a research process repeatedly searches one portal for homes in a neighborhood. A property disappears from the results. Has it sold? Was it withdrawn? Did it cross a filter boundary after a price change? Did the query or the distribution route change?

If the process records only presence or absence, those explanations can become indistinguishable. A search observation has been mistaken for a transaction event.

Now suppose the same home returns under a new record. If a model treats that as a completely new property, it may lose the earlier marketing history. If it automatically joins every similar address, it may incorrectly combine units or separate offerings. The repair requires careful identity matching, not merely confidence in a prediction.

Price histories need comparable care. An asking price at one stage of marketing is not the amount received at closing. An unavailable earlier price is not evidence that no earlier price existed. A blank field must remain unknown unless another source establishes what happened.

For a well-designed system, the questions are quite concrete. Which property is this? Which listing episode? Which event does the record describe? When did the event occur, and when was it observed? Through which source and audience was it visible? What remains uncertain?

This is a proposed discipline for interpreting property data, not a claim that Oftu has measured or resolved every portal discrepancy. Its value is that it keeps the record connected to the question a user is asking.

AI can help reconcile identities, identify conflicts and explain competing definitions. It cannot honestly recover a hidden marketing history simply by producing a fluent answer. A simulation can explore several possible histories, but those branches must remain assumptions rather than be relabeled as observed facts.

The more polished the answer becomes, the more important it is to preserve the boundary between what happened and what the system inferred.

A search observation
is not a sale event.

The observation

A home disappears from results.

The questions
  • Did its sale status change?
  • Did a price or type filter exclude it?
  • Did the boundary or distribution change?
A proposed interpretation discipline, not a verified Oftu capability or a diagnosis of an actual listing. Preserve the source, event date, observation time and uncertainty.

The question after the photograph

Return to the buyer with three tabs open.

The kitchen still looks promising. The stretch of lawn has not changed. But the buyer no longer expects the three screens to provide three independent versions of the truth.

The discrepancies have become specific questions. Search visibility, marketing history, contractual status and representation can each be checked on their own terms.

Instead of letting an unexplained nine-day badge decide the pace, the buyer asks when this sale attempt began, what happened before it and what opportunity exists now. Before discussing the highest offer they could afford, they ask whom the agent represents.

The answer may be reassuring. The discrepancy may be ordinary. The home may still be worth seeing tomorrow.

What changes is the source of the urgency. It comes from confirmed circumstances, not a number whose beginning was never established.

Property portals are valuable precisely because they make distant, complicated markets easier to see. They bring photographs, addresses and possibilities into reach. The task is not to reject that convenience. It is to avoid asking the convenience to do work it has not promised to do.

A home search can show a door. Before we arrange our lives around it, we should know which door is open, what is behind it and who is inviting us through.

A home search
can show a door.

Sources and scope

Research checked October 14, 2026. The study's release date and observation period are distinguished above. We have not conducted an independent address-by-address portal audit.

Portal help pages and company product descriptions explain their published rules and offerings. They are not independent proof of every live implementation or of claimed performance benefits. Older guidance is dated in the relevant section. MLS rules, contractual opportunities and agency arrangements vary locally.

The opening buyer and the $640,000-to-$610,000 listing history are illustrative. The latter uses our explicitly stated counting convention, not a universal cumulative-days-on-market formula. No estimate of a particular seller's motivation, a contract's failure probability or a portal's national superiority is implied.

Read as MarkdownIllustrative timeline CSV
More insights