Agreed. Keep financing as unknown unless confirmed. You can still compare completed deals by apparent marketing period without pretending to know why a particular buyer took longer.
The key distinction is buyer delay versus seller delay. Financing may lengthen a transaction after agreement, while late price cuts extend the period before agreement. Combining them makes the result hard to interpret.
For the energy point, compare similar-age homes within the same neighbourhood and condition category. If the gap disappears, energy performance was probably standing in for those other features.
What exactly counts as an energy-performance outlier here? If that information is inconsistently included in advertisements, missing entries could be mistaken for poor performance.
That is important. “Not stated” needs its own category; it should not be grouped with a low assessment. Otherwise the strongest-looking pattern may just reflect better listing descriptions.
Also note whether energy information appeared from day one or was added later. A seller might improve the advert after weak interest, reversing the apparent direction of cause and effect.
I would avoid making energy the headline until completed deals are matched. Price, micro-location and condition offer simpler explanations for old stock in such a broad Dubai sample.
There is still value in flagging it as a hypothesis. Just test it against comparable properties rather than treating the current online outliers as proof.
Could seasonality be influencing the snapshot? Without several observation dates, “this month” may capture a temporary mix of fresh listings and properties carried over from earlier periods.
Yes, save the current list and repeat the same process later. The important movement is which homes sell, which are withdrawn, which cut price and which simply reappear.
I would contact agents on a small number of disappeared listings and ask only whether they sold, were withdrawn or moved to another agent. Even incomplete answers could prevent obvious classification errors.
Be careful with verbal status updates, though. “No longer available” does not necessarily mean completed. Keep unverified exits separate rather than forcing them into sold or withdrawn.
Absolutely. Three labels are insufficient if the evidence is patchy. Add an unknown-exit category and revisit it if a completed transaction can later be matched.
Another useful split is vacant versus occupied, where clearly stated. Access for viewings and presentation can affect time online without saying anything about the underlying value.
Plot and built area should be captured too if available. Detached homes inside one price range may not be comparable when one offers much more land but needs extensive work.
At this point, 87 days looks suitable as a description of Omar’s current online sample, not an estimate of how long a newly listed home will take to sell.
That distinction should appear in any conclusion: “age of active advertisements” rather than “time to find a buyer.” The latter requires evidence of an agreed or completed sale.
And even completed transactions create survivor bias in the other direction if withdrawn homes vanish from the analysis. Showing every exit type is the cleanest way to frame it.
Has anyone considered seller price changes before the visible advert? If a property was quietly marketed higher elsewhere, the apparent first asking price may already reflect a reduction.
Probably impossible to capture consistently. I would document only observable listing history and state the limitation, rather than expanding the project into unverifiable off-market activity.