Knowing the price at which a house was actually sold is not the same process as estimating its value before putting it on the market. The two exercises rely on distinct sources, with very different levels of reliability. The DVF database, notarial tools, and online simulators do not measure the same thing, and the discrepancies between their results can reach several tens of thousands of euros for the same property.
DVF, Patrim, and notarial databases: three sources of actual prices compared
Three public or semi-public tools allow access to the actual sale prices signed at the notary. Their scopes and limitations do not completely overlap.
See also : The boom of streaming platforms: how to access a wide range of TV channels for free?
| Source | Access | Available data | Main limitation |
|---|---|---|---|
| DVF (Demandes de Valeurs Foncières) | Free, online via data.gouv.fr or app.dvf.etalab.gouv.fr | Sale price, date, area, type of property, address | No details on the condition of the property or the work done |
| DVF+ (enhanced version via Etalab) | Free, integrated into some professional tools | DVF + nature of the work, specific discounts, VEFA | Variable coverage depending on the municipalities |
| Patrim (DGFiP) | Free, via the personal space impots.gouv.fr | Sale price, area, number of rooms, cadastral parcel | Limited to a certain number of queries per month |
Since 2023, DVF+ data includes information absent from the classic DVF database, such as the presence of sales in a future state of completion or discounts related to work done. For those looking to find out how to know the selling price of a house already sold in their neighborhood, DVF+ offers a more usable level of detail than the raw database.
The notarial databases (accessible via the departmental chamber websites) sometimes add qualitative elements that DVF does not mention: presence of a garden, floor, orientation. However, their consultation is often paid or reserved for professionals.
Recommended read : The price of an MSC cruise: everything you need to know
Online estimation and actual price: why the gaps widen
Online estimation simulators (MeilleursAgents, SeLoger, PAP) calculate a theoretical value based on averages per square meter and declared parameters. Their results can diverge significantly from the price actually obtained during the sale.
Three main factors explain these discrepancies.
- Simulators ignore the actual condition of the property. A DPE rated G or F leads to a discount that these tools do not model accurately. Since the reform of the enforceable DPE and the gradual ban on renting thermal sieves, poorly rated properties suffer an increased discount upon resale, especially in large metropolitan areas where buyers anticipate heavy renovation work.
- Reference data can sometimes be outdated. A simulator relying on transactions that are twelve months old in a declining market overestimates the price. Conversely, in a tight market, it may underestimate it.
- The human factor remains absent. The quality of a view, sound insulation, proximity to a nuisance, or the configuration of a plot are not captured by an algorithm fed solely by surface area and postal code.
A simulator provides an order of magnitude. Only cross-referencing with actual prices from DVF allows for verifying the consistency of this estimate.
Duration of sale: a price indicator that databases do not show
Public databases record the final price, never the sales journey. The time between the posting of an ad and the signing of the compromise constitutes a strong signal regarding the relevance of the displayed price.
Several agency networks have documented a change in behavior over the past few years. A property that exceeds a certain time without qualified visits receives price reductions faster than before 2020. This observation means that an overvalued price no longer corrects slowly through negotiation: it corrects quickly, due to a lack of buyers.
For a seller, monitoring the number of visits in the first weeks provides a real-time reading that neither DVF nor a simulator can provide. For a buyer, spotting a property that has been online for a long time without a price drop may indicate a seller who is not willing to negotiate, or a defect not visible in the ad.
DPE and selling price: the variable that weighs the most since 2024
The energy ranking has become a price criterion in its own right, beyond its initial informative role. The gradual ban on renting properties rated G (effective) and then F (scheduled) pushes landlords to sell, which increases the supply of energy-intensive properties on the market.
The direct consequence: an increasing price gap between a property rated D and one rated F or G, with comparable area and location. Notaries in France have documented this trend in their recent economic notes.
To accurately estimate the price of a house, cross-referencing the DPE with DVF transactions in the same area allows for quantifying the local discount. In some areas, it represents a significant part of the price. In others, the effect remains marginal because the market is tight and buyers absorb the renovation cost.
Check the consistency between displayed DPE and asking price
A DPE rated E or lower displayed without a price drop compared to better-rated equivalent properties deserves scrutiny. The seller may have carried out work not yet accounted for in the diagnosis. Asking for the date of the last DPE and recent work invoices can help clarify.
The price of a house cannot be read from a single source. Public databases provide the signed price, simulators offer a theoretical range, and the ground reality (sale duration, DPE, actual condition) corrects both. The most reliable data remains that of DVF+, provided it is read considering the energy ranking and the state of the local market at the time of the transaction.