Each month, our Automated Valuation Model (AVM) processes data on over 15 million properties in England and Wales, to estimate what each is worth.
This guide will cover what an AVM is, how ours works, and also how to view and interpret the valuations.
What is an AVM?
An AVM is a sophisticated piece of software designed to analyse and process vast volumes of data to estimate the value of properties.
Each AVM has a unique modelling process and often utilises different data sources. Accuracy and reliability vary depending on the quality of the data and the logic.
As the process is automated, AVMs offer instant results, low costs, can value multiple properties simultaneously, and aren't subject to human biases. However, they can struggle in some scenarios, such as with unique properties that have few comparables, and properties that have been recently renovated.
Financial institutions often utilise AVMs to value homes as part of the mortgage lending process. Individuals can also use AVMs to gain insights into the property market.
How does our AVM work?
Our AVM gathers data from trusted sources and then processes the data to generate property valuations.
The process is run each month to make use of the latest available data.
Data Collected
- Price Paid Data - This includes information on properties that have been bought or sold. Each record includes the property address, property type, transaction date, and flags such as if the property is a new build.
- Survey Data - These are property details collected by surveyors. Each includes property characteristics such as its size, age, and layout.
- Lender Data - This is data from mortgage lenders on house price trends, based on their lending activity.
Price Modelling
- Repeat Sales Index - Our model creates a house price index, tracking the average price of different property types by area and month. This is done by analysing properties that have sold more than once, using the price difference between sales to gauge the growth rate.
- Initial Valuation - Using the Repeat Sales Index, we calculate initial value estimates for each property. We pair each property with its corresponding index (for instance, a detached house in Birmingham city centre matches with the Birmingham city centre detached house index) and add on the growth trend since the last sale.
- Refinements - Next, we fine-tune the initial valuation by looking for other signals that better reflect the current value. If trends of nearby similar properties consistently deviate from the matched house price index, we adjust the valuation. The model also calculates a valuation based on the property's characteristics, such as size and location. It will use this valuation if there is an indication the property has been redeveloped or if this method usually produces more accurate valuations for similar properties.
- Confidence Range - After establishing a valuation for each property, we produce a confidence score. This score indicates our certainty that the valuation accurately reflects what the property is worth. It's based on factors such as how much data we have available, and the time since the property was last sold.
How do I use the AVM?
Our AVM is designed to be straightforward and user-friendly.
Search by address
Start by visiting the House Prices section. Enter the street or postcode of the property you wish to value. Keep in mind that searching by street address is usually more effective, as it often provides a broader range of comparable properties than a postcode search.
Review properties
A list of properties will appear from which you can locate the specific address you wish to value. If the property doesn't appear in HM Land Registry records, it won't show up in our results. In such cases, consider reviewing similar properties to get an estimation of the property's value.
Review valuation report
The valuation report provides a snapshot of the property and its valuation details.
- Valuation - At the top of the page, you'll find the valuation estimate, along with the value increase since the property's last sale.
- Confidence score - This is represented on the graph as the maximum and minimum estimates, providing a range where the property's market price is likely to fall. Note that a wider range indicates a lower confidence score.
- Property details - The report also includes the property's sold price history and characteristics, such as size and location, which the AVM considers during the valuation process.
Strengths and Limitations of the AVM
Understanding the strengths and limitations of our AVM enables you to effectively interpret its results and know when to seek additional information or expert input.
AVM strengths compared to traditional valuations
- Speed - Our AVM provides instant valuations, updated monthly. Traditional methods can take days or weeks as they represent a fixed point in time.
- Cost - Our AVM is free to use, estate agents and surveyors typically charge a fee for their services.
- Scope - Our AVM can simultaneously calculate the value of multiple properties, making it ideal for portfolio management. Traditional valuations are conducted one property at a time.
- Bias - Our AVM operates purely on data and isn't subject to the risk of human bias in traditional valuations. Estate agents, for example, may overvalue a home to increase the likelihood of them being chosen to market the property.
AVM weaknesses compared to traditional valuations
- Unique properties - Our AVM heavily relies on the quantity of data used in the inputs - it struggles with unique properties with few comparables. A traditional valuation can factor in elements not present in the data.
- Recent renovations - If properties have recently undergone significant renovations or extensions, our AVM may not have access to this information, leading to an undervaluation.
- Market value assumptions - Our AVM often assumes that the last price paid for a property was fair and representative of its value at that time. If a property was sold below or above market value, the AVM's valuation might be underestimated or inflated.
How accurate is the AVM?
We have tested our AVM and found that when a property is sold, the price paid will be within the valuation confidence range around 75% of the time. We calculated this figure by analysing approximately 500 transactions that took place over several months and comparing the sale prices with the AVM's estimates.
Keep in mind that 75% is the overall figure - the reliability increases in areas with many similar properties and for properties sold within the past five years. Conversely, the accuracy decreases for properties that haven't been recently sold, are unique, or are in an area with a low volume of transactions.
Before making any financial decisions, you should speak to a suitably qualified professional. Local market experts, such as estate agents, can provide more accurate assessments by considering factors that an AVM might overlook.
Addressing the AVM's Limitations
Broadening your knowledge of the local property market through analysis of similar properties and sold prices can give you further insight into a property's market value.
- Identify comparable properties - Research similar properties nearby and see their valuations. These 'comparables' can provide a more accurate gauge of the property's value, particularly if they have recently sold.
- Review sold prices - Examine sold prices for a range of properties nearby to understand the premiums people are paying for specific features relative to others. You can estimate a property's value by considering these premiums or discounts, combined with recent sales of other properties on the street.
Should I Buy a House?Understand the financial and personal advantages of property ownership along with the potential challenges
