x
By using this website, you agree to our use of cookies to enhance your experience.

A raft of estate agencies are using a little-known technique called predictive analytics to target touting letters on onwers considered - statistically - most likely to sell their home.

The agencies include Knight Frank, Strutt & Parker, Carter Jonas, Faron Sutaria, Kinleigh Folkard & Hayward, Fine & Country, Engel & Voelkers, Ryder & Dutton and Arun Estates.

Their names appear as clients on the website of Outside View, which has adapted a statistical approach to calculating the likelihood of specific properties being sold in a six-month period.

Scatter-gun flyers through letterboxes asking owners if they want to sell sometimes work, but they're a blunt instrument admits Anne Currell, chief executive of Islington estate agency Currell - the latest firm to commission Outside View.

She's put her money where her mouth is by scrapping an order of 5,000 leaflets for a blanket' campaign, and will later this month instead put flyers through only 100 letterboxes of homes recommended by The Outside View.

The homes are individual addresses, not necessarily next to each other. To my eye, they have no connection with each other but the algorithm on which predictive analytics are based suggest they are the ones most likely to sell in the near future says Currell.

The technique of analysing past behaviour to judge future behaviour was first been used in this way in 1980 in US politics. Americans have used the tactics ever since: for example, Barack Obama employed scores of analysts crunching numbers to work out which swing-voters were most likely to vote for him if prompted by leaflets or phone calls or TV ads.

The Outside View - which advises organisations in a number of industries, not just those involved in selling property - is not disclosing the exact details of what statistics are used when the make their assessment. However, most are deived from Land Registry data.

It's easier in Britain than anywhere else in the world because there's so much data about house sales. Land Registry stats go back 20 years, there are long-standing portals like Rightmove and Zoopla, and estate agents hold data explains TOV founder Rob Symes.

He says there may be triggers' which are factored in to the algorithm; for example, there is an increased likelihood of someone selling their home if a similar property in the same street has sold recently at a good price.

The end result is that we've got a sophisticated analysis of what kind of owner in what kind of property is most likely to sell in any period up to 180 days from now claims Symes.

Noel Flint, head of London residential sales for Knight Frank, appears in a promotional video for The Outside View on the TOV website, admitting the firm's analysts were more accurate than one of his local agents at predicting which homes would sell over a year.

Comments

  • icon

    Really interesting. I guess unless people just move at completely random times, it can be possible to predict when, with a certain accuracy.

    • 13 April 2015 13:36 PM
  • icon

    Haha! Very true.

    • 13 April 2015 10:15 AM
  • icon

    Wow. those who live at an address longer are MORE LIKELY to move than people who have lived where they are for five minutes.

    If that needs pointing out surely one needs to reevaluate their chosen line of work...

    • 13 April 2015 08:21 AM
MovePal MovePal MovePal