Sale price premium cannot be claimed by advertising alone

Last week I challenged the claim made by the NZ Herald that the power of Herald Homes print advertising added an extra 20% to the sale price of property featured in Herald Homes. I wrote a detailed article "Can advertising generate extra sales price for property?" challenging their assertion which I firmly believed was misleading.

I wrote to the NZ Herald to request answers to questions regarding the research, in order to enable me to better understand the research they quote in support of their claim. This was their response I received :

The research was carried out on our behalf by TNS Research, a well-respected global research provider.

TNS purchased a list of recently sold properties from CoreLogic – all properties were residential dwellings in the Auckland area (Franklin to Rodney), and had sold in the most recent 6 month period (1st August 2013 – to 31st January 2014) for over $1 million. This list contained fields for address, land area, dwelling area, number of bedrooms, CV, sale price, days on market, agency and agent.

This list was then supplied to the relevant agencies to populate with marketing data for each property. We asked for spend by media options (Herald, Property Press, TradeMe, etc.), open home visits, enquiries and numbers of auction bidders where relevant. This took a reasonable amount of time to collate but resulted in a completed list of 252 properties. Success measures were based on:
· Sale price to CV ratio
· Premium achieved (sale price minus cv)

This allowed TNS to look at similar properties (based on the data from CoreLogic) that included / didn’t include Herald Homes in their marketing mix (based on the additional agency data). Our advertising is based on the results of this analysis. We are absolutely confident in the methodology –as we have based our claims on actual sales and market data.
— NZ Herald : Brad Glading, Head of Research & Insights : NZ M & E Ltd 9 October 2014

This response is certainly illuminating. I have no problem from a research perspective as the sample size of 252 properties is sufficiently robust and the methodology and factual basis is without question. The one issue I do have is that they chose to limit the sample to only those properties that sold for over $1m. This fact is not detailed in the supportive advertising when disclosing the research. It should be disclosed as in my mind it creates a bias in the advertising campaign as people would rightly believe the claim applied to any property sold with Herald Homes advertising.

As a point of reference and context. Using Barfoot & Thompson sales data for Auckland $1m + property sales represented just 1 in 8 of all property sales in the period from August 2013 to January 2014. Therefore this selective sample set of $1m + sales is actually a small fraction of the market. 

Far more important than the bias of the sample set is the belief by Herald Homes that the premium price achieved at sale is solely due to the advertising. The data they have analysed certainly shows a correlation between premium sale price and advertising BUT their assertion is that the advertising is causing the resultant premium sale price - this is not provable.

The NZ Herald have fallen into the classic trap of believing that "Correlation implies causation" where in fact the opposite is true. The fact is that a correlation between two variables does not necessarily imply that one causes the other.

The fact is there are just too many variables at work in the factors influencing the sale of a property for any one component to be isolated to be proven to be the cause of a sale price premium. I am certain that with the same set of data Trade Me could argue with confidence that advertising on Trade Me would result in a sale price premium or that auction sale process caused a premium or that real estate agents with blue logos caused a sale price premium.

On a more serious note I believe that there is a alternative hypothesis to explain the results of the research undertaken on behalf of the NZ Herald.

Alternative Hypothesis

 A key factor in the Auckland property market over the past few years especially of property  sales value over $1m is the prevalence of inner city renovated properties. These are properties which have achieved significant premiums over their respective CV due to the substantial investment in renovation of many hundreds of thousands of dollars. I detailed this situation in an article from earlier this year "Property statistics can be misleading".

I believe that such properties which have been renovated and then come onto the market attract a higher level of marketing investment which clearly will involve the Herald Homes together with other advertising in print and online. These properties also attain a higher margin over CV because effectively they are new homes and the CV relates to the former home on the site. These type of properties are more than likely to be skewing the results of the research and creating the correlation.

To prove my assertion I undertook a detailed piece of research of my own. I took the inner city suburb of Grey Lynn and with the assistance of a friendly agent accessed the CV and sale price for a total of 62 properties sold in the suburb between August 2013 and January 2014. Of this total, 46 properties had a sale price in excess of $1m. I then went through these properties one-by-one to identify which had been sold following a renovation. There were 16 properties which had been renovated.

Here is the resultant sale price premium for renovated and non renovated property compared to CV.

Renovated properties: 16 - average sale price over CV $403,125

Un-renovated properties: 30 - average sale price over CV $335,826

The margin is 20%! 

So I could quite legitimately say that the 20% premium claimed by the NZ Herald as attributed to the power of Herald Homes advertising is as likely to be the result (in the case of Grey Lynn) to be down to properties having undergone a renovation.

I judge that the NZ Herald has leapt to a interpretation of causation from their commissioned research simply by identifying a correlation potentially the result of significant renovations and therefore the advertising claim is without foundation and therefore fundamentally misleading.