Digital Technologies Challenge Real Estate Valuers
How will emerging technologies such as artificial intelligence and data analytics change the professional sphere of real estate valuers?
Digitisation is changing the occupational profile of the real estate appraiser, say Andreas Kunert and Stephan Halling, two experts on the field of digitisation in real estate. In their article published in Immobilien Zeitung they predict, that future technologies will increase the precision and efficiency of architectural appraisal. Thus, the interaction with technologies behind the tools will rise and with it, valuers will become more like consultants to banks and investors.
There is a number of online tools which promise to determine the real estate value at the touch of a button — without the appraiser ever having visited the location. Is this the future of property valuation? Will the real estate appraiser be replaced by artificial intelligence?
PropTechs offering data for valuation improvement
This is something that could be well assumed looking at the large number of PropTechs (What is PropTech?) that already are engaged in automated real estate valuation is increasing. They are called realxdata or Archilyse.
Archilyse, a PropTech startup from Zurich, processes addresses and available floorplans (as 2D images or IFC models) of real estate portfolios or valuers into simulations and calculations of around 300 features. Accessible through an API the results are then used to sharpen the statistical valuation models.
The additional data points delivered by Archilyse include e.g. view analyses of addresses or distances to points of interests. The advantages for valuers are two-fold: Providing qualitative and semantic features on the one hand reduces the bias, while on the other, having new features available the flexibility of the model increases. All in all, the rich dataset reduces the variance and helps to minimise residuals. Pilot studies with an external auditor showed that the provided features could improve the accuracy and precision of their valuation models drastically while the valuation process is also fully automated.
The potential of digital tools for valuation
That big data and artificial intelligence will play a big role in the future is also what Andreas Kunert, Director Research at vdpResearch, predicts: “You cannot get around it”. Stephan Halling, Head of Valuation and Consulting at HypoVereinsbank, further explains that many forecasts state, that especially in the realm of real estate valuation digitisation offers great potential. This is because it enables to process large amounts of data and produce high-speed results, to generate forecasts over three to five years and to forecast the market value for large portfolios.
Despite all new technologies and developments, Kunert says, that the key tasks of a valuer will remain the same: collecting data, analysing it and determining the market value. Also, the legal requirements for value determination provide the framework and limit the use of digitized valuations. However, the approach to the individual work tasks will get an additional technical focus, which not only has to be applied by the future and old-established valuers, but also needs to be understood. “The classic real estate management skill set will no longer be sufficient”, says Thomas Herr, EMEA Head of Digital Innovation at the real estate consultant CBRE.
Change in the collection and use of data
In the future it will be indispensable not only to use the data, but also to validate it beforehand, to question the sources and thus to estimate the quality of the data. “Unfortunately, it is difficult to see whether an online tool is good or not,” says Birger Ehrenberg, managing partner of ENA-Experts in Mainz. “The quality of the database, which can be evaluated by the tool, is often not disclosed. Because it makes a difference whether actual selling prices or just bid prices serve as input and whether these come from the last six months or from the last three years.”
Kunert estimates that in about ten to fifteen years, thanks to the availability of complete BIM models, even more data on the property will be accessible at the push of a button. This will add some important facts to the datasets. According to Kunert, the task of the evaluator will be to verify this data. “Clipboard and camera belong to the past”.
Technical and methodological skills to analyse data
Apart from collecting data, the great challenge lies in analysing it, which becomes an increasingly technical process. Machines can deal with large amounts of data faster and more precisely. HypoVereinsbank for instance, launched a digital tool more than ten years ago, with which the value of standard real estate such as single-family homes, townhouses and semi-detached houses is estimated. It changed the work of HypoVereinsbank. “By not having to do any part of the job by hand, we can build our analytics faster and have more time to evaluate more complex real estate projects”, says Halling.
However, entering data into a computer is not enough. “You need to know what you are doing,” says Kunert. Methodological competence in data analysis and a certain understanding of the functioning of algorithms are indispensable. The valuer does not have to be able to calculate himself, but he has to understand what is happening there, Kunert emphasises. In addition, people have to tell the machine what to do. Therefore, Kunert asks for a greater emphasis on mathematical statistics in the training — a topic that is already unpopular among prospective real estate evaluators. But this is not so easy as all future models are predicted to be much more complex than the models used today. Kunert refers to the neural networks, which are a black box for many. They have a good predictive quality and a high level of accuracy. The functionality is highly complex. But their usage must be well understood by experts because he must be able to prove his calculations and make them comprehensible, even if they are the result of a machine.
Data for different asset classes
To determine the market value is ultimately the key skill of the evaluator. As not only the visible facts count for that. A good market perception and the assessment of factors that can not be rationalized play important roles here. Especially in the field of standard real estate such as privately used real estate, digital tools are already applied. According to Herr, many banks often take the rough estimates of the computers in terms of financing as a benchmark. In Spain and Australia, where transaction data in the housing market are more transparent, the valuation is already fully automated. In Germany or Switzerland, however, the data situation is still very opaque.
However, the market of commercial real estate and large-volume transactions is a lot more interesting for valuers. Here, the precise assessment of the market value is more important, because it often involves special buildings or larger portfolios. Much more data and circumstances that can not be grasped by machines enter into play, which can significantly influence the value. For example, an appraiser can assess whether the transaction was a friendship price, how the city center could develop once the shopping mall in the outskirts opens, or how solvent the large tenant is, and whether bankruptcy rumors are in the headlines.
Additional benefits of automated evaluation
Herr assumes that “maybe there will be a far fewer evaluators” because the data-collection business will be highly automated. However, this does not necessarily mean a weakening of the profession; rather there will be a specialisation in regards to the understanding and application of future technologies and to the different asset classes. “The evaluator becomes a consultant,” says Kunert. Customers can be banks or investors with large portfolios. Halling explains that the time gained through automation can flow into the quality of the assessment. Legal and regulatory requirements could also be met more efficiently. Furthermore, additional detailed real estate ratings would be possible, resulting in numerous control and analysis options for real estate portfolios.
Do you also think that the digitisation will change the valuation process and the real estate business in general? Let me know your opinion on that!
Check out our articles about Architecture Analysis on Medium or follow us on Twitter.
