Linguistic Styles and Linguistic Style Matching in Online Real Estate Descriptions: Does Language Matter?
Brian L. Hoenstok 11833831
BSc Business Administration, University of Amsterdam Management in the Digital Age
Dr. F.B.I. Situmeang 22 June 2020
Statement of Originality
This document is written by student Brian Lorenzo Hoenstok, who declares to take full
responsibility for the contents of this document. I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
While the effect of persuasive language on prospective buyers has been plentifully explored in other academic fields, the same cannot be said for online real estate descriptions. To fill this lacuna, this paper sets out to examine and provide insight into how language in real estate descriptions on Funda.nl can affect the selling process and selling time of a property. Using LIWC dictionaries and custom word filters, based on expert interviews with real estate agents, this thesis examines the effect of varying linguistic styles and linguistic style matching in Funda real estate descriptions. Although not all independent variables showed significant change when the selected words were present, significant change on the average selling time was found in at least one variable for all property types. The results showed that the use of certain linguistic styles may influence the selling time of properties on Funda. Additionally, for three of the six property types examined, a significant difference in selling time was found when at least one of the words from the custom filter was used versus none. For future research, it is recommended to look at other aspects of online real estate listings which might influence the selling time of a property.
Linguistic Styles and Linguistic Style Matching in Online Real Estate Descriptions: Does Language Matter?
In 2018, the market for existing real estate properties in The Netherlands was worth a staggering amount of €62,765,000,000. This value was spread out over 218,491 properties, with an average selling price of €287,266 (CBS, 2020). The Dutch real estate market is of enormous interest to individuals looking to buy or sell property, and Funda – a popular Dutch online real estate platform that attracted over 46 million visitors per month in 2018 – plays a major role in how properties are sold. With approximately 71% of all houses for sale in The Netherlands being listed on Funda, with listing descriptions, pictures, floor plans, a neighborhood map and property features as the main selling points, it is a prime location for data collection and data analysis on the Dutch housing market. However, the wealth of insight that this platform can provide on the specificities of the online housing market has so far been relatively underexplored. In particular, while the effects of language on prospective buyers in other fields – think for instance of online reviews, offline real estate advertisements and commercial product descriptions – has been plentifully explored in academic scholarship (Petty, Cacioppo & Schumann, 1983; Ludwig et al., 2013), the same cannot be said for online real estate descriptions.
To fill this lacuna, this paper sets out to examine how language in real estate descriptions on Funda.nl affect the sales process and overall rate at which a property is sold. In particular, using LIWC analyses to measure the effect of varying linguistic styles and the use of linguistic styles matching, this thesis aims to determine whether affective language used in Funda real estate descriptions can positively influence the selling time of a property. To measure this, the following research question was established: “To what extent does the use of certain linguistic
styles, including the use of style matching to correspond to various different housing types, affect properties' selling time on Funda.nl?”
To answer this research question, an extensive literature review, expert interviews and statistical tests will be conducted. To begin, the broader scholarship on the relevant models and studies within the fields of linguistics and real estate, such as the elaboration likelihood model by Petty and Cacioppo (1986) and rhetorical devices such as ethos, logos and pathos, will be
reviewed in the theoretical framework. This section will also discuss how several core concepts, including linguistic style and linguistic style matching, are defined in this paper and define the research hypotheses. Following the theoretical framework, the methodology section will discuss how the Funda data was collected, demonstrate how the dependent and independent variables were selected, and explain the reasoning behind which statistical tests – such as t-tests and the Kendall’s Tau-b test – were selected. In addition, the methodology section will explain how the filters for linguistic style matching were created, and how three expert interviews with real estate agents fed into this process. In the results section, the outcome of the various statistical tests for six different property types – eengezinswoning, vrijstaand, herenhuis, twee-onder-een-kap, appartement, villa – will be analyzed. Finally, the implications for future research in the field of online buyer persuasion and language in real estate listings will be discussed, as well as the strengths and limitations of this thesis, and the research question will be answered.
Theoretical Framework
To set out the broader context in which this research is situated, the hegemonic theories and scholarship within the field of persuasion and the use of linguistic devices will first be explored.
The Elaboration Likelihood Model
Several established theories for persuasion exist in current literature, perhaps the most prominent one being the ‘elaboration likelihood model’ (ELM) by Petty and Cacioppo (1986). This model presumes that there are two ways to persuade an individual and bring about a change in attitude towards a message. The first option is the central route, in which an option is
thoroughly evaluated based on the facts and quality of the given information. The second route is the peripheral route, in which an option is evaluated in a more superficial manner, with more attention paid to minor, irrelevant or superficial aspects such as the attractiveness of the option. Generally speaking, the central route of processing is used if someone both has the motivation and ability to process the option, while the peripheral route is used when these factors are absent. Research has shown that when an option or a message is evaluated through the central route, the results of the change in attitude towards this option last longer and are more resistant than under the peripheral route.
The elaboration likelihood model has been used to study the impact of language in many different fields, such as advertising, customer reviews and civic crowdfunding (Petty, Cacioppo & Schumann, 1983; Sher & S.H. Lee, 2009; C.H. Lee et al., 2019). In the aforementioned work by C.H. Lee et al. (2019), which draws heavily on the ELM, significant evidence is found for a positive relationship between positive affective language – i.e. language that evokes positive emotions – and funding success within the context of civic crowdfunding. Conversely, the authors also found significant evidence for a reduction in the likelihood of success when an extensive use of social language was found. In further research on how individuals process information, Alter et al. (2009) have also found that words that can easily be processed, are
usually perceived as being more accurate and persuasive. In other fields such as advertising and online product reviews, various scholars have also proven that linguistic styles can influence the level of persuasion when a message is conveyed to an individual (Ludwig et al. 2013; Peng et al. 2004).
Applying the premise of the elaboration likelihood model to the sales process in the online real estate market, has significant implications for how real estate descriptions,
terminology and communication influence the manner in which persuasion takes place. Indeed, this raises the question whether language can bring about a change in attitude – in this case related to whether a buyer is persuaded to purchase a property. Different linguistic styles might, for example, yield more successful results for certain types of properties or certain types of buyers.
Linguistic style and style matching
Different definitions of the term ‘linguistic style’ exist within recent literature. Put simply, linguistic style can be defined as the use of specific words and language, such as the level of formality and the use of words that generate positive emotions, to convey a message. Another definition is provided by Holtgraves (2002, as cited in Blankenship & Craig, 2011), who defines linguistic style as “a set of pragmatic features that can modify the intended assertion in a message”. Blankenship and Craig (2011) add to this that linguistic styles might not change the actual message in terms of content, but can have a substantial effect on how both the message and the messenger are perceived. In this regard, linguistic styles can affect the persuasiveness of a message, even if the content remains unchanged (Holtgraves, 2002). Applying this
different linguistic styles to communicate about the property may yield different results in how the property is perceived.
Linguistic style matching (LSM) is the phenomenon in which individuals communicate using similar linguistic styles, for instance adapting the way they communicate to use similar phrasing and register. Ireland and Pennebaker (2010) have shown that utilizing LSM leads to an increase in the feeling of connectedness, credibility and shared perceptions. In their research, Ludwig et al. (2013) demonstrate that LSM cannot only be utilized in real life conversations and interactions, but can also be applied to written online text. The authors found that online reviews which were written in the same linguistic style as used by the product target audience, led to higher sales conversion rates. In addition, they discovered a strong positive effect of positive affective content on conversion rates, and a strong negative effect of negative affective content. At the same time, however, an increase in positive affective content did not have as significant an effect on conversion rates, as an increase in negative affective content, which indicates that different linguistic devices have varying degrees of impact on an individual’s perception of a message. Building on the fundamental elements of LSM, Ng and Bradac (1999) add that persuasion can also be increased by utilizing a linguistic style that is perceived as more prestigious.
Bridging both linguistic style and the elaboration likelihood model, Blankenship and Craig (2011) found that linguistic styles can have different effects under different levels of elaboration, which is also known as the level of motivation and ability possessed by the receiver of a message. When an individual has a low level of elaboration, linguistic styles are recognized as simple cues that have a limited effect on the individual’s perception. This is also in line with
the peripheral processing route, as set out in the ELM by Petty, Cacioppo and Schumann (1983). When an individual employs a moderate level of elaboration, the use of specific linguistic styles can influence the way the reader reflects on the text, and which elements are or are not
highlighted from their point of view. And finally, when individuals have a high level of
elaboration, linguistic styles can have a significant effect on how the individual recognizes and processes the methods of persuasion and how the validity of the presented arguments is assessed.
The various models set out in this section demonstrate that the manner in which a message is communicated, can substantially impact how the message is ultimately received by the reader. Considering the online sales process on Funda.nl, in which real estate agents communicate their message through language and pictures, this premise can also be applied to the way in which listing descriptions (i.e. the ‘message’) are received, perceived and processed by the receiver (i.e. the ‘buyer’).
Ethos, logos and pathos
In addition to models such as the ELM and the scholarship on linguistic styles, which both attempt to map the relationship between language and persuasion, one can also look at persuasion through Aristotle’s framework of ethos, logos and pathos. As described by Oates and Pryce (2007), ethos can be defined as an appeal made by a speaker or writer to establish
credibility and trust. Logos, on the other hand, can be defined as an appeal to logic or reasoning, while pathos can be defined as an appeal to the emotions and feelings of the audience. With their systematic analysis of language uttered by real estate agents, Oates and Pryce researched the usefulness of ethos, logos and pathos and the extent to which these are used within property listings. Based on their sample, the authors found no substantial evidence that ethos plays an
important role in the language used in property descriptions. It is, however, mentioned that ethos could play a part in other aspects of the real estate sales process, independent of property
descriptions, such as their general advertising and branding. The scholars find that the majority of listings tend to employ logos in their description, emphasizing rational facts and figures. Furthermore, they find that pathos also occurs frequently in the descriptions of property listings, prompting appeals to emotion that are not related to tangible characteristics of the property.
Arndt et al. (2013) conducted research to see whether real estate agents can use pathos to influence property perceptions, analyzing eight different audiovisual tours. By using different voices and different pictures to portray the real estate agents, Arndt et al. set out to research whether there was a difference in how the agents were perceived. The scholars found that, contrary to popular thought, agents who used pathos instead of logos did not come across as less trustworthy than those who did not use pathos. In addition to that, they found that when the tour was given by attractive female agents, the use of pathos significantly altered how respondents perceived the property, which demonstrates that the effectiveness of pathos depends on a variety of factors.
Going beyond how pathos influences buyers’ perceptions, the use of pathos in different regions and time periods has also been researched (Oates & Pryce, 2007). The scholars found that while holding all the other factors constant, the use of pathos went up in a buoyant and fast housing market and increased in spring and summer. In contrast to this, Robertson and Doig (2010) found that language was more emotive in periods in which the market was slower – in other words, they found a higher use of pathos in slower markets rather than in faster markets. According to their research in faster markets, the property listings during those periods were
more likely to focus on certain lifestyle elements. While these two papers are not in agreement on when the use of pathos is higher, they do agree on the fact that the use of this rhetorical device varies under different market conditions.
In the context of property listings, these findings imply that making use of either ethos, logos, pathos, or a combination, could lead to a positive increase in how a listing is perceived. This, in turn, might lead to an increase in viewings, which are a necessary step for selling the house, and could thus lead to a faster sale and a shorter time on the market.
Online searching and listing behavior
In addition to the main theories and frameworks as set out hitherto, there are numerous other aspects that influence how a property is perceived by prospective buyers. This section will cover these additional elements as distinguished in the literature.
Scholars typically assumed that people looking to buy property weigh their options systematically, meaning that they are not affected by persuasion tactics (Smith et al., 2006). However, a vast amount of research nowadays suggests that real estate agents actively try to influence property perceptions, be it by using certain language in property listings or in their real-time communication during viewings (Robertson & Doig, 2010; Oates & Pryce, 2007). Since buyers generally only make an offer or decide to purchase a property after viewing it, it is important for real estate agents to convince buyers to come for a viewing. Arndt et al. (2013) found that – following common sense – buyers only schedule viewings if the real estate listing overall sufficiently matches their wishes and gives them a generally positive impression of the property.
Research has also been conducted to see whether there have been changes over time in how properties are marketed. Analyzing Swiss listings between 1870 and 2007, Kriese and Scholz (2012) found that since the late 20th century, there has been a shift from highlighting class-based and socio-economic aspects in listings, to highlighting lifestyle aspects.
Finally, various authors have explored whether the inclusion of certain information in property listings affects the selling price of the property. Haag et al. (2000), for example, found that naming facts and figures in the property listing has a positive effect on the final selling price. Other research has also tried to establish whether there is a difference in selling price between when a property is sold by a real estate agent versus when a property is sold by the owner of the house (Curto et al., 2014; Levitt & Syverson, 2008). Curto et al. (2014) discovered that when a house was sold by its owners, there were certain factors that could influence the selling price which were not mentioned in the property listing. When a house was sold by a real estate agent, certain factors did not influence the selling price, but were still mentioned in the listing. Levitt and Syverson (2008) found that when the owner is a real estate agent by profession, they are able to sell their own house at a better price than the houses of their clients. The authors argue that this is caused by the fact that real estate agents are better-informed about the features and characteristics of their own house, but simultaneously found that real estate agents leave their own house on the market for approximately 9.5 days longer.
These studies, once again, show how the use of language and the way in which different aspects are marketed in property listings can affect the selling price of a property, as well as its time on the market. Funda – the platform of focus in this thesis – does not allow individuals to sell their own properties without having a realtor license. However, there are certain companies
which can upload the property listing to Funda for you in exchange for financial compensation. Usually, however, these companies offer to write the description for you. This implies that most, if not all, of the listings on Funda are written by professionals within the real estate field, thus lowering the plausibility of the phenomenon described by Curto et al. in the case of individuals selling their own house.
Concepts
Building on the concepts used in previous scholarship, as set out in the previous
paragraphs, this paper will utilize several core concepts related to linguistics and persuasion. To begin, building on the work of various scholars (Holtgraves, 2002; Blankenship & Craig, 2011), the term ‘linguistic style’ in this paper encompasses the use of affective language (pathos), social language, cognitive language, perceptual language and informal language. The concept
‘linguistic style matching’, in this paper, refers to the use of certain vocabulary and terminology by real estate agents in online listings, to reflect and match the type of language they believe a prospective buyer of that specific property type would utilize. This definition stems from previous research by Ludwig et al. (2013), who describe linguistic style matching as linguistic “congruence with the product interest group's typical linguistic style”.
Hypotheses
To answer the previously established research question – “To what extent does the use of certain linguistic styles, including the use of style matching to correspond to various different housing types, affect properties' selling time on Funda.nl?” – the following hypotheses were formulated.
Hypothesis 1. “The use of linguistic styles in the listing description influences properties' selling time on Funda.nl.” This hypothesis is based on various research on the effects of
linguistic styles and buyer perception or conversion rates, both inside and outside the real estate industry. The use of pathos in real estate sales has mostly been looked at in an offline context, but the finding that pathos and other linguistic style elements can influence buyers – and
consequently affect the selling time – is expected to hold true in an online context as well (Arndt et al., 2013).
Hypothesis 2. “The use of linguistic style matching affects properties' selling time on Funda.nl.” Research by Ludwig et al. (2013) demonstrates that when a higher degree of
linguistic style matching was employed in online product reviews, a higher conversion rate was found. This mechanism is expected to yield similar results in the online real estate market.
Methodology Data collection
The data for this research was provided by Dr. F.B.I. Situmeang (Faculty of Economics and Business, University of Amsterdam), who collected the datasets by scraping the website Funda.nl. The data consists of all the listings that were published on Funda in the year 2018, and which also had a date of sale occurring in 2018. A total of 209,338 listings were collected. The data was split up into six datasets, based on the six property types that could be deduced from the Funda data. The table below shows the number of properties (without any missing values that were of importance) that were analyzed in each dataset. It is important to note that some
properties were categorized under multiple, non-mutually exclusive, property types. All property types are described in table 1.
Table 1
Distribution of property types
Property type Number of listings Percentage of total listings
Eengezinswoning/ Single family property
126,628 62,9% Vrijstaand/ Detached house 33,320 16,5% Herenhuis/ Town house 8,565 4,3% Twee-onder-een-kap/ Double house 28,568 14,2% Appartement/ Apartment 54,187 26,9% Villa 5,757 2,9%
Total number of properties 201,459 100%
The descriptions of these listings were then analyzed using Linguistic Inquiry and Word Count (LIWC) software. This software is widely used in the field of social sciences, and the words that were included in the software’s dictionary and the corresponding variables these
words are attributed to were judged by an independent jury. The dictionaries’ and the variables’ internal reliability and external validity were also tested (Pennebaker et al., 2007).
Furthermore, interviews were conducted with three experts in real estate property sales, the transcripts of which can be found in Appendix A. First, Rogier de Vries from RE/MAX PRO was interviewed, who has handled the sales and purchase processes of approximately 200
properties. This respondent can be seen as an authority in this field as he has worked as a real estate agent – dealing with both selling and purchasing properties – for the last five years. Secondly, an interview was conducted with Marc van der Linden from Agterberg Makelaardij. This respondent is an assistant-realtor and has, by approximation, handled around 120 properties over the past 2 years. Lastly, Pascal de Roo from ZO Makelaars was interviewed. Having
approximately 15 years of experience as a realtor, being the owner of a real estate agency, and being a board member of the regional division of the NVM – the biggest Dutch real estate organization, with over 4000 connected real estate agencies – he too can be seen as an authority on the subject. Over the course of his real estate career, this respondent has handled
approximately 1,000 properties.
The interviews were conducted in a semi-structured manner, and ranged from 13 to 30 minutes in duration. While a set of questions had been created upfront which each respondent was asked to answer, there was also room for the interviewees to divert from these questions. For some responses, follow-up questions were asked. The interviews first established the real estate agents’ credibility and authority on the subject, after which the sales process – from receiving a request to sell a property, to signing the official documents at the notary – was discussed. The respondents were then asked whether they believe, in their professional capacity, that the
descriptions in Funda listings play a significant role in selling a property. Finally, the
interviewees were asked to describe which words, terms or characteristics they would generally use to describe the six different property types in table 1.
The purpose of these interviews was to gain insight into the broader context in which the real estate sales process takes place, and in particular to situate the role played by listing
descriptions on Funda.nl within this process. In addition, the words, terms and characteristics which were suggested by the respondents upon asking which vocabulary they associated with specific property types, fed into the word filters that were created to analyze the descriptions. Variables
Dependent variable. The dependent variable in this research is dateDiff, this variable was created by determining the difference between the date of posting the listing and the date on which the sale of the property occurred. As a rule, realtors connected to the NVM have to post every listing on Funda, even if the property has already been sold before being listed online. This led to a peak in dateDiffs, measured in the number of days, ranging from negative values to one. As it is highly likely that these properties were already sold before the listing was posted, all listings with a dateDiff of one day or lower were omitted from the dataset.
Independent variables. The chosen independent variables were based on both the LIWC software and the expert interviews. Following the LIWC software, the following variables were inferred: social words, affective words, cognitive processes, perceptual processes and informal speech (Pennebaker et al., 2007). For each variable, a dictionary – a list of words – was created to filter the text. The number of times that words within a certain variable are detected in the listing description, is shown as a percentage of the total word count of the description. Per
property type, they were normalized to values between 0 and 1, based on the maximum value of the variable in the data file for that specific property type.
The variable social words consists of the sub-variables family, friends and humans, and examples of words in these variables include ‘mate’, ‘talk’ and ‘buddy’. Affective words is comprised of the variables positive emotion and negative emotion; some examples of words included in these variables are ‘love’, ‘nice’ and ‘ugly’. Cognitive processes consists of the sub-variables insight, causation, discrepancy, tentative, certainty, inhibition, inclusive and exclusive, and examples of words within this sub-variable include ‘because’, ‘would’ and ‘perhaps’. The variable perceptual processes consists of the variables see, hear and feel, and examples of words included are ‘observing’, ‘feeling’ and ‘touch’. And lastly, informal speech is made up by the variables swear words, assent, nonfluencies and Fillers, examples of which include the terms ‘agree’, ‘yes’ and ‘damn’.
In addition to the variables produced by the LIWC software, also drawing on the expert interviews and the manual analysis of 5-10 listings per property type, filters were created to capture which words, terms or characteristics would be highlighted per property type. These filters counted the occurrence of every word that was found both in the filter and in the listing. These were added up and then turned into percentages of the total word count of the description. As is the case for the LIWC variables, these filter variables were normalized to values between 0 and 1 as well. The variables and the words they filter are presented in table 2 below.
Table 2
Property filters and their dictionaries
Variable names Words in filter
eengezinswoningFrequency
familie, gezin, kindvriendelijk, kinderen, veilig, spelen, school, scholen, ov, openbaar vervoer
vrijStaandeFrequency
tuin, rustige buurt, rustig gelegen, rustige ligging, rustige wijk, privacy, perceel, groene omgeving, bosrijk, ruim, kavel, lichtinval, garage, oprit
herenhuisFrequency
klassieke stijl, ornamenten, authentieke details, originele details, monumentaal, jaren 30, chique, chic, elegant, stijlvol, statig, hoge
plafonds, hoog plafond, royale, royaal, ruim
tweeInEenwoningFrequency
ligging, parkeren op eigen terrein, parkeergelegenheid, licht, ruim, oprit
appartementFrequency
meerjarenonderhoudsplan, mjop, meerjarenplan, gezond, actieve, actief, buitenruimte, balkon, terras, tuin, serre, gelijkvloers,
gerenoveerd
villaFrequency
luxe, afwerking, afgewerkt, alarm, camera, beveiliging, ruime kavel, ruim opgezet, parkeren op eigen terrein, garage, tuin, rustige buurt, rustig gelegen, rustige ligging, rustige wijk, privacy, perceel, groene
Analysis
Both hypotheses were tested by running Kendall’s Tau-b correlation test and t-tests on each individual property type data file. These tests established whether there is a correlation with dateDiff and the variable, and whether there is a significant difference in selling time between a variable having a value of either 0 or higher.
Results
The results section will be split into six sections, corresponding to the different property types that were analyzed using Python and Statistical Package for the Social Sciences version 26 (SPSS 26). For each property type, a correlation test and t-test were executed. While plotting all the independent variables with the dependent variable – dateDiff – in Python for all six data sets, no linear relationships could be seen, these plots can be found in Appendix B. As the assumption for a linear relationship is not met, Pearson's correlation test would not yield reliable results. Therefore, correlation is tested with Kendall's Tau-b test, the results of which can be found in Table 3, which shows the correlation of dateDiff with all the independent variables for all six property types. The assumptions for Kendall’s Tau-b – having variables that are measured on at least an ordinal scale and which represent paired observations – were met. While many
correlations are significant, all of them can be considered relatively weak correlations (Field, 2013).
The t-tests split the data into two groups. In the first group, the variable had a value of 0 – in other words, no occurrences of the words in the LIWC dictionary of the variable were found in the text. In the other group, the variable had a value of higher than 0 – in other words, at least one occurrence of a word from the LIWC dictionary of the variable was found. These t-tests and
the descriptives of the variables in each dataset can be found in Appendix C and Appendix D. For the t-tests, the assumptions of independence of observations and approximately normal distributions were met. The assumption that no significant outliers occurred was not met, but the decision was made to not remove these outliers, as this could have led to valuable data being lost. Furthermore, the assumption of homogeneity of variances was not met for some of the t-tests, but was met for others. Significance for the t-test was measured at a p-value of below .05.
Table 3
Correlations between dateDiff and independent variables for all six datasets
Variable dateDiff eengezins- woning dateDiff vrijstaand dateDiff herenhuis dateDiff twee-onder-een -kap dateDiff appartement dateDiff villa eengezinswoning- Frequency -.073** - - - - -vrijStaandeFrequency - -.012** - - - -herenhuisFrequency - - -.017* - - -tweeInEenwoning- Frequency - - - -.042** - -appartementFrequency - - - - -.030** -villaFrequency - - - -.046** socialNormalized .015** .016** .076** .046** .032** .036** affectNormalized -.043** -.022** -.046** -.028** -.028** -.022* posemoNormalized -.045** -.023** -.048** -.031** -.029** -.023** negemoNormalized -.005* .008 -.009 -.001 -.015** .008 cogmechNormalized .002 -.009* .004 .002 .039** -.007 perceptNormalized -.035** -.010** -.054** -.021** -.023** -.010 informalNormalized -.024** .004 -.010 -.023** -.033** .012
**. Correlation is significant at the .01 level (2-tailed). *. Correlation is significant at the .05 level (2-tailed).
Eengezinswoning
After omitting listings with missing relevant variables, the eengezinswoning dataset comprised a total of 126,628 listings. The average selling time, i.e. the duration of a property being on the market, was approximately 63 days for properties of this type. To further explore the effects of the independent variables on the selling time (dateDiff) of the house, t-tests were run, the results of which can be seen in Table 1 in Appendix D. The biggest significant change in means can be seen for eengezinswoningFrequency and posemoNormalized. For both of these variables, the t-test showed a significant decrease in the average selling time of more than 16 days.
Vrijstaand
After removing listings with missing relevant variables, a total of 33,320 listings remained. The average selling time of this property type was approximately 98 days. T-tests were conducted again, the results of which can be seen in Table 2 in Appendix D. From the results in Table 2, it can be seen that no significant changes in means were found. This means that we cannot state with a 95 percent confidence level that any of the independent variables had an effect on the average selling time of this property type.
Herenhuis
The herenhuis dataset consisted of 8,565 listings, after removing the listings with missing variables. The average selling time of this property type was approximately 81 days. Again, t-tests were run, the results of which can be seen in Table 3 in Appendix D. The most noteworthy significant correlations for this property type were affective language and perceptual processes.
A decrease of at least 34 days in the selling time of this property type occurred for these two variables.
Twee-onder-een-kap
The tweeInEenwoning data set contained 28,568 listings after removing the listings with missing variables. A property of this type was, on average, sold within 72 days. The t-tests for this data set showed that amongst others, posemoNormalized had a significant effect, with a decrease of 28 days in the mean selling time.
Appartement
After eengezinswoning, appartement was the most frequently occurring property type, with 54,187 listings remaining after the data was cleaned up. The average selling time for this property type was approximately 49 days. The main differences in means for this property type could be seen in appartementFrequency, posemoNormalized and cogmechNormalized. All of these variables showed a decrease of at least 25 days in the mean selling time.
Villa
Finally, the property type villa consisted of 5,757 listings after cleaning. Villas were on average sold within 107 days. posemoNormalized and informalNormalized showed significant correlations. Where posemoNormalized decreased the mean selling time by 38 days,
informalNormalized actually increased the mean selling time with 9 days. Discussion
This paper set out to answer the following question: “To what extent does the use of certain linguistic styles, including the use of style matching to correspond to various different
housing types, affect properties' selling time on Funda.nl?” To answer this question, the two hypotheses and the extent to which they are met will be analyzed.
Hypothesis 1
“The use of linguistic style matching affects properties' selling time on Funda.nl.” The first hypothesis predicted that the use of linguistic styles in the listing description would have an effect on the selling time of a property on Funda.nl. The analysis does indeed confirm that there are significant changes in the mean selling time between when no occurrence of the words from the LIWC dictionaries takes place, and at least one occurrence of one of the words is found in the description. This change in mean goes both ways. For example, when words that are associated with negative emotions were used in the appartement type property, the mean selling time increased by 11 days. Conversely, the use of social words and positive emotions in the descriptions decreased the mean selling time by at least 14 days for this property type. In all property types, at least one of the variables showed a significant change in average selling time. All variables showed a significant change in at least one of the properties. As expected, the variables socialNormalized, affectNormalized, posemoNormalized, cogmechNormalized, and percept only showed significant decreases in selling time. The variable negemoNormalized showed a significant increase in selling time, which was also expected. Finally, the variable informalNormalized showed mostly decreases in the average selling time, but showed an increase in average selling time for the property type villa. This might suggest that prospective buyers of villas, being in the higher segment of the market, respond negatively to informal language. Based on these findings, the null hypothesis – asserting that the use of certain
linguistic styles does not have an effect on the selling time – can be rejected. For all property types, H1 can be supported.
Hypothesis 2
“The use of linguistic styles in the listing description influences properties' selling time on Funda.nl.” The second hypothesis predicted that the use of linguistic style matching in the listing description would have an effect on the selling time of a property on Funda.nl. A
significant change in selling time could be found when one of the words included in the custom filters for the eengezinswoning, twee-onder-een-kap and appartement property types was found in the description. As expected, the average selling time of all three of these property types decreased when one or more words included in the filter were found. In line with previous scholarship, for example by Ludwig et al. (2013), these results imply that the use of linguistic style matching in online description may positively influence the selling time of a property. Due to the lack of a significant test result, the effect of the filters for vrijStaand, herenhuis and villa could not be determined. Thus, the null hypothesis – stating that linguistic style matching does not have an influence on the selling time of a property – can be rejected for properties of the eengezinswoning, twee-onder-een-kap and appartement type. Instead, for properties of this type, support is found for H2.
During the expert interviews with various real estate agents, the respondents held
divergent views on the importance of Funda.nl descriptions within the broader sales process. One respondent argued that adapting Funda descriptions to a buyer’s wants and needs is a key
element in persuading a buyer. Other respondents, however, believed that buyers generally are not attentive to the language used in Funda descriptions and that linguistic style matching has
little influence on the selling time. However, the results of this research indicate that the presence of certain linguistic styles and the use of linguistic style matching, can influence buyer
persuasion in the sales process of a property leading to a shorter time on the market. Limitations
It is important to note that one cannot conclude that this change in average selling time can solely be attributed to the independent variables in this research. Other factors that are not taken into account in this thesis could certainly also be at play and have an impact. For example, during the expert interviews several real estate agents stated that the pictures in the Funda listings had a greater impact on buyer perception and persuasion. It could be the case that the inclusion of more or less pictures, or uploading the pictures in a higher or lower quality, goes hand in hand with the other aspects that were found to have a shorter selling time. In this case, the change in average selling time could be caused by the photos, rather than the descriptions of Funda listings. Other factors that may contribute to buyer persuasion and a shorter selling time include property characteristics, such as whether a property has its own parking space or whether a garden is present, as well as the listing price and the neighborhood in which a property is located. Many factors can contribute to buyer persuasion and, consequently, the selling time of a property. It is likely that a combination of these factors is at play.
Furthermore, Funda’s algorithm could also have an impact on the selling time of a house. Kerste et al. (2012) found that Funda’s algorithm firstly lists listings posted by realtors connected to the NVM association, which are only then followed by listings of realtors that are not
perhaps properties with a more prominent page placement on Funda have a faster selling time simply because they are seen more.
An additional limitation of this thesis may be found in the custom filters that were created. As H1 was only partially supported, this could mean that the created filters were either incomplete or inaccurate. The filters were developed using three expert interviews and manual analysis of some of the listings included in the dataset, but the words included might not be all-encompassing.
In addition to the aforementioned limitations, the assumptions for outliers and
homogeneity of variances were violated, which means that the results of the t-test might not be fully reliable.
Recommendations for future research
For future research, it is recommended to perform this research on the real estate markets of other countries. This would allow us to draw transnational conclusions and see whether differences can be found between different cultures and nations.
Additionally it is advisable to run more statistical tests to further exactly specify how each variable plays into the selling time of a property and also find the effects of certain
variables combined. This could yield valuable insights on which linguistic styles should be used together to sell a property the quickest.
It is also advised that future research focuses on optimizing the created property filters. This can be done for example by conducting a quantitative survey research with many realtors as respondents. More expert interviews could also be conducted to gain a more complete and better understanding of what to include in these filters and to gain a better understanding of what other
forces might be in play within the sales process of a property. Lastly, it is advisable that future research relies on more and perhaps also alternative statistical tests.
Conclusion
To conclude, this research found significant changes in the average selling time of properties based on the independent variables. There is a change in average selling time between when none of the words included in the LIWC dictionaries for each independent variable occur, and when at least one of the words is found in the description. The custom filters created upon conducting expert interviews showed a significant decrease in selling time for properties of the eengezinswoning, twee-onder-een-kap and appartement property types. Thus, it can be
concluded that certain linguistic styles and linguistic style matching, affect the selling time of a property. This yields interesting insights for further research into how language may impact different aspects of the online real estate sales process.
References (n.d.). Retrieved from
https://opendata.cbs.nl/statline/#/CBS/nl/dataset/83913NED/table?ts=1533209588332 Alter, A. L., & Oppenheimer, D. M. (2009). Uniting the Tribes of Fluency to Form a
Metacognitive Nation. Personality and Social Psychology Review, 13(3), 219-235. doi:10.1177/1088868309341564
Arndt, A., Harrison, D. M., Lane, M. A., Seiler, M. J., & Seiler, V. L. (2013). Can Agents Influence Property Perceptions Through Their Appearance and Use of Pathos? Housing Studies, 28(8), 1105-1116. doi:10.1080/02673037.2013.802292
Blankenship, K. L., & Craig, T. Y. (2011). Language Use and Persuasion: Multiple Roles for Linguistic Styles. Social and Personality Psychology Compass, 5(4), 194-205.
doi:10.1111/j.1751-9004.2011.00344.x
Curto, R., Fregonara, E., & Semeraro, P. (2015). Listing behaviour in the Italian real estate market. International Journal of Housing Markets and Analysis, 8(1), 97-117. doi:10.1108/ijhma-01-2014-0003
Field, A. (2013). Discovering statistics using IBM SPSS statistics. sage.
Haag, J., Rutherford, R., & Thomson, T. (2000). Real estate agent remarks: help or hype?. Journal of Real Estate Research, 20(1-2), 205-215.
Holtgraves, T. (2002). Language as social action: Social psychology and language use. Erlbaum. Ireland, M. E., & Pennebaker, J. W. (2010). Language style matching in writing: Synchrony in
essays, correspondence, and poetry. Journal of Personality and Social Psychology, 99(3), 549-571. doi:10.1037/a0020386
Kriese, U., & Scholz, R. W. (2012). Lifestyle Ideas of House Builders and Housing Investors. Housing, Theory and Society, 29(3), 288-320. doi:10.1080/14036096.2011.629679
Lee, C. H., Bian, Y., Karaouzene, R., & Suleiman, N. (2019). Examining the role of narratives in civic crowdfunding: Linguistic style and message substance. Industrial Management & Data Systems, Ahead-of-print(Ahead-of-print). doi:10.1108/imds-08-2018-0370
Levitt, S. D., & Syverson, C. (2008). Market Distortions When Agents Are Better Informed: The Value of Information in Real Estate Transactions. Review of Economics and Statistics, 90(4), 599-611. doi:10.1162/rest.90.4.599
Ludwig, S., Ruyter, K. D., Friedman, M., Brüggen, E. C., Wetzels, M., & Pfann, G. (2013). More than Words: The Influence of Affective Content and Linguistic Style Matches in Online Reviews on Conversion Rates. Journal of Marketing, 77(1), 87-103.
doi:10.1509/jm.11.0560
Ng, S. H., & Bradac, J. J. (1999). Power in language: Verbal communication and social influence. Sage Publ.
Oates, S., & Pryce, G. (2007).Pathos and patter in real estate parlance. SHEFRN: Scottish Housing Economics and Finance Research Network (Discussion Paper).
Peng, K., Fan, Y., & Hsu, T. (2004). Proposing the content perception theory for the online content industry – a structural equation modeling. Industrial Management & Data Systems, 104(6), 469-489. doi:10.1108/02635570410543780
Petty, R. E., Cacioppo, J. T., & Schumann, D. (1983). Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement. Journal of Consumer Research, 10(2), 135. doi:10.1086/208954
Petty, R. E., & Cacioppo, J. T. (1986). The Elaboration Likelihood Model of Persuasion. Communication and Persuasion, 1-24. doi:10.1007/978-1-4612-4964-1_1
Robertson, K., & Doig, A. (2010). An Empirical Investigation of Variations in Real-Estate Marketing Language over a Market Cycle. Housing, Theory and Society, 27(2), 178-189. doi:10.1080/14036090903159994
Sher, P. J., & Lee, S. (2009). Consumer skepticism and online reviews: An Elaboration
Likelihood Model perspective. Social Behavior and Personality: An International Journal, 37(1), 137-143. doi:10.2224/sbp.2009.37.1.137
Smith, S. J., Munro, M., & Christie, H. (2006). Performing (Housing) Markets. Urban Studies, 43(1), 81-98. doi:10.1080/00420980500409276
Appendix A: Transcripts expert interviews (provided in Dutch) Respondent 1
Transcriptie van een interview met Rogier de Vries van RE/MAX PRO, geleid door Brian Hoenstok.
Rogier de Vries: Goeiemorgen. Brian Hoenstok: Goeiemorgen. RV: Hoe gaat het?
BH: Ja gaat prima, gaat prima! Het is mooi weer vandaag, dus denk dat ik straks maar even ga genieten van de zon.
RV: Moet je zeker doen, ik lig op het strand. BH: Oh lekker, heel lekker.
RV: Ja, dus als je het niet erg vindt, doen we geen video. Is dat oké? BH: Nee is prima! Is prima.
BH: Even kijken hoor. In ieder geval bedankt dat je de tijd wil nemen om wat vragen voor mij te beantwoorden.
RV: Tuurlijk, geen probleem.
BH: Ik had al aangegeven dat ik het gesprek ga opnemen zodat ik het later kan transcriberen. RV: Ja.
BH: En verder is het gesprek heel open. Ik heb bepaalde vragen waar ik graag antwoord op zou willen, maar je mag mij altijd onderbreken als je iets toe wil voegen, of als je een andere vraag of iets dergelijks hebt.
RV: Ja, oké.
BH: Ik wil heel graag beginnen met wat basisvragen. Allereerst, voor welk bedrijf werk je? RV: Bij RE/MAX PRO, Amsterdam.
BH: En hoe lang werk je al in de makelaardij? RV: Nu vijf jaar.
RV: Ik ben aan- en verkoopmakelaar. Dus ik begeleid mensen bij de verkoop van hun woning, en ik help zoekers bij het vinden van een woning.
BH: Oké. En doe je dit al de volledige vijf jaar of heb je nog een andere positie gehad in de makelaardij?
RV: Ik doe het inderdaad al vijf jaar.
BH: Oké. En als je een schatting moet maken, hoeveel huizen heb je behandeld? RV: In vijf jaar?
BH: Ja.
RV: Even denken hoor. In het begin deed ik er minder dan nu, nu doe ik er zestig per jaar. Dus het zijn er nu rond de tweehonderd.
BH: Zou je me wat meer kunnen vertellen over het hele proces vanaf het moment dat je een verzoek krijgt om hun huis te verkopen, tot het tekenen bij de notaris? En dan ben ik vooral ook geïnteresseerd rondom het deel van het plaatsen van een advertentie op Funda. Dus het
verzamelen van de informatie en dergelijke. RV: Oké. Gewoon helemaal vanaf de start?
BH: Ja, als je me er snel doorheen kan leiden, graag.
RV: Oké. Nou, we beginnen met een intakegesprek, waarin we de behoeftes en wensen van de potentiële verkopers bespreken, we leggen dan onze dienstverlening uit. Als we dan de opdracht hebben tot verkoop, dan gaan we naar de voorbereidingsfase. In die voorbereidingsfase moeten we alle media voor de woning verzamelen, dus dan komt er een fotograaf langs, die maakt foto’s, video, plattegronden en dat soort zaken. En wat we moeten doen, is dat we een dossier maken met allemaal documenten over de woning, bijvoorbeeld VvE-stukken, kadasterstukken, al dat soort zaken. Op het moment dat we het hele dossier compleet hebben, dan gaan we in de verkoop. Dus dan gaan we online, op Funda, social media, eigen website, dat soort zaken. Op het moment dat we online zijn, gaan we bezichtigingen houden. De meeste woningen, daar komen zoveel aanvragen voor bezichtigingen binnen, dat we een à twee weken bezichtigingen hebben. En dan dat we de inwoning op inschrijving verkopen, dus dat er een deadline is voor biedingen. BH: En voor het plaatsten van die advertentie op Funda, ik neem aan dat je daar bepaalde informatie voor nodig hebt? Is er een soort vragenlijst in het intakegesprek of iets dergelijks, of verzamelen jullie dat allemaal zelf?
RV: Dat is, zeg maar, in die voorbereiding, dat we het dossier compleet moeten maken. Heel veel halen we uit het kadaster, we laten we laten mensen ook wat vragenlijsten invullen over hun woning. Op die manier. Heel veel is al beschikbaar en kunnen wij gewoon online vinden.
BH: Oké, dan zou ik nu graag iets dieper in willen duiken op de beschrijvingen op de websites zoals Funda plaatsen. Hoe belangrijk is een Funda-omschrijving – naar jouw mening – voor het verkopen van een huis?
RH: Niet belangrijk.
BH: Wat denk je dat belangrijker is dan? Foto’s of…?
RV: De foto’s zijn het allerbelangrijkste. En wat heel veel mensen doen, is dat ze de tekst niet lezen, maar dat ze meteen naar beneden scrollen naar 'bijzonderheden'. Dat is waar de
kenmerken opgesomd staan. Vooral de jongeren doen dat, de ouderen lezen wel de tekst. Heel veel jongeren weten dat die tekst eigenlijk een mooi verkooppraatje is, maar meer ook niet. BH: En pas je zo’n tekst dan ook aan op of je iemand hebt die een wat jonger of ouder publiek aantrekt? Dat je er meer aandacht aan geeft bij een ouder publiek en minder aandacht bij een jong publiek misschien?
RV: Ja klopt, dat doen we wel. Maar we laten sowieso de tekst door een extern bureau schrijven, door het bedrijf dat ook onze foto’s en video maakt. Die maken de woningtekst, en die doen dat door middel van een… Die sturen een vragenlijst naar de verkopers toe, die vullen dat in, met wat kenmerken waarom ze er met zoveel plezier... Wat zijn de bijzonderheden, dingen die echt verteld moeten worden in de tekst. En dan maakt de tekstschrijver de tekst.
BH: En worden er nog bepaalde methodes of trucjes gebruikt in die beschrijvingen, om ze aantrekkelijker te laten overkomen of ze sneller te verkopen?
RV: Ze hebben natuurlijk vaste dingen die ze gebruiken, het lijkt altijd mooier in de tekst dan het is natuurlijk. Ja, gewoon standaard-verkoopteksten. Oude dingen zijn… Alles wordt een beetje mooi gemaakt.
BH: Ja. En in hoeverre denk je dat het taalgebruik en het gebruikmaken van bepaalde woorden of termen, een invloed heeft op de verkooptijd?
RV: Ik denk minimaal.
BH: Oké, dan zijn we er bijna. Funda maakt een onderscheid in verschillende soorten woningen en verschillende woningtypes – uit mijn hoofd zij het er een stuk of vijftig. Ik weet niet of je dan weet over welke ik het heb? Ik kan wel een lijstje sturen.
RV: Heb je het dan over een rijtjeswoning, hoekwoning, losstaande woningen, dat soort zaken? BH: Ja, precies.
BH: Ik zou er dan graag een aantal willen bespreken, en dan zou ik graag willen weten wat voor eigenschappen of woorden jij denkt dat belangrijk zijn in de omschrijving voor zo’n specifiek huis.
RV: Welke woorden belangrijk zijn?
BH: Of eigenschappen of termen, of eigenlijk ook die kenmerken waar je het eerder over had. Alles waarvan jij denk dat belangrijk is voor dat woningtype, om het makkelijker te maken om dat huis te verkopen.
BH: Dus de eerste is een vrijstaande woning.
RV: Ja, dan gaat het om de tuin die eromheen ligt, rustige buurt, vaak privacy, groene omgeving, luxe, zoiets denk ik. Ik ben er heel eerlijk in, ik maak die teksten nooit zelf, en ik vind ze niet heel belangrijk. Ik kijk, heel eerlijk, niet eens zo naar de teksten die gemaakt worden. Ik schrijf alleen nog of de verkopers nog verbeteringen zien, en dan geven zij nog vaak een of twee dingen die ze nog anders willen, en dan passen we dat aan. En hetzelfde voor die kenmerken, die doen we eigenlijk wel wat meer uitgebreid.
BH: De volgende is twee-onder-een-kapwoning.
RV: Dan is de staat van de woning vaak belangrijk, hoeveel slaapkamers. In Amsterdam is bij alle typen woningen de grond belangrijk, eigen grond of erfpacht. Tuin is natuurlijk ook belangrijk, de ligging. Parkeergelegenheid.
BH: Oké. Dan is de volgende hoekwoning/tussenwoning. RV: Ja dat is hetzelfde als twee-onder-een-kapwoning.
BH: Dus je hebt niet extra kenmerken die je voor twee-onder-een-kapwoning zou zetten. RV: Nee.
BH: En voor een eengezinswoning?
RV: Een eengezinswoning is eigenlijk ook niet anders. Het enige waar die nog wel anders is, is bij een appartement bijvoorbeeld. Bij een appartement heb je nog de VVE die belangrijk is, en een eventuele buitenruimte.
BH: Dat was er inderdaad ook eentje die ik erop had staan. En dan de laatste die ik erop heb staan is 'villa'. Heb je daar nog bijzondere kenmerken of dingen die opvallen of die je er vaak bij zet?
RV: Ja, daar gaat het natuurlijk veel meer op luxe en afwerking.
BH: En licht je dan bijvoorbeeld ook dingen toe zoals een alarmsysteem of is dat niet echt van belang?
RV: Doen we wel, maar is niet heel erg van belang.
BH: Als het om die prijsklasse gaat dan is het waarschijnlijk zo dat die mensen dat zelf ook wel kunnen plaatsen.
RV: Precies, ja.
BH: Precies. Dan heb ik nog één laatste vraag. Jullie zijn vooral werkzaam in Amsterdam neem ik aan?
RV: Ja, klopt.
BH: Heel veel internationale mensen. Is het bij jullie standard practice om het zowel in het Engels als in het Nederlands te doen, die omschrijvingen? Of is het altijd of alleen in het Nederlands of alleen in het Engels?
RV: Nee, allebei. Je kan aan de achterkant van Funda selecteren of je meerdere talen wil
invoeren, en als je naar de Engelstalige Funda gaat, dan krijg je ook de Engelstalige tekst te zien. BH: Oh oké, dat wist ik niet, dat is wel interessant.
RV: Ja, dus Funda selecteert automatisch wat voor tekst je krijgt.
BH: Oké, top! Dan zijn we er eigenlijk een beetje doorheen. Is er nog iets wat je nog toe zou willen voegen of iets waar ik niet naar heb gevraagd, maar waarvan je wel denkt dat het relevant is voor mijn onderzoek?
RV: Wat is precies het onderzoek dat je doet?
BH: Ik wilde je inderdaad niet beïnvloeden, dus ik kon nog niet te veel loslaten. Maar wat ik in principe wil onderzoeken is of er bepaalde elementen, woorden, termen of schrijfstijlen zijn, die invloed hebben op bepaalde woningtypes voor de verkoop van een huis, voor de omloop. RV: Ik denk dat je… Ik ben natuurlijk een wat jongere makelaar, dus ik ben van mening dat het helemaal niet uitmaakt. Waarschijnlijk als je een wat oudere makelaar vraagt... Vroeger was het gewoon veel belangrijker. Die zullen waarschijnlijk zeggen dat het misschien wel belangrijk is, maar ik ben van mening dat het niet uitmaakt.
BH: Oké. Nou ja, interessant in ieder geval om de verschillende perspectieven te horen. RV: Ja. Want heb je anderen gesproken?
BH: Nee, ik heb nog wel twee andere verdere interviews ingepland. Dus ik ben benieuwd of ik daar een ander verhaal krijg te horen of hetzelfde. Dan wil ik je heel erg bedanken voor je tijd en je informatie.
RV: Ja, heel veel succes ermee!
BH: Dan kan ik mijn scriptie met je delen zodra die af is. RV: Ja is goed, leuk.
RV: Ja, zal ik doen.
BH: Snelle vraag, ik weet niet of je dit mag vertellen en of je dit wilt vertellen, maar je zei dat een bedrijf voor jullie de huisomschrijvingen maakt. Zou je eventueel de naam van dat bedrijf kunnen noemen?
RV: Topr. BH: Super. RV: Yes?
BH: Yes! Hartstikke bedankt, en geniet nog van het mooie weer. RV: Jij ook.
BH: Fijne dag, doei.
Respondent 2
Interview met Marc van der Linden van makelaarsbedrijf Agterberg, geleid door Brian Hoenstok.
ML: Goedemorgen, met Marc van der Linden.
BH: Goedemorgen, u spreekt met Brian Hoenstok. Ik ben momenteel bezig met mijn scriptie, en daarin doe ik onderzoek naar de online verkoop van woningen, en dan specifiek via websites zoals Funda. En ik wilde vragen of jij of een van je collega’s mij wellicht zou kunnen helpen met een kort interview van ongeveer vijftien minuutjes. Dat zou mij namelijk enorm helpen met het verzamelen van kwalitatieve data, die ik zou kunnen verzamelen voor mijn analyse.
ML: In welke zin online verkoop? Dus zonder bezichtigen of…?
BH: Het gaat voornamelijk om advertenties die op Funda worden geplaatst. ML: Oké, nou schiet op, dan gaan we hem er doorheen jassen.
BH: Super, hartstikke bedankt. Vind je het erg als ik het gesprek opneem? ML: Ja, dat hoef ik niet. Liever niet.
BH: Het is puur voor het uittypen van het interview voor mijn scriptiebegeleider. ML: Oké, dan is het goed.
BH: Allereerst, bedankt dat je de tijd wil nemen. Ik wil beginnen met wat basisvragen stellen. Allereerst, hoe lang werk je al in de makelaardij?
ML: Twee jaar nu.
BH: En wat is je huidige functie en wat doe je op een dagelijkse basis?
ML: Ik ben nu assistent-makelaar, en dat komt omdat ik mijn kandidaat-papieren nog niet heb, daar ben ik mee bezig. Maar verder doe ik alles, dus bezichtigen, verkoopgesprekken,
aankoopbegeleiding, koopcontracten maken.
BH: En in de huidige functie zit je dus al twee jaar. Of heb je hiervoor nog wat anders gedaan in de makelaardij?
ML: Nee, dat is het, ja.
BH: Duidelijk. En als je een schatting moet maken, hoeveel huizen heb je dan behandeld? Dus als in het verkoopproces begeleiden, of bezichtigingen of dergelijke.
BH: Oké. Zou je me wat meer kunnen vertellen over het hele proces vanaf het moment dat je een verzoek krijgt om een huis te verkopen, tot het tekenen bij de notaris?
ML: Wij krijgen een lead doorgestuurd via een programma. Daar gaan we bij die mensen op gesprek. Als we bij dat huis zijn geweest, dan kunnen we een waarde afgeven van de woning. Dan sturen wij naar aanleiding van dat verzoek een mail naar die mensen, met onze
prijsindicatie, en wat wij daar allemaal voor doen. Als die mensen dat dan accepteren, dan gaan wij het dossier opstarten, dus dan gaan wij de mensen om documenten vragen –
eigendomsinformatie, VvE stukken als dat van toepassing is, dat soort zaken. Als wij dat helemaal ontvangen hebben, dan gaan we foto’s inplannen bij een extern bedrijf. Zij maken foto’s en een plattegrond. En als de woning het toelaat, dan laten we ook 360 graden en een video maken. Als dat allemaal ontvangen is, dan maken wij een aanbiedingstekst. Dan zetten we dat allemaal in ons programma en gaat dat op Funda.
Dan krijgen wij vanuit Funda aanvragen voor bezichtiging, en bellen wij die mensen op voor een afspraak in te plannen. Dan sturen wij een bevestiging van die afspraken met al de relevante stukken van de woning, dus kunnen mensen zich van tevoren al inlezen bij de woning. En dan krijg je ook hele andere gesprekken bij de woning, dus dan gaan ze zelf VvE dingen vragen. Dus dan weet je wel dat mensen echt geïnteresseerd zijn. En als ze dan verder willen gaan, dan gaan ze een bod doen met voorwaarden erin, dus de koopsom, voorbehoud financiering of niet,
voorbehoud bouwkundige keuring of niet, en binnen wat voor termijn ze willen afnemen. Als we dan een prijsovereenstemming hebben bereikt, dan gaan we een koopovereenkomst maken. Die wordt dan door de koper getekend, sorry, eerst door de verkoper dan door de koper. En dan gaat de bedenktijd van drie dagen in voor de koper. Als dat verstreken is, dan gaat het
financieringsvoorbehoud in. Dus je krijgt maar zes of acht weken voor het aanvragen van een financiering, en als hij dan is goedgekeurd, dan is het vaak nog twee weken wachten totdat die wordt gepasseerd bij de notaris. Dan ga je naar de notaris en wordt de hypotheekakte opgesteld, de akte van levering opgesteld, en ingeschreven in het kadaster.
BH: Oké, top, helemaal duidelijk. Dan zou ik nu graag iets dieper in willen gaan op Funda, en dan specifiek de omschrijvingen die jullie erbij plaatsen. Hoe belangrijk zou je zeggen dat de Funda-omschrijving is voor de verkoop van een huis?
ML: Wij hechten er heel veel waarde aan, puur omdat wij het professioneel over vinden komen – naar onze opdrachtgever, maar ook naar potentiële kandidaten. Alleen wij merken wel vaak dat mensen de tekst niet lezen en alleen naar de foto’s kijken.
BH: En zijn er ook niet bepaalde delen waar mensen dan wel naar kijken in de tekst, of skippen ze hem helemaal?
ML: Wat wij vaak doen is kenmerken toevoegen onderaan, dat zijn gewoon bulletpoints met belangrijke informatie over de woning – dus hoeveel vierkante meter het is, is het gelegen op eigen grond of erfpacht, dat soort zaken. Dat lezen mensen vaak wel.
BH: Dus dat zijn dan inderdaad bepaalde dingen waardoor je wel de aandacht kan trekken naar de omschrijving. En in hoeverre denk je dat het taalgebruik en het gebruik van bepaalde woorden of termen invloed heeft op de verkooptijd? Dus taalgebruik en het gebruik van bepaalde woorden in die omschrijvingen?
ML: Nee, dat heeft niet zo veel… Kijk, tuurlijk, het moet gewoon een vlot lopende tekst en een goede tekst zijn. Wat vooral belangrijk is, is dat er geen taalfouten in staan of schrijffouten. Maar ik denk niet dat als er beter of minder goed wordt geschreven, dat er dan sneller verkocht wordt op basis van de tekst.
BH: Oké, duidelijk. En zijn er naar jouw weten nog bepaalde termen of woorden, of taalgebruik waarvan je hebt ondervonden dat ze goed kunnen werken in omschrijvingen?
ML: Nou dan zie je vaak 'ruim', 'licht', 'luxe', dat soort termen.
BH: Dan zijn we er bijna. Funda maakt onderscheid in verschillende soorten woningen,
bijvoorbeeld vrijstaande woning, of twee-onder-een-kapwoning, tussenwoning, hoekwoning. Ik zou graag even deze woningen door willen gaan, een aantal types. En dan zou ik graag willen weten wat voor kenmerken of eigenschappen je naar voren zou laten komen in de omschrijving. ML: Ja.
BH: Dus de eerste is vrijstaande woning.
ML: Nou ja, op een ruime kavel gelegen vaak. Veel lichtinval vanwege de vele raampartijen. Ruim.
BH: En dan twee-onder-een-kapwoning?
ML: Ook licht, vaak ook ruim. Vaak een eigen parkeergelegenheid op eigen terrein. BH: En dan tussenwoning of hoekwoning?
ML: Fijne gezinswoning. Ja, wel moeilijk, je moet wel de woning zien, wil je daar echt makkelijk antwoord op willen geven.
BH: Ja, er zit natuurlijk ook veel variatie in dat woningtype. En appartement?
ML: Ja, heel erg afhankelijk van wat voor appartement je hebt, want je hebt natuurlijk
portiekwoningen veel in Den Haag, of je hebt een appartement in een appartementencomplex, en dan kan het heel anders zijn. Kijk, als je een appartementencomplex hebt op de elfde verdieping, en je hebt een appartement op de elfde verdieping bedoel ik, dan ga je heel anders omschrijven dan een portiekwoning op de eerste etage.
