Exploring the Power of Content and Visitor Sentiment: A Study of Web Traffic Dynamics in South Africa’s Residential Real Estate Landscape

Abstract

The real estate sector has increasingly shifted toward digital platforms, where content sentiment plays a crucial yet understudied role in driving user engagement. While sentiment analysis has been widely applied in retail and finance, its impact on real estate web traffic remains poorly understood, particularly in competitive digital marketplaces. This study examines the relationship between sentiment in web content and the traffic it attracts on residential real estate websites in South Africa. Specifically, it examines how different sentiments associated with the type of content (articles versus property listings) influence total monthly web traffic and user engagement. A quantitative analysis of six years (2017–2023) of scraped data from Property24, Remax, and Private Property employed R (rvest, sentimentr, and stats packages) for web scraping, sentiment analysis, and ANOVA testing to evaluate relationships between content sentiment, type (listings vs. articles), and web traffic metrics. The analysis revealed a significant impact of sentiment on web traffic, indicating that the sentiment of web content influences visitor numbers. Specifically, property listings generated a total of 16,780,623 monthly visitors, significantly surpassing the 13,407,521 visitors attracted by articles. This study contributes empirical evidence regarding the influence of content sentiment and content type on web traffic within the South African real estate market. It highlights the critical role of sentiment in shaping web traffic and potentially user engagement and provides actionable insights for real estate developers and marketers seeking to optimize their content strategies to improve user attraction and retention.

1. Introduction

1.1. Background

The residential real estate sector in South Africa has undergone a significant transformation in recent years, largely driven by advancements in technology and the increasing reliance on digital platforms for marketing properties [1,2,3]. Websites have evolved from mere listings into dynamic, interactive platforms that provide potential buyers with valuable resources, including detailed property information, virtual tours, and immersive experiences [4,5,6]. This evolution reflects a broader global trend in which the real estate market is adapting to the digital age, emphasizing the importance of a robust online presence. In addition to this technological shift, recent legislative reforms such as the 2023 amendments to the Property Practitioners Act have introduced stricter consumer protection and transparency standards, further transforming how real estate services are delivered and perceived.

In the South African context, the real estate market is not only dynamic but also highly competitive, characterized by diverse consumer preferences and varied regional characteristics [7,8]. Real estate companies continually strive to enhance their digital marketing strategies to attract web traffic and convert online visitors into prospective homebuyers [9,10]. As most property searches commence online, understanding the multifaceted factors that drive web traffic to real estate websites becomes crucial for companies seeking a competitive edge. These factors encompass website design, user experience, marketing strategies, and, importantly, the quality and sentiment of the content presented. Emerging trends such as fractional property ownership and tokenization of assets are also gaining traction, presenting new challenges and opportunities for online engagement.

One critical element influencing web traffic is the sentiment expressed in the website content [11]. Sentiment analysis, a specialized field within natural language processing (NLP), enables the assessment of emotional tone and attitude within textual content [12]. This analytical approach has found extensive applications in various domains, including marketing and customer feedback analysis [13]. For instance, positive sentiment such as “charming family home with ample natural light and secure parking” may attract users more than neutral or negative phrasing like “two-bedroom house with limited amenities.” Content types commonly found on real estate websites include listing descriptions, blog articles, customer testimonials, and area guides. For real estate websites in South Africa, the attributes of property listings and articles can significantly impact visitor perceptions and engagement [14]. Positive sentiment about the attributes of companies’ websites can create an inviting atmosphere that encourages inquiries and potential sales, while negative sentiment may deter potential buyers, leading to lost opportunities [15]. These dynamics have become increasingly important in the post-COVID-19 environment, where consumer behaviour has shifted toward cautious and research-intensive online interactions [16], influenced by macroeconomic pressures such as rising interest rates and inflation [17,18].

Unlike other industries, such as retail, finance, and hospitality, where sentiment analysis is widely adopted as a tool for managing consumer perceptions and optimizing customer engagement [11,19], sentiment’s influence on web traffic in the South African real estate sector remains underexplored. In the retail sector, sentiment analysis is used to gauge customer satisfaction and adjust marketing strategies based on real-time feedback [13]. The finance industry leverages sentiment analysis to monitor market trends and investor confidence, directly informing investment strategies [20]. Hospitality businesses apply sentiment insights to manage online reviews and enhance guest experiences, tailoring services to better meet customer expectations [13]. While these sectors have embraced sentiment analysis, real estate research has traditionally focused more on price forecasting, market cycles, and investor behaviour, with less attention to digital engagement dynamics such as user traffic and interaction. This pattern is especially pronounced in Africa, where sentiment-related real estate research remains scarce. Most existing studies are concentrated in developed economies such as the U.S., U.K., and Europe [21,22], with Africa and South Africa in particular, largely underrepresented in this area.

Focusing on South Africa is, therefore, both timely and necessary. As one of the continent’s most advanced real estate markets with robust digital infrastructure, South Africa offers a relevant testing ground for examining how digital content sentiment affects consumer engagement in an African context [23]. The interplay between evolving policy, shifting consumer preferences, and economic volatility highlights a unique confluence of factors that warrants urgent scholarly attention.

Although the results may not generalize across all African countries due to socioeconomic and infrastructural diversity, they offer critical insights into the emerging dynamics of digital real estate marketing on the continent. This study, thus, contributes to bridging a geographic and methodological gap in the global literature.

The objective of this research is to investigate how the emotional tone and types of online real estate content influence web traffic and user engagement, and by extension, contribute to potential sales conversions. By analyzing sentiment impact, the research seeks to uncover insights into how positive or negative content influences web traffic and subsequently impacts property sales.

1.2. Research Problem

Many studies have reviewed and explored the effect of sentiment on real estate returns both from investor and occupier perspectives [20,21,22]. As noted by [20], market sentiment is a driver of asset prices, and since the advent of artificial neural networks and other machine learning algorithms, emphasis has been placed on how content sentiment also affects the price of real estate.

However, much of this research has concentrated on financial markets or macroeconomic-level data, with little focus on how sentiment in digital property content influences micro-level user behaviours, such as web traffic or site engagement. Despite the growing relevance of sentiment analysis in digital marketing, a notable gap exists in understanding its specific impact on web traffic within the South African residential real estate sector. While existing literature has highlighted the importance of web traffic in determining the success of real estate companies [15,24], there is limited research focusing on how the sentiment of online content shapes traffic patterns and influences potential sales.

More importantly, the global research landscape reveals a strong bias toward developed economies, with Africa and South Africa in particular, often excluded from sentiment-focused real estate studies [23]. Where African markets are included, findings suggest sentiment plays a relatively smaller but potentially growing role compared to economic fundamentals [11]. South Africa, with its relatively mature real estate sector and increasing digital penetration, represents a critical stepping stone for examining sentiment-related dynamics in African contexts.

To address this gap, this study explores the relationship between website content sentiment and web traffic in the South African residential real estate market. Recent studies have demonstrated the effectiveness of sentiment analysis in predicting user behaviour and preferences across various digital platforms [11,13]. However, the extent to which sentiment influences web traffic and ultimately affects property sales remains an area that warrants further exploration. There is also a methodological gap in local research that incorporates NLP-driven sentiment scoring in conjunction with actual web traffic analytics from South African real estate firms. This research aims to quantify the impact of content sentiment on web traffic in real estate companies, equipping real estate companies with the insights necessary to optimize their content strategies and enhance their outreach to a broader online audience.

1.3. Research Questions

The research questions derived from the defined objectives are as follows:

• What is the impact of website content sentiment on web traffic of real estate companies in South Africa?
• How do different content types (articles or property listings) affect web traffic on real estate websites?

1.4. Hypotheses

Hypothesis 1 (H1).

Website content sentiment has a significant impact on web traffic for real estate websites.

Hypothesis 2 (H2).

Different content categories have a significant influence on web traffic for real estate websites.

1.5. Significance of the Research

This study holds significant value both academically and practically, particularly in the realm of digital marketing for real estate. Academically, it addresses gaps in understanding how website content sentiment and content types (articles or property listings) influence web traffic, offering a deeper insight into user behaviour in the South African real estate market. By examining these factors, the research contributes to the broader knowledge of content optimization and digital engagement strategies in real estate marketing. Practically, the study provides real estate companies and marketers with actionable insights to refine their content strategies, boost user engagement, and improve business performance through more targeted and data-driven decision-making. These findings can guide companies in enhancing user experiences and driving higher traffic to their websites, ultimately contributing to business growth.

2. Literature Review

2.1. Overview of the Real Estate Industry in South Africa

The real estate industry in South Africa holds a pivotal role in the national economy, contributing approximately 11.4% to the Gross Domestic Product (GDP) in 2023 [24,25]. The sector encompasses various market segments, each catering to distinct property types and investor preferences.

The historical evolution of the South African real estate market can be traced back to colonial times, with significant transformations occurring over the years. The sector was notably influenced by apartheid-era policies, which enforced land restrictions and created racially segregated neighbourhoods [22]. The post-apartheid era has ushered in significant changes, promoting more inclusive property ownership and opening new avenues for investors and homebuyers [10,26].

Despite its resilience, the real estate market in South Africa faces several challenges, including affordability issues, disputes over property rights, urbanization pressures, housing shortages, and market inequality. Addressing these challenges requires innovative solutions and proactive policy measures to ensure sustainable growth [14,17]. The legal framework in South Africa robustly supports property rights, with constitutional guarantees for secure property ownership and protection against arbitrary expropriation, fostering a stable investment environment [23]. This supportive legal landscape provides a foundation for market confidence, encouraging both domestic and foreign investment. However, the evolving economic conditions and socio-political dynamics necessitate continuous adaptation within the sector. Understanding these factors is crucial for stakeholders seeking to navigate the complexities of the South African real estate market and capitalize on emerging opportunities while addressing existing challenges [11]. A matrix of key empirical literature reviewed in this section is presented in Table 1.

Table 1. Summary of key literature reviewed.

No.	Author(s)	Year	Title	Objective/Purpose	Context/Setting (Residential, Commercial, etc.)	Methodology	Key Variables	Key Findings/Results
1	[27]	2021	Lifestyle, opportunity and attraction images: real estate platforms and the digital remediation of space	To analyze the impact of real estate platforms on the presentation of properties and to critique how these platforms frame homes as commodities, influenced by neoliberal ideology.	Residential real estate; advertising across platforms (Domain and realestate.com.au) with a focus on buyer interaction with property listings.	Textual, visual, spatial, and content analysis of online real estate advertising.	Presentation strategies of properties (lifestyle image, opportunity image, attraction image); geographical and social differences in presentation.	Identification of three dominant promotional strategies (lifestyle, opportunity, attraction images), highlighting variations by geography, class, and subculture in the portrayal of properties.
2	[7]	2018	Gender roles and online marketing communications in real estate markets	The aim of the article is to identify contemporary activities undertaken in marketing communications in the real estate market, focusing on gender roles of suppliers and customers. And to analyse the use of the internet in real estate marketing communications.	real estate market, specifically focusing on marketing communications involving property developers and real estate agents.	literature analysis, web content analysis provided by real estate vendors, and telephone interviews with representatives from property developers and real estate agents.	Gender roles in marketing communications. Use of online marketing (e.g., websites, social media). Marketing tools (e.g., images vs. texts, mobile applications, augmented reality).	Online marketing is the most important communication tool in the real estate market. Real estate vendors frequently use social media in marketing. There is a trend toward using images instead of text in communications. Few property developers and real estate agents consider gender-related differences in marketing. There is a need to focus marketing communications on younger customers, leading to the use of new media sources like mobile apps, augmented reality, and virtual reality.
3	[9]	2023	Digital Marketing Strategies in Real Estate for Property Pistol Realty Pvt.Ltd, Bengaluru	The aim of the paper is to investigate and analyse the digital marketing strategies employed by real estate professionals, agencies, and developers to achieve their objectives in the digital age.	real estate industry, with a focus on digital marketing strategies used for residential, commercial, and industrial purposes.	analyzing the effectiveness of digital marketing strategies to reach target audiences and generate both qualitative and quantitative data.	Digital marketing strategies (e.g., search engines, corporate websites, social media profiles). Effectiveness in reaching target audiences. Qualitative and quantitative data on marketing effectiveness.	The real estate industry has been significantly transformed by the rise in digital marketing. Potential real estate customers prefer conducting online research. Real estate companies need to ensure that search engines, corporate websites, and social media profiles strengthen their image and improve communication with customers.
4	[10]	2022	Learning the Real Estate Market Resilience: The Effectiveness of Internet Platforms in Marketing Strategy	The aim of the research is to study the effectiveness of Internet platforms by investigating the contents’ ranking of these platforms for marketing strategy by real estate agents.	real estate industry, with a focus on digital marketing strategies employed by real estate agents, particularly during the COVID-19 pandemic.	The study adopts a quantitative approach, specifically using Relative Importance Index (RII) analysis to evaluate the effectiveness of different Internet platforms for marketing strategies.	Effectiveness of Internet platforms for real estate marketing. Ranking of contents on platforms (e.g., Facebook, other digital platforms). Social media engagement and its impact on sales.	Facebook is the most important platform for real estate marketing, especially for sales and property promotion. To drive sales, real estate agents should regularly post relevant content on social media to achieve engagement and maximize profits. The pandemic influenced the need for more creative marketing strategies in the real estate sector, highlighting the impact of digital platforms on business.
5	[23]	2018	Towards Improved Performance in Marketing: The Use of Property-based websites by Estate Surveyors and Valuers in Lagos, Nigeria	The aim of the study is to investigate the marketing methods most used in Lagos’ metropolis and assess the extent to which estate surveyors and valuers in Nigeria have adopted property-based websites for marketing their real estate services.	the real estate sector in Lagos’ metropolis, Nigeria, focusing on the marketing practices of estate surveyors and valuers.	The study used a random sampling method to gather data from estate surveyors and valuers in Lagos. The data were analyzed using SPSS V24, applying Friedman’s Test and One-Way Analysis of Variance (ANOVA).	Marketing methods used by estate surveyors and valuers. Adoption of property-based websites for marketing. Comparison of conventional/traditional methods vs. digital methods.	Conventional/traditional marketing methods dominate the activities of estate surveyors and valuers in Lagos. The limited adoption of property-based websites for marketing weakens agency practice in Nigeria in the context of global marketing trends.
6	[26]	2019	Changing realty: altering paths of brand discovery for real estate websites in India	To explore brand awareness for Indian real estate websites, focusing on the modified processes of brand discovery and its antecedents.	Real estate websites in India.	Analyzed brand discovery behaviour and traffic metrics (direct and indirect) to assess brand awareness.	Brand awareness (firm-generated, user-generated, experience-initiated). Website traffic as an awareness indicator.	Brand awareness is influenced by firm-generated, user-generated, and experience-initiated factors. Search engines are the key discovery tool, with brand experience impacting awareness and visit intent.
7	[8]	2019	Effectiveness of Applying Marketing Tools in Real Estate Trading	To evaluate the effectiveness of marketing tools used by real estate agents in Krakow.	Real estate market in Krakow.	Diagnostic survey and questionnaire surveys with a cross-sectional analysis of various marketing tools and their effectiveness.	Marketing tools (online advertising, social media, classified ads, property visualizations, and descriptions). Effectiveness of marketing activities.	Online advertising and classified websites are the most commonly used marketing tools. Social media’s role in targeting customers is growing. Modern visualization techniques (e.g., virtual tours, drone photos) are underused. Low-quality property photos and descriptions reduce attractiveness.
8	[24]	2021	Development of real estate marketing—trends for the future	To explore the new potential of digital channels for real estate marketing and their impact on reach and intensity.	Real estate marketing in the Cologne/Bergisch-Gladbach market.	Perspective of real estate agents with a focus on digital marketing channels.	Digital marketing channels, online platforms, social media, video streaming platforms, property branding, international sales.	Digital technologies have fundamentally changed real estate marketing, offering more marketing channels and presentation options. Social media and video streaming platforms are expected to further revolutionize property marketing, enhancing property branding and international sales potential.
9	[14]	2022	Construction of Real Estate Network Marketing System based on mobile Internet	To design and develop a real estate marketing system based on mobile Internet and Python technology to support online marketing and improve sales strategies.	Real estate marketing in the digital age, using web-based platforms and technologies.	Design and development of a real estate marketing system using Python (version 3.7.3) technology, mobile Internet, web adaptation, and webVR for virtual reality viewing.	Sales performance, consumer behaviour records, user needs for housing information, virtual reality viewing, system data collection.	The system supports multi-device login, collects reliable data on sales and consumer behaviour, and uses webVR for virtual housing viewing. It improves communication between customers and sales teams, enhances efficiency, and reduces costs.
10	[28]	2021	Advertising through UAVs: Optimized path system for delivering smart real-estate advertisement materials	To explore innovative advertisement methods for transforming the real estate sector into “smart real estate,” focusing on a UAV-based system for delivering advertisement materials.	Kingsford suburb, Eastern Sydney; the real estate industry in Australia.	Systematic literature review of 58 articles on real estate advertisements, followed by the proposal of a UAV-based smart advertising delivery system with optimized paths using four Java-run algorithms (greedy, interroute, intraroute, and Tabu).	58 articles reviewed for literature analysis; six case studies of real estate advertising delivery (three for rent, three for sales) in the Kingsford suburb.	Real estate advertising methods (print media, electronic media, mixed methods), UAV-based delivery system, optimization algorithms (greedy, interroute, intraroute, Tabu), and performance metrics (advertisement delivery efficiency).
11	[21]	2021	Improving Information Efficiency in Housing Markets Using Online Listings	To examine the effects of online listing portals on the success rates of sales and the performance of real estate agents in Singapore’s private housing market.	Private housing markets in Singapore, focusing on the changes brought about by online listing portals after 2013.	The study merges online listing data with property transaction data to assess the impact of online listings on sales success and agent performance. The analysis is based on real estate transaction and listing data.	Success rates in converting listings into sales, agent performance (price premiums, time on market), online listing adoption, portal selection, writing style, and effort in listings.	Online portals increase the success rate in converting listings to sales. Real estate agents using online listings achieve an average premium of 3.26% on sales prices, which remains significant at 1.98% when controlling for property and location attributes. Online listings improve agent performance, reducing market time and increasing price premiums through better effort, writing style, and portal selection. The use of online listings reduces friction in the housing market, yielding an estimated economic benefit of approximately S$1.71 billion (US$1.18 billion).
12	[20]	2021	Can Web Search Queries Predict Prices Change on the Real Estate Market?	To explore whether the intensity of internet searches, as indicated by Google Trends Search Volume Index (SVI), can predict changes in real estate prices.	Polish real estate market, aiming to evaluate the predictive capabilities of Google Trends data in this specific context.	The research combines cross-correlation analysis with machine learning classification algorithms. It applies a multi-class HPI values classifier to predict housing price increases, decreases, or stability based on Google search data related to ‘real estate’ and ‘credit to buy real estate’.	Google Trends Search Volume Index (SVI) for real estate and credit-related searches, housing price indices (HPI), and the directional shift in housing prices.	The study confirms that Google search data (SVI) can provide valuable predictive insights into future housing price changes (increase, decrease, or stability). SVI is associated with real estate price movements, offering a time-lagged predictor that can inform property purchase or sale decisions. The approach demonstrates the extra predictive value of including Google search volume data in models based on traditional housing price determinants.
13	[13]	2022	Online communities and offline sales: considerations on visiting behaviour dimensions and online community types	To examine how consumer visiting behaviours in online communities influence offline purchasing behaviour, with a focus on the moderating roles of visiting depth and breadth.	online real estate platform and its effects on offline housing sales.	Empirical data was collected on consumer visits to an online real estate platform and local housing sales. The study uses regression analysis in Stata 17.0 to test the relationships, with additional robustness checks.	Consumer visiting behaviours (depth and breadth), visits to online communities (competitive and complementary), offline sales of newly built housing.	Consumer visits to an online real estate community positively influence offline housing sales. Visiting breadth negatively moderates the relationship between online visits and offline sales, while visiting depth shows no significant moderating effect. Visits to competitive online communities positively affect offline sales of focal products, while visits to complementary communities have no significant impact.
14	[11]	2022	Impact of Digital Marketing on Customer Buying Behaviour (Residential Real Estate)	To examine the impact of digital marketing on customer buying behaviour in residential real estate, focusing on the role of digital channels in purchase decisions.	residential real estate, specifically for customers or leads of Vue Residences.	The study is based on a survey of 148 respondents out of 180+ distributed questionnaires. A rating scale was used to assess customer perceptions and attitudes.	Digital marketing channels, customer buying behaviour, site visit bookings.	Customers are aware of digital marketing and are influenced by it in their purchase decisions. A preference for booking site visits through digital channels was observed among customers.
15	[14]	2022	Construction of Real Estate Network Marketing System based on mobile Internet	To design and develop a real estate marketing system that leverages mobile internet and Python technology, addressing the current shifts from traditional offline marketing to online marketing in the real estate industry.	the real estate industry’s marketing transition and user needs for housing information.	The paper proposes a system based on mobile internet and Python technology. It integrates features such as webpage adaptation for different devices and the use of webVR technology for virtual reality housing views.	Sales performance, consumer behaviour records, virtual reality viewing, communication efficiency.	The proposed system improves communication between customers and sales agents by transitioning interactions online, increasing efficiency and reducing costs. The integration of webVR technology enhances consumer understanding of housing information through virtual reality.
16	[25]	2023	Research on the Impact of the Internet Economy on the Development of the Real Estate Industry	To examine how the Internet economy, through its open platform and integration with e-commerce, is influencing the real estate market by enhancing data exchange, labour productivity, and technological advancements.	integration of real estate and e-commerce, particularly in the context of commercial housing development and buyer consulting services.		Data exchange, labour productivity, e-commerce integration, commercial housing promotion, buyer consulting services.	The Internet facilitates successful data exchange, providing comprehensive and in-depth information. Technological advancements and increased labour productivity have had a significant impact on the real estate market. The real estate sector, influenced by the Internet economy, is becoming more integrated with e-commerce, providing new opportunities for growth and development. Developers can more efficiently promote commercial housing and offer consulting services to buyers.
17	[17]	2021	Research and Analysis of a Real Estate Virtual E-Commerce Model Based on Big Data Under the Background of COVID-19	To explore the impact of big data and e-commerce on the real estate industry and to design a more effective virtual e-commerce model for the sector based on big data technology.	Focuses on the real estate industry and its transformation through big data, e-commerce, and virtual communities.	Conceptual design and analysis of a virtual e-commerce model based on big data, with a focus on data collection, analysis, and processing through virtual platforms to understand customer needs and enhance sales.	Big data, e-commerce, virtual communities, customer needs, real estate sales model.	The real estate industry, despite being a significant economic pillar, still operates on traditional sales models and lacks an effective e-commerce model. The advent of big data and IT breakthroughs has opened up new possibilities for the real estate sector to adopt e-commerce models. A virtual e-commerce model, combining e-commerce platforms and virtual communities, can accurately capture customer needs and offer more reasonable options to improve sales. The model designed in this paper is considered more effective and can contribute to the healthy development and sales reform of the real estate industry.
18	[29]	2018	Impact of Social Media on Real Estate Sales	To investigate the impact of Facebook activities on real estate sales, including the effect of time lags between Facebook engagement and transactions.	real estate industry, particularly focusing on Facebook as a tool for promoting sales.	The study analyses Facebook activities (Likes, links, stories, sentiment) and their relationship with real estate sales. Time lags between social media activity and transactions are also considered in the analysis.	Facebook Likes, links, stories, sentiment score of posts, time lag, real estate sales.	Facebook activities, including Likes, links, and stories, are positively associated with real estate sales. The sentiment score of Facebook posts is negatively associated with real estate sales. Time lag plays a significant role in how Facebook activities impact real estate sales, suggesting that the effect may not be immediate.
19	[15]	2022	Internet traffic and firm performance in big-ticket sectors: there are two sides of the coin	To explore the relationship between internet traffic on company websites and their sales volumes and profitability, particularly in the context of the COVID-19 pandemic.	Russian firms in the car manufacturing and real estate development industries during 2017–2022, a period marked by the COVID-19 pandemic.	Panel data regression analysis and the Prais-Winsten regression model with panel standard error adjustment are applied to a sample of 268 observations from 67 firms.	Internet traffic, sales volume, profitability, firm age (younger firms), website content.	Internet traffic positively impacts firm sales volumes. The positive effect of internet traffic on sales is accompanied by a negative impact on profitability. The negative effect on profitability is more pronounced for younger firms, suggesting they may prioritize sales over profitability. Younger firms facing resource constraints may need to focus on cost-effective strategies, including website content improvements, to build customer trust.
20	[22]	2019	The Impact of Internet Real Estate Intermediary Platform on the Real Estate Market	To examine the impact of e-commerce platforms on the real estate market, particularly their influence on housing prices and transaction fluctuations.	real estate market in Beijing, China, particularly in response to the People’s Bank of China’s interest rate cut policy in October 2015.	The paper uses a Difference-in-Differences (DID) model to analyse real estate transaction data from Beijing (2015–2016) and policy impacts.	Housing prices, transaction fluctuations, e-commerce real estate intermediary platforms, traditional intermediaries, interest rate cut policy.	The introduction of internet trading platforms significantly reduced housing price increases and transaction fluctuations compared to traditional intermediary methods. After analysing the data using the PSM-DID method, the results confirmed that e-commerce platforms help stabilize housing prices, showing no significant difference from the initial DID results.
21	[30]	2025	Deep Learning based Seasonality and Trend Detection in Sales Forecasting	To develop a Hybrid Deep Learning (HDL) model combining CNN and LSTM to improve sales prediction accuracy.	Commercial (sales forecasting)	Time-series modeling using CNN, LSTM, and hybrid CNN-LSTM models; seasonal decomposition; sliding window for data structuring	Sales volume, seasonality, trend components, model performance metrics (RMSE, MAE, MAPE)	The hybrid CNN-LSTM model outperformed individual models with RMSE of 2171.38, MAE of 1219.79, and MAPE of 538.18, effectively capturing patterns and trends.
22	[31]	2025	AI integrated framework for stability of functions in persistent homology	To identify factors influencing stability in persistent homology functions and propose an AI-integrated framework for stability.	General AI systems with cyber threat applications	Topological Data Analysis (TDA); analysis of point cloud data; examination of function stability in noisy datasets	Triangularability, topological invariance, stability functions, accuracy in cyber-threat detection	Stability of persistent homology functions depends on properties like triangularability; AI framework improves stability and accuracy in cyber-threat contexts.

2.2. The Significance of Real Estate Websites in Real Estate Marketing and Sales

Real estate websites play a crucial role in modern marketing strategies by leveraging digital technologies to enhance the visibility, appeal, and accessibility of properties, thereby transforming how real estate is marketed and sold [24]. Studies [15,21,23] show that increased online activity on business websites directly influences buyer interest, engagement, and ultimately, sales and overall market performance.

• Enhanced Visual and Interactive Experience

Real estate websites provide high-quality images, videos, drone shots, and 3D walkthroughs, which create compelling visual narratives of properties. These elements help potential buyers visualize the space and its potential, making the house-hunting experience more engaging and informative [27,32]. Advanced technologies like virtual reality (VR) and augmented reality (AR) are increasingly used to offer virtual tours, allowing users to explore properties remotely, which is particularly beneficial in the current digital age [14].

• Broader Reach and Accessibility

Real estate websites serve as comprehensive online platforms that provide 24/7 accessibility to property listings, making it easier for a diverse audience to explore real estate options [10,26]. These online platforms enable real estate agents and developers to reach a wider audience beyond geographical limitations. Social media and search engine optimization (SEO) are critical tools for increasing visibility and driving traffic to real estate websites [7,9,29]. The use of mobile applications and responsive web design ensures that property listings are accessible on various devices, enhancing user convenience and engagement [14].

• Data-Driven Marketing Strategies

Real estate websites collect valuable data on user behaviour and preferences, which can be analyzed to tailor marketing strategies and improve customer targeting. This data-driven approach helps in creating more personalized and effective marketing campaigns [8]. These data insights can be further enhanced by integrating deep learning models to uncover patterns in user engagement over time, enabling more accurate forecasting and personalization strategies [30]. The integration of algorithms and interaction design on these platforms provides users with customized property recommendations, enhancing the overall user experience and increasing the likelihood of successful transactions [13].

The following table provides a synthesis of key academic studies that inform the broader research landscape on digital real estate marketing. These foundational works explore critical themes such as the strategic framing of property listings, the use of online marketing tools, and evolving consumer engagement patterns. The findings summarized below in Table 1 below offer valuable theoretical and methodological insights relevant to understanding the dynamics of web traffic and visitor sentiment in residential real estate markets.

• Social Media and Online Advertising

Social media platforms are extensively used for real estate marketing, allowing agents to engage with potential buyers through targeted ads and interactive content. This approach not only increases brand awareness but also fosters a sense of community and trust [9,29]. Online advertising, including pay-per-click (PPC) campaigns and classified ads, remains a dominant method for promoting real estate listings. These ads are often complemented by high-quality visual content to attract and retain user interest [32,33].

• Adaptation to Market Trends

The real estate market is continuously evolving, and digital marketing strategies must adapt to these changes. For instance, the COVID-19 pandemic has accelerated the shift towards online marketing, with real estate agents becoming more creative in their digital outreach efforts [29]. The need to cater to younger, tech-savvy customers has driven the adoption of new media sources and technologies, ensuring that marketing efforts remain relevant and effective [7].

Real estate websites have, therefore, revolutionized the marketing landscape by providing enhanced visual experiences, broader reach, data-driven strategies, and effective use of social media and online advertising. These platforms are essential for modern real estate marketing, offering innovative tools and methods to attract and engage potential buyers and sellers.

• Increased Sales Success Rates

Online listing portals have been shown to increase the success rate of converting listings into sales. This is attributed to the reduced market frictions and enhanced information efficiency provided by these platforms [21].

• Price Premiums

Properties listed online tend to sell at a premium. In Singapore, real estate agents using online listings achieved an average price premium of 3.26%, which remained significant at 1.98% even when controlling for property and location attributes [21].

• Enhanced Agent Performance

The use of online listings has been linked to improved performance of real estate agents, including higher price gaps and shorter time on the market. This improvement is driven by better effort, information presentation, and portal selection [21].

• Predictive Power of Web Searches

The intensity of internet searches, as measured by Google Trends, can predict changes in real estate prices. This predictive power can help in making informed decisions regarding property purchases or sales [20,25].

• Impact of Online Communities

Consumer visits to online real estate communities positively impact offline sales. However, the breadth of visiting different online communities can have a negative moderating effect, while the depth of visits does not show a significant moderating effect [13].

• Digital Marketing Influence

Digital marketing significantly influences customer buying behaviour in residential real estate. Customers are increasingly using digital channels to book site visits and make purchase decisions, indicating a strong preference for digital engagement in the buying process [11].

Web traffic and online activities, therefore, have a profound impact on residential real estate sales. Online listings increase sales success rates and property prices, while digital marketing and online communities play crucial roles in shaping consumer behaviour and purchase decisions. Additionally, web search trends can serve as valuable predictors for real estate market changes, aiding in strategic decision-making.

2.3. Sentiment Analysis in Digital Marketing

Sentiment analysis, or opinion mining, has emerged as a transformative force in digital marketing, revolutionizing how businesses interpret and respond to customer sentiments. This approach relies on advanced computational linguistics and machine learning techniques to distil subjective information from text, categorizing sentiments into positive, negative, or neutral. Refs. [27,34] elucidate the principles of natural language processing (NLP), showing how algorithms discern emotional nuances in written language. In digital marketing, sentiment analysis serves as a strategic asset, providing insights into consumer sentiments. Ref. [9] details how applying sentiment analysis to customer reviews, social media interactions, and textual content enables marketers to understand audience perceptions and tailor their strategies accordingly.

A key application of sentiment analysis lies in tracking brand perception. Ref. [26] emphasize its significance in gauging public sentiments toward a brand, guiding marketers in maintaining a positive image and addressing potential issues. The role of sentiment analysis in social media listening strategies is paramount. Ref. [29] highlight that real-time monitoring and analysis of social media sentiments allow marketers to respond swiftly to emerging trends, ensuring agility in adapting marketing strategies. Furthermore, understanding the sentiment associated with various content types informs content strategy optimization [35]. Refs. [8,35] demonstrate how sentiment analysis assists marketers in identifying content that resonates positively with audiences, guiding content creation and distribution.

Sentiment analysis is instrumental in analyzing customer feedback. Ref. [23] emphasizes that categorizing sentiments in customer reviews helps businesses identify strengths and weaknesses, facilitating targeted improvements in products or services. Additionally, sentiment analysis extends to competitor analysis. Ref. [13] show how businesses can gain insights into consumer perceptions of competitors, informing benchmarking strategies and identifying unique selling points for strategic positioning in the market. As a dynamic analytical tool, sentiment analysis empowers marketers to navigate the emotional landscape of their audience, enabling data-driven decision-making and fostering proactive engagement.

2.4. Data-Driven Content Optimization

Data-driven recommendations involve systematically analyzing quantitative and qualitative data to derive actionable insights [15]. In real estate website content, these recommendations are tailored suggestions based on empirical evidence and patterns identified through comprehensive data analysis. In the dynamic landscape of digital marketing, data-driven recommendations are crucial for optimizing real estate website content to enhance web traffic [24]. Understanding target audience preferences and behaviours is fundamental [7]. Data-driven recommendations enable real estate companies to tailor content effectively for specific demographic groups [10]. By identifying patterns in user interactions, companies can create content that aligns with potential clients’ interests and needs. Advanced AI-based approaches, including persistent homology and deep learning, are increasingly being explored to analyze unstructured and temporal web activity data. Such models offer a robust framework for tracking user sentiment and behaviour across digital real estate platforms [30,31] search engine algorithms play a pivotal role in determining the visibility of real estate websites [14]. Emerging AI-integrated models provide a new layer of precision in optimizing for search visibility by incorporating both structural and behavioural web metrics [31]. Data-driven recommendations assist in optimizing content for search engines, ensuring that articles and property listings are strategically crafted to enhance organic search rankings [22]. This involves keyword analysis, content structuring, and other SEO strategies based on data-driven insights. User experience also critically influences web traffic and engagement [25] by analyzing user behaviour data, real estate companies can make informed recommendations to enhance the overall user experience [17], including optimizing navigation, improving load times, and creating a seamless browsing experience.

Real estate markets are dynamic, and consumer preferences evolve [21]. Data-driven recommendations empower companies to adapt content strategies in response to changing trends and market demands [32]. Whether shifting focus between article topics or adjusting emphasis on specific property features, these recommendations ensure content remains relevant and impactful. Metrics such as click-through rates, time spent on pages, and social media shares indicate content engagement [20]. Data-driven recommendations guide real estate companies in refining their content to boost these metrics [11] through strategies like A/B testing of headlines and experimenting with multimedia content.

The integration of data-driven recommendations in content optimization is instrumental in navigating the competitive landscape of real estate digital marketing [21]. By harnessing the power of data, companies can attract increased web traffic and cultivate a more meaningful online presence in residential real estate marketing [21]. In the competitive realm of digital marketing within the real estate sector, implementing best practices in content optimization is vital for attracting and retaining web traffic [21,36]. Strategic keyword use is integral to content optimization [32]. Keyword research, as demonstrated by [24], enables real estate companies to align content with terms potential clients use during searches, enhancing website discoverability and organic traffic.

Crafting compelling headlines and meta descriptions is endorsed by content marketing experts [26]. These elements improve search engine ranking and entice users to click on links, driving traffic to websites [27]. Optimizing these components for relevance and clarity aligns with user expectations, encouraging higher click-through rates. Visual content, including images, infographics, and videos, significantly contributes to content optimization [8]. High-quality visuals enhance user engagement and dwell time, recognized factors in search engine algorithms [24]. Recent studies [8,11] show that real estate websites featuring visually appealing content are more likely to attract and retain visitors.

Ensuring the responsiveness of real estate websites across devices, particularly mobile devices, is fundamental [14]. Mobile optimization caters to the growing number of users accessing websites through smartphones and is acknowledged by search engines as a ranking factor [25]. Adopting a mobile-first approach enhances the user experience, positively impacting web traffic. Personalizing content based on user preferences resonates with contemporary digital marketing strategies [9]. Utilizing data-driven insights to tailor content to target audience interests enhances relevance and encourages return visits [13]. Personalization contributes to increased engagement and web traffic.

Incorporating interactive elements, such as quizzes, polls, and calls-to-action (CTAs), encourages user participation [32]. Engaged users are more likely to explore multiple pages, positively impacting dwell time and overall web traffic [15]. Strategically placed CTAs guide visitors through the conversion funnel, increasing the likelihood of meaningful interactions. Regularly updating content is crucial [26]. Search engines favour fresh and relevant content, leading to improved rankings [32]. Studies [8,24,26] demonstrate that real estate websites consistently providing valuable and up-to-date information experience sustained increased web traffic.

Implementing best practices in content optimization forms a holistic approach to enhancing web traffic for real estate websites. By combining keyword strategies, engaging visuals, mobile optimization, personalized content, interactive elements, and maintaining relevance through regular updates, real estate companies can create an online presence that effectively attracts, retains, and converts visitors. The landscape of real estate marketing has shifted, with digital platforms central to outreach strategies. A review of existing studies reveals the multifaceted relationship between content optimization strategies and increased web traffic for real estate websites. Research by [21] underscores the pivotal role of content optimization in enhancing search engine visibility. The study reveals a direct correlation between strategic keyword use, meta descriptions, and improved search rankings, leading to increased organic web traffic.

The impact of visual appeal on user engagement has also been investigated [27]. The study illustrates how high-quality images and visually appealing content contribute to extended engagement, consequently driving higher web traffic. Real estate websites incorporating such visuals witness a significant increase in their visitor base. Regular updates and fresh content are imperative for maintaining relevance and attracting returning visitors [26]. Consistently refreshed content signals to search engines that the website is active, which enhances visibility and attracts ongoing web traffic.

2.5. IT Governance in Developing Economies and Digital Real Estate in South Africa

In developing economies like South Africa, IT governance plays a critical role in structuring and enhancing digital real estate platforms. IT governance frameworks offer mechanisms to manage risks, ensure compliance, and optimize digital platform operations, all essential for meeting the unique demands of emerging markets [26,31]. South Africa’s real estate market, shaped by its local economic context and technological adoption rates, faces distinct challenges in web traffic and user engagement dynamics. The high levels of inequality and income diversity within South Africa create a varied consumer base, which is further influenced by disparities in technological infrastructure across regions, affecting digital platform access and engagement [15,23].

IT governance frameworks such as COBIT (Control Objectives for Information and Related Technologies) and ITIL (Information Technology Infrastructure Library) have been adapted in various sectors in South Africa to address challenges in technology-driven industries, including real estate. These frameworks emphasize data security, compliance with local and international regulations, and customer-centric policies, which collectively enhance user trust and engagement on digital platforms [17,24]. Ref. [17] underscores the importance of robust data protection and transparent policies in fostering customer loyalty and boosting platform traffic in digital real estate markets.

Operational efficiency, security, and content quality are central components of IT governance in digital platforms, significantly impacting user retention and traffic. A seamless user experience, which involves enhanced security protocols, optimized site performance, and easy navigation, is vital for attracting and retaining diverse users in the South African real estate market [8,14]. These measures are crucial for real estate platforms in managing online user interactions, particularly in contexts with varying levels of digital literacy and access [31].

Furthermore, applying global IT governance frameworks to fit local economic and technological realities can position South African digital real estate platforms to better address their user base’s needs. This approach provides a competitive edge in the digital marketplace, enabling platforms to navigate regulatory landscapes while promoting equitable access and a trustworthy online environment [15,25]. As displayed in Figure 1, addressing local challenges such as economic inequality, urban-rural digital divide, reliance on mobile internet, and low levels of digital literacy, is necessary for enhancing inclusive engagement. Research by [31] highlights that customized IT governance strategies support not only compliance and security but also address specific market challenges like inconsistent internet access, thus fostering sustainable digital engagement. Incorporating deep learning techniques and AI-driven governance frameworks can further enhance predictive capabilities and platform resilience, especially when tailored to local infrastructural realities [30,31].

Figure 1. Local factors shaping web traffic and implications for IT Governance.

Figure 1 highlights key local factors shaping digital real estate engagement in South Africa and their implications for IT governance. Economic inequality calls for content and pricing strategies that serve both premium and budget users. The urban–rural divide demands mobile reach and offline accessibility, while the dominance of mobile internet use requires a mobile-first design with low data usage. Varying digital literacy levels necessitate intuitive interfaces and embedded learning tools. Informal property transactions highlight the need for trust-building and verification systems, and compliance with POPIA requires strong data protection and transparent consent policies. Together, these factors underline the need for localized, adaptive IT governance to ensure inclusive and sustainable digital real estate platforms.

2.6. Theoretical Framework

In this study, two key theoretical frameworks are used to explore how website content sentiment impacts web traffic in real estate marketing: Information Foraging Theory and the Elaboration Likelihood Model (ELM).

Information Foraging Theory, proposed by [37] views users as adaptive foragers seeking valuable information, much like animals searching for food. In the context of real estate websites, this theory is relevant as it explains how users are drawn to different types of content, such as property listings or articles, based on perceived value. The theory provides insights into how content attributes can be optimized to guide users through the site more effectively, thereby increasing engagement and web traffic.

The Elaboration Likelihood Model (ELM), developed by [38], distinguishes between two routes of persuasion: central and peripheral. In this study, sentiment serves as a peripheral cue that influences user engagement. The model’s relevance lies in its ability to explain how the emotional tone of real estate listings or articles affects users’ cognitive responses and behaviours. By understanding how sentiment impacts users’ decision-making processes, the study can offer strategies for improving content that drives higher traffic to real estate websites. Both theories provide a comprehensive lens through which to analyze and optimize the impact of content on user engagement and web traffic.

2.7. Conceptual Framework

Figure 2 presents a conceptual framework illustrating the hypothesized relationship between web content features and digital engagement, specifically web traffic. The model proposes that two independent variables, Web Content Sentiment and Content Category influence the dependent variable, Web Traffic.

Figure 2. Conceptual Diagram.

Hypothesis H₁ suggests that the sentiment expressed in online content (positive, neutral, or negative) has a significant effect on the amount of traffic the web page receives. This could relate to emotional tone, user perception, and content relatability.

Hypothesis H₂ posits that the category of content, such as news, promotional listings, or informational articles also plays a key role in attracting visitors and shaping online behaviour.

3. Methodology

This section outlines the research methodologies, encompassing research design, data collection, and analysis, aimed at understanding the influence of website content sentiment on web traffic within the residential real estate sector.

3.1. Research Design

A quantitative approach was adopted, leveraging statistical tools to discern patterns, trends, and correlations in large datasets. This approach is justified by its capacity to provide an objective perspective on the relationships being examined [20]. Specifically, the study employed various statistical techniques, including chi-square tests, linear regression, and analysis of variance (ANOVA), to examine the relationship between sentiment and content type on web traffic. ANOVA was selected to determine statistically significant differences in web traffic among different content categories, allowing a detailed assessment of how various types of content, such as property listings and articles, affect website engagement. Chi-square tests were utilized to analyze the association between content sentiment and category type, aiding in understanding distribution patterns. Linear regression analysis quantified the influence of content sentiment on total monthly visitors, providing insights into how sentiment impacts web traffic over time. These methods are well-established in sentiment analysis and web traffic studies for their capacity to handle large datasets and accurately model relationships between categorical and continuous variables [15,30].

3.2. Data Collection and Pre-Processing

Data was collected using R (version 4.5.1) through web scraping from prominent real estate websites, such as Property24, Remax, and Private Property, from 30 June 2017 to 30 June 2023. This six-year timeframe was selected to capture evolving trends in web traffic and sentiment in the South African real estate market, ensuring that the dataset reflects contemporary behaviours and patterns [22]. Scraping was performed using R packages such as rvest and httr, employing versions that were fully compatible with R version 4.5.1 at the time of analysis, which enabled automated extraction of data from structured HTML elements.

The dataset included four key variables:

• Title: the headline or description of a listing or article;
• Category: the content type (e.g., “Property Listing”, “Real Estate Blog”, “News Update”);
• Sentiment: the tone of the content (classified as Positive, Neutral, or Negative);
• Total Monthly Visitors: a numeric field representing the estimated monthly traffic for the specific page or content piece, derived from SimilarWeb, a site analytics tool.

Content categories were assigned based on the context and structure of each webpage or article. For instance, pages containing specific property features were categorized as Property Listings, while longer-form editorial content about buying tips or market trends was grouped under Articles.

For sentiment classification, they were as follows:

• Text data (title + description body) was processed using the textdata, tidytext, and sentimentr packages in R.
• Sentiment scores were calculated based on polarity dictionaries and tokenized text, and then mapped into three broad categories:

  ο

  Positive (e.g., “charming”, “spacious”, “secure”);

  ο

  Neutral (e.g., “located in the heart of the city”, “three bedrooms”);

  ο

  Negative (e.g., “requires renovation”, “limited access”).

• This categorization was validated manually for a sample of entries to improve label accuracy and account for contextual variations in real estate language.

Data preprocessing involved the following steps:

• Removing duplicates and null values;
• Normalizing text (lowercasing, removing HTML tags, punctuation);
• Handling outliers (especially extremely high or low visitor counts);
• Encoding categorical variables numerically for modelling purposes.

This rigorous preprocessing ensured high data quality and model-readiness for statistical analysis.

3.3. Analytical Framework and Tools

The analysis followed the CRISP-DM (Cross-Industry Standard Process for Data Mining) framework, a structured and widely used approach in data mining projects [39]. The CRISP-DM framework’s phases—Business Understanding, Data Understanding, Data Preparation, Modelling, Evaluation, and Deployment provided a comprehensive approach to handling the dataset systematically. In line with approaches by [32], this study employs sentiment analysis techniques to evaluate user engagement with digital real estate content. This approach facilitated a deeper understanding of the data, aiding in the identification of relevant patterns and trends.

In the Modelling phase, several R packages were utilized to enhance model performance and ensure robust statistical analysis aligned with the chosen techniques of chi-square tests, linear regression, and analysis of variance (ANOVA). These packages include the following:

• stats: This base R package provides functions for performing core statistical techniques, including chi-square tests, linear regression (lm function), and ANOVA (aov function). These tools are fundamental for examining the relationships among sentiment, content type, and web traffic.
• car: The Companion to Applied Regression package (version 3.1-2) is used for diagnostic tests and enhancements in linear models. It aids in performing ANOVA analyses with Anova () and enables post hoc tests, which are useful for understanding detailed differences in content type and sentiment categories.
• MASS: Provides robust linear regression tools and diagnostics for assessing model assumptions. It is particularly valuable for handling potential outliers or heteroscedasticity issues that may arise in web traffic data analysis.
• caret: Used for managing model training and validation processes, particularly helpful for setting up the data into training and testing sets, optimizing model parameters, and performing cross-validation, which ensures the reliability of linear and ANOVA models in predictive analysis.
• sentimentr, tidytext, and textdata: Applied during the sentiment analysis stage to score and classify textual data into sentiment categories.

The choice of these tools was informed by their efficacy in statistical computing, and their specific suitability to regression tasks commonly used in sentiment analysis and web traffic modelling. This analytical framework allowed the study to accurately assess the impact of sentiment and content type on web traffic, contributing valuable insights into user engagement dynamics in the digital real estate market.

4. Results

The study explores the relationship between sentiments on web content of real estate websites and the traffic it attracts. This chapter presents the analysis of the data in line with the objectives.

4.1. Data Exploration

4.1.1. Analysis of Missing Values

The dataset’s “Total Monthly Visitors” variable was assessed for missing values using the is.na function, and the result indicated that there are no missing values. The output “Number of missing values: 0” affirms that the column is complete, with no instances of missing data, and each observation includes a numeric value for the total monthly visitors.

The “Category” and “Sentiment” variables were converted to factors, and the dataset was checked for missing values using the any(is.na()) function. The result, denoted by FALSE, signifies that each observation in these categorical variables has been successfully converted, and there are no missing values in the entire dataset after these conversions.

The absence of missing values in the key variables is a positive indication of data completeness. The successful conversion of categorical variables to factors further ensures the dataset’s readiness for subsequent analyses or visualizations. This robustness in data quality contributes to the reliability of any findings or insights drawn from the dataset.

4.1.2. Summary Statistics

The summary statistics for the key variables have been compiled and presented in Table 1. This table encapsulates essential information about the dataset, providing insights into its characteristics and central tendencies.

The “Title” variable, representing the titles of web pages, has a length of 14,907 entries. The data type is characterized as a character, emphasizing the diverse and textual nature of the titles. The “Category” variable, indicating the category types of the titles, also has a length of 14,907 entries. Similarly to “Title,” the data type is a character, highlighting the categorical and diverse nature of the content. In the case of the “Sentiment” variable, reflecting the emotional tone of the titles and content, the length aligns with the overall dataset at 14,907 entries. The data type remains character, emphasizing the qualitative nature of sentiment categorization. The “Total Monthly Visitors” variable, quantifying the number of monthly visitors to each web page, is characterized by numeric values. The summary statistics reveal the minimum monthly visitors as 105, the maximum as 3,017,000, with a mean of approximately 2,279,380. This numeric representation offers a quantitative perspective on web traffic across the dataset.

4.1.3. Traffic Summary per Category

Table 2 presents a summary of the total monthly visitors based on the category of the titles, revealing key patterns in web traffic for real estate websites.

Table 2. Total Monthly visitors per Title category.

Category	Total Monthly Visitors
Article	13,407,521
Property Listing	16,780,623

Source: Data Analysis Results from [40].

Property listings attract significantly higher traffic, with a total of 16,780,623 monthly visitors compared to articles, which receive 13,407,521 visitors. This indicates that property listings are more effective in driving web traffic compared to articles, highlighting the preference of users for direct property-related content over informational articles.

4.2. Impact of Sentiment of Web Traffic

The aim of this analysis was to assess the impact of sentiment (Positive, Neutral, Negative) on the number of total monthly visitors to real estate websites in South Africa from June 2017 to June 2023. Table 3 below presents the ANOVA results.

Table 3. ANOVA Results for Sentiments and Web Traffic.

	Df	Sum Sq	Mean Sq	F Value	Pr(>F)
Sentiment	2	1.677 × 1014	8.384 × 1013	103.9	<2 × 10−16 ***
Residuals	14,904	1.203 × 1016	8.072 × 1011

Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1. Source: Data Analysis Results from [40].

The ANOVA results shown in Table 3 indicate that sentiment has a statistically significant effect on total monthly visitors to real estate websites, as shown by the high F-value (103.9) and an extremely low p-value (<0.001). The asterisks (***) next to the p-value (<2 × 10⁻¹⁶) for Sentiment indicate a highly statistically significant finding. These findings allow us to reject the null hypothesis, which posited that sentiment in real estate listings has no impact on web traffic. Therefore, we conclude that the sentiment of real estate listings significantly influences user engagement on digital platforms, with different sentiments (positive, neutral, and negative) attracting varying levels of visitor traffic.

This finding is consistent with previous studies, such as those by [15,29], which found that positive content generally enhances user engagement across digital platforms. Additionally, ref. [17] observed that sentiment in e-commerce platforms can drive user behaviour and engagement, a result mirrored in the South African context, where users respond meaningfully to sentiment in real estate listings.

In global digital behaviour literature, positive sentiments are often linked to increased engagement, as seen in sectors like e-commerce and social media [11,13,29]. Interestingly, South African users also respond to negative sentiment, potentially as a form of information-seeking behaviour. This nuanced reaction may reflect the unique socioeconomic environment in South Africa, where consumers engage with content not only to identify attractive opportunities but also to remain informed about market conditions [22].

4.3. Impact of Content Category on User Engagement for Real Estate Websites

An analysis of variance (ANOVA) was conducted to evaluate the effect of category (article or property listing) on the total monthly visitors to real estate websites. The model focused on the category as the explanatory variable, with residuals accounting for the unexplained variance in the model.

The high F-statistic of 426.9 and a p-value of less than 0.001 as shown in Table 4 indicate a strong, statistically significant effect of content category on total monthly visitors. These results confirm that the category (whether an article or property listing) significantly impacts user engagement, measured by the volume of web traffic, allowing us to reject the null hypothesis. Therefore, we conclude that content type has a critical role in shaping user interaction with real estate websites. This finding aligns with global research that emphasizes the role of content type in driving user engagement. For instance, previous studies e.g., refs. [10,21] have shown that detailed, product-oriented content, such as property listings, tends to attract more focused engagement than general articles. Specifically, research in the real estate sector by [23] supports the idea that property listings tend to generate higher user interaction, as they directly address the information needs of users in the market. In contrast to regions where real estate content may equally balance informational articles and property listings, the strong preference for property listings in South Africa reflects a unique market orientation toward actionable content. As noted in studies of emerging economies, consumer engagement often prioritizes practical and transaction-related information [9]. This preference suggests that, in South Africa, users primarily seek content that facilitates decision-making rather than general market insights.

Table 4. ANOVA test result for Influence of category listing on web traffic.

	Df	Sum Sq	Mean Sq	F Value	Pr(>F)
Category	1	3.393 × 1014	3.393 × 1014	426.9	<2 × 10−16 ***
Residuals	14,875	1.182 × 1016	7.948 × 1011

Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1. Source: Data Analysis Results from [40].

5. Conclusions

The analysis of sentiment and content category on web traffic for real estate websites in South Africa provides key insights with practical implications for real estate marketers, web developers, and business strategists. First, sentiment was found to significantly influence monthly web traffic, as listings categorized by positive, neutral, or negative sentiments had a measurable impact on user engagement, as evidenced by the significant F-statistic (103.9) and a p-value less than 0.001. This suggests that the emotional tone conveyed through listings plays a crucial role in attracting users, with positive or strong sentiments likely resonating more deeply with users seeking relevant information.

Similarly, the category of content, whether article or property listing, significantly affected user engagement, with property listings consistently attracting higher web traffic than articles. This finding, supported by an F-statistic of 426.9 and an extremely low p-value, points to a clear preference among users for transactional, property-oriented content over general informational articles. The hypothesis that content category does not influence web traffic was, therefore, rejected, emphasizing the importance of tailored content strategies.

In practical terms, these insights suggest several applications for the South African real estate sector. Real estate marketers can enhance web traffic by optimizing listing content to reflect strong, clear sentiment, especially positive tones, which could capture user attention more effectively. Incorporating sentiment analysis tools into content development processes could help tailor messages that emotionally resonate with diverse market segments, from first-time buyers to luxury investors. By analyzing user sentiment and engagement patterns, marketers can create targeted strategies that resonate with audience emotions and needs. Web developers and business strategists can also leverage these findings to structure websites with a greater emphasis on property listings, which are shown to drive higher engagement, rather than relying equally on articles. This could involve prioritizing listing content on landing pages, integrating smarter search and filter functions, and enabling mobile-responsive listing views to meet the high mobile internet usage in South Africa. Developing streamlined, visually appealing listing pages could capitalize on users’ preference for direct property information, thereby improving user experience and engagement rates. Additionally, integrating features that simplify navigation for users with lower digital literacy, and providing offline accessibility options for under-connected regions, could further broaden audience reach and inclusion.

This study contributes to digital marketing research by demonstrating that both sentiment and content category are integral to influencing web traffic on real estate platforms. By showing that emotional tone and content type significantly impact user engagement, it highlights how strategic content structuring can optimize user interactions. For business strategists, these findings present actionable insights for aligning content strategy with user behaviour trends, thereby supporting data-driven decisions on partnership development for investment and content prioritization. Beyond real estate, these insights have broader applications for digital content optimization across industries where sentiment and content type affect user behaviour.

Future research could expand on this study by exploring longitudinal analyses to understand how trends in sentiment and engagement evolve over time in response to market conditions. Additionally, replicating this research in other developing economies would offer comparative insights into the role of sentiment and content type in digital engagement, which could enrich understanding of cultural and economic differences in user behaviour across various regions. This research provides a foundation for further studies on how digital content can be refined to meet user expectations, thereby enhancing engagement and contributing to strategic business growth.