SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations

69 SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations Rubén Alcaraz-Martínez Universitat de Barcelona, Spain https://orcid.org/0000-0002-7185-0227 Mari Vállez Universitat de Barcelona, Spain https://orcid.org/0000-0002-3284-2590 Raquel Escandell-Poveda Universidad de Alicante, Spain https://orcid.org/0000-0002-8398-1873 Diana Bajaña-Cedeño Universitat de Barcelona, Spain https://orcid.org/0009-0008-7395-3028 Alcaraz-Martínez, R., Vállez, M., Escandell-Poveda, R., & Bajaña-Cedeño, D. (2025). SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations. In J. Guallar, M. Vállez, & A. Ventura-Cisquella (Coords). Digital communication. Trends and good practices (pp. 69-84). Ediciones Profesionales de la Información. https://doi.org/10.3145/cuvicom.06.eng 70 SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations Rubén Alcaraz-Martínez; Mari Vállez; Raquel Escandell-Poveda; Diana Bajaña-Cedeño Digital communication. Trends and good practices 1. Introduction In the information age, data-driven journalism has emerged as a powerful instrument for structuring, analysing, and telling stories based on large volumes of data. This journalistic speciality, which combines data analysis techniques with visual narratives, not only helps to unravel the complexity of data but also makes it accessible and understandable to the public (Córdoba-Cabús, 2020). Its evolution has been driven by the rise of digitisation and the availability of open data, marking a transformation in the practice of traditional journalism that is increasingly driven by data (Radcliffe & Lewis, 2019). Although its origins date back to the second half of the 20th century, the modern practice was consolidated in the late 2000s and early 2010s (Ferreras Rodríguez, 2015), driven by the proliferation of available open data, particularly from public administrations, the emergence of new technological tools focused on the analysis of large datasets and the creation of interactive visualisations, alongside a media context in which audiences demanded innovative new narratives. Some of its advances are the result of the investigation into the so-called Panama Papers of the International Consortium of Investigative Journalists (2016), or applications stemming from this approach, such as those published by the Civio Foundation (Dónde van mis impuestos or the Indultómetro) (Ferreras Rodríguez, 2015). The integration of departments or units specialised in data journalism within cybermedia became a reality during the 2010s (Peiró & Guallar, 2013). During the COVID-19 pandemic, data journalism played a crucial role, becoming an essential intermediary between institutions and society. The abundance of statistics and their lack of consistency highlighted the importance of this approach in organising, interpreting, and communicating relevant information (Córdoba-Cabús et al., 2020). Alexander and Vetere (2011) propose a list of parameters to ensure the quality of content in data journalism, such as trust in the data presented, the integration of the surprise factor, transparency in methodological processes, the identification of a leader, the use of accessible language, and the possibility of access to the original data. The work of Córdoba-Cabús (2020) analyses some of the most internationally recognised data visualisations emerging from the journalistic sector, with the aim of synthesising a list of common elements in this journalistic practice. Its variables for analysis include subject matter, narrative style, story types, Abstract Storytelling data visualisations have gained prominence in digital media, enabling the communication of complex information in an accessible and engaging manner. However, their impact and reach on search engines largely depend on their visibility in search results. Based on this premise, this chapter examines the key ranking factors for such information products and proposes a set of best practices and a methodology to evaluate its SEO optimisation through a list of indicators, referred to as SEOVisFrame. Keywords Storytelling data visualizations; Data-driven journalism; Media outlets; Search Engine Optimisation (SEO); Search engine visibility. 71 SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations Rubén Alcaraz-Martínez; Mari Vállez; Raquel Escandell-Poveda; Diana Bajaña-Cedeño Digital communication. Trends and good practices dimension, elements of focus, purpose, method of communication, source, type of analysis performed, types and ratio of visualisations. The quality and effectiveness of these narrative products depend on parameters such as transparency, and the selection and treatment of sources, as well as the appropriateness and innovation of the forms selected to present the information (Córdoba-Cabús, 2020). Since the late 1990s and early 2000s, when both major international and national newspapers introduced their respective digital versions, search engine optimisation (SEO) has become an indispensable tool for journalism, as it directly influences the visibility and accessibility of news content. Given that search engines are the primary access point for finding information on the web, the capacity of journalistic content to rank high in search results is crucial for reaching a wider readership. SEO practices not only enhance the discoverability of articles but also ensure that the content is presented in a way that meets the evolving demands of both algorithms and users. Given the competitive nature of the online media landscape, effective SEO strategies are vital for news organisations seeking to maintain relevance and engage audiences in an increasingly crowded digital environment. The number of specific SEO ranking factors and their relative weight in search engine algorithms remain largely unknown, posing a daily challenge for professionals in the field. This tacit knowledge has been primarily disseminated through a range of specialised books (Maciá Domene, 2019; Enge et al., 2023; Lewandowski, 2023; Vicente et al., 2024), as well as through less academic channels (Semrush, 2024; Dean, 2024) and, to a lesser extent, scientific papers (Chotikitpat et al., 2015; García-Carretero et al., 2016; Lopezosa et al., 2018; Ziakis et al., 2019; Almukhtar et al., 2021), some of which are specifically focused on the online media (Dick, 2011; Giomelakis & Veglis, 2015; Lopezosa et al., 2019; 2020; Giomelakis, 2023). In addition to these sources, the technical documentation published by search engines is crucial, with Google being the most prolific. This documentation covers topics ranging from general search engine fundamentals (Google, 2024a) to more specific aspects, such as site optimisation aligned with Google’s EEAT (Experience, Expertise, Authority, and Trust) guidelines (Google, 2024b) or SEO-focused recommendations for image optimisation (Google, 2024c). The articles published under the umbrella of data journalism are complex products that incorporate multiple forms of data visualisation, such as tables, interactive statistical charts, maps, infographics, and rich media animations, accompanied by narratives that typically form a storyline (Domínguez, 2016). This approach has led to formats such as long-form journalism and techniques like scrollytelling (Seyser & Zeiller, 2018). Therefore, addressing the SEO optimisation of such products requires the analysis of various dimensions (textual and graphic elements), each of which is linked to different ranking factors, as well as their visibility in different search engines (e.g. Google Search and Google Images). The most common way to classify SEO factors is based on the degree of control and influence that can be exerted over them. According to this classification, a distinction can be made between on-page and off-page SEO ranking factors (Lewandowski, 2023). On-page SEO can be further divided into technical SEO and content SEO (Escandell-Poveda et al., 2021), two fundamental yet distinct areas that impact a website’s performance and visibility. 72 SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations Rubén Alcaraz-Martínez; Mari Vállez; Raquel Escandell-Poveda; Diana Bajaña-Cedeño Digital communication. Trends and good practices Figure 1 Classification of ranking factors for storytelling data visualisations. Technical SEO focuses on optimising ranking factors related to page crawling and indexing, mobile compatibility, redirect management, and overall site performance and loading speed (Krstić, 2019; Vicente et al., 2024). Moreover, the use of the HTTPS protocol has been a ranking factor since 2014 (Google, 2014). When implementing HTTPS on a domain, it is important to ensure that all pages, as well as the resources required to render their content, such as images and other elements, are loaded using this protocol. Google (2014) recommends using relative URLs for resources that reside on the same secure domain to ensure the correct protocol. This can also improve performance if absolute URLs are misconfigured (e.g., using HTTP instead of HTTPS). In such cases, browsers can resolve relative URLs more quickly, as they avoid potential redirects. The integration of Schema.org structured data has been another key technical SEO element on many websites since its introduction by Google in 2011. Structured data is a standardized format for providing information about a page and its content. Structured data is implemented using in-page markup on the page to which the information applies. The inclusion of structured data markup enhances news indexing and improves the accuracy of search results. Ambiguity in content interpretation by search algorithms is also reduced (Salem et al., 2025). When a news article page has structured data, Google can also use that information to display a rich snippet for the piece. Although literature findings indicate that ranking remains the most critical factor in perceived relevance, in certain cases, the richness of snippets can capture user attention (Marcos et al., 2015). In other words, these elements can significantly influence users’ behaviour and decision-making (Rodas et al., 2016). Structured data must be representative of the main content of the page, syntactically correct, ensure that the referenced content is visible to users, and comply with the guidelines for each specific structured data feature (Google, 2025a). In this case, NewsArticle structured data is the most 73 SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations Rubén Alcaraz-Martínez; Mari Vállez; Raquel Escandell-Poveda; Diana Bajaña-Cedeño Digital communication. Trends and good practices relevant. However, structured data for images can also be useful, as it enables Google Images to display additional details such as the creator, rights, or credit information. Good page performance is also crucial with websites increasingly being accessed via mobile devices. To assess this, Google introduced the Core Web Vitals (CWV) in 2020, incorporating them into its algorithm as a ranking factor in 2021. These are a set of metrics designed to measure the performance of web pages (Maciá Domene, 2019; Alcaraz Martínez, 2022; Vicente et al., 2024), including the LCP (Largest Contentful Paint), INP (Interaction to Next Paint), and CLS (Cumulative Layout Shift). The LCP, which measures the load time of the largest visible text or image element in the viewport, is an indicator of perceived load speed. INP measures the responsiveness of pages to user interactions (such as clicks or keystrokes). Finally, CLS is a measure of visual stability and describes unexpected layouts shifts in the viewport during page loading. Storytelling through data visualisation presents significant CWV challenges, as it often integrates multiple complex images on a single page. The optimisation of these files (selection of the appropriate bitmap format, compression, dimensions, etc.) or the creation of crawlable, search engine-friendly vector images can be challenging. On the other hand, content SEO aims to meet users’ needs by creating texts and other types of content (such as images, multimedia and interactive content) optimised for relevance based on keywords and the search intent expressed by users (Alcaraz Martínez, 2022). The most important on-page content ranking factors for any page are equally significant for the storytelling data visualizations under study. The inclusion of the keyword in the title meta tag, meta description, URL, h1-h3 headings, or within the body of the page itself with an appropriate density, are fundamental factors (Chotikitpat et al., 2015; Almukhtar et al., 2021; Maciá Domene, 2019; Lewandowski, 2023; Dean, 2024; Google, 2024a; Semrush, 2024). Although no longer as important as it once was, incorporating keywords into the <title> tag remains an important on-page SEO signal. While meta descriptions were historically an important ranking factor, their influence has diminished. In recent years, Google has increasingly generated snippets directly from the page content. However, Google may still use the meta description HTML element if it provides users with a more accurate summary of the page than content extracted directly from the site. All pages should include an <h1> heading tag with content closely aligned to the <title> meta tag. Unlike the <title> meta tag, the <h1> tag does not contribute to the generation of SERP snippets. Consequently, strategies such as incorporating the branded keyword (e.g., the name of the media outlet) are commonly applied in meta titles but not in top-level headings. Optimising the <h1> tag for SEO involves crafting story titles that align with the search terms for which visibility is sought. This does not preclude creativity, as media virality and news trends can sometimes render titles such as The Panama papers highly effective. Within the <main> element of a page, multiple levels of headings can be included, ranging from <h2> to <h6>. From an SEO perspective, those higher in the hierarchy, particularly the <h2> and <h3> tags, carry greater relevance. According to Mueller (2015), coordinator of Google Search Relations, “heading tags in HTML help us to understand the structure of the page.” While a single page may aim to rank for multiple keywords, structuring the content into distinct thematic sections, each with its own specific headings, is a widely adopted SEO practice (Maciá Domene, 2019). 74 SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations Rubén Alcaraz-Martínez; Mari Vállez; Raquel Escandell-Poveda; Diana Bajaña-Cedeño Digital communication. Trends and good practices Images must be optimised for accessibility to crawlers and be easily indexed. Bitmap images should be provided via HTML elements such as <img>, <figure> or <picture>, but never as CSS background images, which is quite common when implementing Parallax effects typical in storytelling. In terms of relevance, keywords should be included in the alternative text, file name, image caption, and in the contextual content (Chotikitpat et al., 2015; Maciá Domene, 2019; Lewandowski, 2023; Semrush, 2024). The alternative text is the most important attribute for providing additional details about an image (Google, 2024c), while also improving accessibility for users who cannot see images on web pages. The primary keyword of the page does not need to be added to all images, and excessive keyword usage (keyword stuffing) should be avoided. Selecting appropriate images to illustrate content is a crucial first step. When done effectively, all images will align with the article’s subject matter, reinforcing its overall relevance and enabling the creation of natural, well-optimised alternative texts that serve both SEO and accessibility purposes (Alcaraz Martínez, 2024). Google recommends using descriptive filenames rather than generic ones, even localising them according to the language of the content (Google, 2024c). Although the exact impact of this factor remains unknown, its status as a recommended practice suggests it is beneficial. Images should be placed near relevant text and on pages that align with their subject matter (Lopezosa et al., 2018). While no official guidelines address the proximity of image-related content, logic suggests considering the paragraphs immediately before and after each image. Another important element, given its direct semantic relationship with the image, is the <figcaption> element. The <figcaption> element within a <figure> grouping associates one or more images with a visible description (image caption) as follows: Figure 2 Example of the use of the <figcaption> tag. Metadata is best to be included in IPTC format, a technology that Google uses to extract certain data from images, as it can improve their visibility in Google Images (Google, 2025b). IPTC metadata is embedded directly into the image, allowing both the image and its metadata to remain intact when used across different pages. Unlike structured data, IPTC metadata only needs to be added once per image, regardless of how many pages it appears on. The properties extracted by Google include Copyright Notice, Creator, Credit Line, Digital Source Type, Licensor URL, and Web Statement of Rights. According to Google (2024c) “high-quality photos appeal to users more than blurry, unclear images”. Also, sharp images are more appealing to users in the result thumbnail and can increase the likelihood of getting traffic from users.It is important to strike a balance between quality and weight in order to meet the CWV metrics discussed above and to satisfy user expectations and user experience. In 2020, Google introduced the max-image-preview robots meta tag, allowing publishers greater control over how images from their sites appear in Google Discover. When the maximage-preview:large meta tag is added to the header of a webpage, it signals that Google may display the site’s images in larger formats. This feature enhances the visual presentation of content on Search surfaces, such as Discover, creating a more engaging user experience. 75 SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations Rubén Alcaraz-Martínez; Mari Vállez; Raquel Escandell-Poveda; Diana Bajaña-Cedeño Digital communication. Trends and good practices SVG images are embedded in an HTML page in two main ways: either as an external file referenced by the <img> element or by directly incorporating the SVG code within the HTML document. The first method is no different from the practice of using bitmap images within the same element and requires the inclusion of alternative text, along with the other elements mentioned above. Embedding SVG code within the page offers several advantages such as the ability to manipulate SVG elements using scripts and to modify their appearance via CSS. Additionally, certain SVG elements and attributes contribute to accessibility, such as <desc>, <title>, and <text>, which provide textual descriptions for visually impaired users. These elements can also be indexed by search engines (Ferraz, 2017). The <desc> and <title> elements are specifically designed to describe images. While they are not visually rendered on the page, the <title> element may appear as a tooltip when hovered over, depending on the browser. The <text> element defines a graphical text component within the SVG. It can include attributes that control visual properties such as text direction, positioning, and fill. Unlike <desc> and <title>, <text> is rendered visibly on the page, and its placement depends both on its position in the code and on the X and Y coordinates defined within the SVG container. Figure 3 SVG vector bar chart example with XML markup, including <title>, <description>, and <text> elements. In the previous example, the aria-labelledby attribute of the Google-recommended (2024c) W3C WAI-ARIA specification, is also included. This allows an element, in this case the visualisation through its <svg> tag, to be related to a textual description. The value of the aria-labelledby attribute refers to the HTML element (in this case, <desc> containing the description). Originally published in 2013 and periodically updated (Google, 2024b), the Search quality rating guidelines serve as the reference document for understanding what Google considers when assessing the quality of web page content, based on the criteria used by its reviewers. These guidelines introduce the concept of EEAT, which is not only linked to factors such as a page’s ability to satisfy search intent or its compatibility with mobile devices but also to various elements associated with the experience, expertise, authority, and trust conveyed by a site through its content. This is particularly relevant for media and sensitive-category websites, including those related to economic or health issues. The direct experience of the author and their familiarity with the subject they write about constitutes the first key factor. Being a prestigious media outlet associated with data or investigative journalism is one way to fulfil this criterion. Expertise refers to the level of specialisation of both the author and the publication in a specific subject, reflecting the accumulated experience that qualifies them as specialists. Authority, in this context, pertains to the reputation of the media outlet (and 76 SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations Rubén Alcaraz-Martínez; Mari Vállez; Raquel Escandell-Poveda; Diana Bajaña-Cedeño Digital communication. Trends and good practices its website) as well as that of the content author. This authority is assessed based on factors such as the search volume associated with branded keywords (related to the media outlet and author), inbound links, and social mentions. Finally, trustworthiness determines the reliability, honesty, and accuracy of the published information. This is evaluated through technical aspects such as the use of HTTPS, the presence of legal information pages about the media outlet, dedicated author pages featuring a biography relevant to their area of expertise, and links to their published content and professional social media profiles. In the context of data journalism, it is also essential to provide information on the origin or source of the data, as well as the methodology employed. Figure 4 El Nacional.cat has a section (/firmas/) dedicated to all its editors. Each article links to its authors, whose profiles include a biography, social media links, content tags, and related authors. Finally, off-page ranking factors are factors that cannot be managed through the site’s content or technical configuration but that influence its authority and ranking. These factors include the number of inbound links (Chotikitpat et al., 2015; Lopezosa et al., 2019), certain authority metrics applicable at the domain or page level (domain authority and page authority) (García Carretero et al., 2016), traffic volume (Krstić, 2019), and others generated in response to user interaction (such as click-through rate, dwell time, bounce rate or pogo sticking) (Maciá Domene, 2019). It is imperative to consider the influence of external links and the diversity of referring domains when evaluating a page’s ranking potential, given their impact on search engine algorithms. These factors serve as indicators of authority and relevance, contributing significantly to visibility in search results. Domain authority, page authority, and estimated monthly traffic have been identified as key benchmarks that help to explain why some pages achieve strong rankings despite not excelling in certain aspects of on-page optimisation. While metrics such 77 SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations Rubén Alcaraz-Martínez; Mari Vállez; Raquel Escandell-Poveda; Diana Bajaña-Cedeño Digital communication. Trends and good practices as domain and page authority are not direct ranking factors used by search engines, they offer valuable insights into a site’s potential for optimal performance. High estimated traffic can be indicative of a page’s capacity to attract users, which may in turn correlate with good rankings. When comparing the positioning of two journalistic resources competing for the same keywords, these metrics can help explain the greater ranking potential of certain content or media over others. 2. Proposal SEOVisFrame is a proposal based on a set of indicators for evaluating narrative data visualisations. These indicators encompass SEO ranking factors as well as other elements that can indirectly enhance the visibility and ranking potential of such articles. They may be understood both as a set of best practices and as criteria for assessing existing content. A total of 34 factors were identified, and are listed in Table 1. Each factor is assessed using a three-point qualitative Likert scale, coded from 0 to 2, to facilitate the calculation of average scores and enable comparisons. Table 1 Indicators for the analysis of SEO optimisation of data visualisations. ID Factor Description Scope of applicability Analytical method / Rating scores 1Keyword in title meta tag The keyword is included in the <title> meta tag. All Manual <title> evaluation. Rating scores 0: The page does not have a title, or the keywords are not included. 1: The title is of partial relevance. 2: The title is fully relevant. 2 Keyword in meta description The keyword is included in the meta description. All Manual meta description evaluation. Rating scores 0: The page does not have a meta description, or the keywords are not included. 1: The meta description is of partial relevance. 2: The meta description is fully relevant. 3 Keyword in URL The keyword is included in the URL. All Manual URL evaluation. Rating scores 0: The keywords are not included in the URL. 1: The URL is of partial relevance. 2: The URL is fully relevant. 4Keyword in <h1> tag The keyword is included in the <h1> tag. All Manual <h1> tag evaluation. Rating scores 0: The page does not have an h1 tag, or the keywords are not included. 1: The <h1> tag is of partial relevance. 2: The <h1> tag is fully relevant. 84 SEOVisFrame: best practices and method for evaluating the SEO optimization of storytelling data visualisations Rubén Alcaraz-Martínez; Mari Vállez; Raquel Escandell-Poveda; Diana Bajaña-Cedeño Digital communication. Trends and good practices Lopezosa, C., Codina, L., & Gonzalo-Penela, C. (2019). SEO off page y construcción de enlaces: estrategias generales y transmisión de autoridad en cibermedios. Profesional de la Información, 28(1). https://doi.org/10.3145/epi.2019.ene.07 Lopezosa, C., Codina, L., Gonzalo Penela, C., & Pont Sorribes, C. (2020). SEO y periodismo: Marco de optimización global como parte del emprendimiento en cibermedios. http://hdl. handle.net/10230/44923 Maciá Domene, F. (2019). SEO avanzado: casi todo lo que sé sobre posicionamiento web. Anaya Multimedia Marcos, M. C., Gavin, F., & Arapakis, I. (2015, September). Effect of snippets on user experience in web search. In Proceedings of the XVI international conference on human computer interaction (pp. 1-8). https://doi.org/10.1145/2829875.2829916 Mueller, J. (2015). English Google webmaster central office-hours hangout. https://youtu. be/i_xnKznRNCc?t=24m5s Peiró, K., & Guallar, J. (2013). Introducció al periodisme de dades. Característiques, panoràmica i exemples, Ítem: Revista de Biblioteconomia i Documentació, 57, 23-37. https://raco.cat/ index.php/Item/article/view/269701 Radcliffe, D., & Lewis, S. C. (2019). The datafication of journalism: strategies for datadriven storytelling and industry-academy collaboration. In The data journalims handbook 2. https://datajournalism.com/read/handbook/two/training-datajournalists/the-datafication-ofjournalism Rodas, C. M., Vidotti, S. A. B. G., & Monteiro, S. D. (2016). Interface de busca do Google e Yahoo: a experiência do usuário sob o olhar do eye tracking. Informação & Sociedade, 26(2). https://doi.org/10.22478/ufpb.1981-0695.2017v12n1.34043 Salem, H., Salloum, H., Orabi, O., Sabbagh, K., & Mazzara, M. (2025). Enhancing News Articles: Automatic SEO Linked Data Injection for Semantic Web Integration. Applied Sciences, 15(3), 1262. https://doi.org/10.3390/app15031262 Seyser, D., & Zeiller, M. (2018). Scrollytelling: An analysis of visual storytelling in online journalism. In 22nd International Conference Information Visualisation (pp. 401-406). https://doi. org/10.1109/iV.2018.00075 Semrush (2024). Google ranking factors. https://go.semrush.com/Ranking-Factors.html Vicente, J.E., Saquete, R., Rodríguez, R., & Maciá Domene, F. (2024). SEO técnico: aprende a optimizar tu web como un profesional. Anaya Multimedia. Ziakis, C., Vlachopoulou, M., Kyrkoudis, T., Karagkiozidou, M. (2019). Important factors for improving Google search rank. Future Internet, 11(32). https://doi.org/10.3390/fi11020032