Shuichiro Ogawa
日本語

Notes · updated 2026-10-07

What Value Does Data Visualization Have in Industry? Market Valuation, Where Value Arises, and What Changes after Generative AI (from Industry Sources)

This note examines the industrial and business value of data visualization (dashboards, business intelligence, news graphics, and the visualization of official statistics), drawing on industry sources: statutory filings and earnings calls of listed companies, records of the UK statistics regulator and health agency, Ja…

Contents (8)
  1. High valuations and screens that go unused
  2. The market values visualization together with the data platform
  3. Value was clearly created where a single message was conveyed to the public
  4. Value is harder to create with unused screens and numbers without context
  5. What is changing after generative AI
  6. Where does the value move?
  7. Where this account does not apply
  8. Footnotes

High valuations and screens that go unused

This note is about industry practice. The sources are statutory filings and earnings calls of listed companies, records of statistics regulators and health agencies, surveys by research firms, and statements by the makers of visualization tools and news graphics; academic research was not collected1. The difference between data visualization and data art is covered in What Data Visualization and Data Art Share, and Where They Differ, and their treatment in education in How Data Visualization and Data Art Have Been Treated in Creativity Education.

The market places a high value on visualization. In August 2019, Salesforce completed its acquisition of Tableau through an exchange offer of 1.103 shares of Salesforce common stock for each share of Tableau (Salesforce 2019). In December of the same year, Alphabet obtained all regulatory clearances necessary to close its $2.4 billion acquisition of Looker, a unified platform for business intelligence, data applications, and embedded analytics, which was to join Google Cloud (Alphabet 2020). In January 2025, Microsoft said that Power BI had more than 30 million monthly active users, up 40% from the previous year (Microsoft 2025).

However, the share of employees in organizations who actively use BI tools averaged 25% (BARC 2022). BARC, which conducted the survey, notes that this share has shown minimal growth over the seven years it has tracked it. The amount paid for visualization and the spread of people who actually use it are not moving in the same direction. Where does its value arise, and where does it not?

The market values visualization together with the data platform

Each of the large acquisitions bundled visualization with a data platform or business applications. Looker joined Google Cloud (Alphabet 2020), and Power BI is deeply integrated with Microsoft’s data analytics platform, Fabric (Microsoft 2025). On its October 2025 earnings call, Microsoft said that Fabric revenue grew 60% and that it had 28,000 paid Fabric customers (Microsoft 2025b). In its 2026 annual report, Salesforce describes Tableau, built on its AI agent platform, as enabling customers to visualize, analyze, and act on data, and to continuously surface trends, predict outcomes, and deliver recommendations (Salesforce 2026). In Japan as well, IDC Japan reported that the analytics and BI software market grew 15.3% year on year in 2024 (IDC Japan 2025).

The figures for stand-alone BI vendors differ. Domo’s revenue was flat, at $308.6 million in the fiscal year ended January 2023, $319.0 million in fiscal 2024, and $317.0 million in fiscal 2025, while its annual recurring revenue net retention rate fell from 106% to 96% to 89% (Domo 2025). Domo attributes the decline in part to macroeconomic conditions and to challenging renewals from customers with COVID-19 use cases. Alteryx, a data preparation and analytics company, signed a merger agreement in December 2023 to be taken private by investment firms (Alteryx 2024). Visualization is valued where it is built into a data platform or a business workflow, and companies that sell visualization alone are struggling. However, the stand-alone examples are only two companies and do not establish an industry-wide trend.

Value was clearly created where a single message was conveyed to the public

The cases in which value showed up in numbers are concentrated in charts aimed at the public. As of March 2021, the UK COVID-19 dashboard drew about 19 million page views a week and recorded 76.5 million hits in 24 hours at the peak of the third lockdown (UK Health Security Agency 2021). The team wrote that the latest data were available to the public at the same time as they were seen in Downing Street. In user surveys, the share who trusted the data rose from 73.4% in November 2020 to 90.3% in March 2021, whereas in the summer of 2020 fewer than half had said they had confidence in the statistics. The UK statistics regulator concluded that dashboards were a hugely popular way of communicating statistics to the public during the pandemic, and recorded that Public Health Scotland’s dashboard received over 120,000 daily views at its peak (Office for Statistics Regulation 2022).

The same happened in journalism and research. John Burn-Murdoch of the Financial Times said that a COVID-19 chart intended as a one-off became something readers wanted updated every night, and that “we accidentally created a product” (Gibson 2020). Burn-Murdoch also said that by showing the differences between measures and methodologies, the paper allowed people to make an informed decision on whether to believe what they were hearing or to interpret the information for themselves. Max Roser of Our World in Data said that monthly users grew from 2 million before COVID-19 to 10 million, 25 times the annual downloads of the World Bank’s World Development Report (400,000) (Wiblin 2021).

The makers place the value of such charts in the author’s ability to state the point. Mike Bostock, the creator of D3, said that if the author of a visualization has nothing interesting to say about it, the graphic probably should not be published (Schwabish 2020). Amanda Cox, then graphics editor at the New York Times, said that most graphics should stand on their own without the accompanying text, and that the words in a graphic should highlight the relevant pattern or an expert’s interpretation rather than merely say “here is some data” (Simply Statistics 2012). The UK government’s analysis function likewise asks analysts to reconsider their choice of chart if they cannot write down its message in a few sentences (Government Analysis Function n.d.).

For decisions inside organizations, public bodies set the same goal. Japan’s Digital Agency states that its dashboard guidebook aims to make data easy to see, establish a correct common understanding among stakeholders, improve the quality of decision-making, and lead to better actions (Digital Agency 2026a). Describing itself as taking the lead in data-based and evidence-based visualization of policy making and its effects, the agency publishes a list of policy dashboards (Digital Agency 2026b). In July 2025, the Cabinet Office and the Digital Agency renewed the portal for visualizing economic and fiscal indicators as the Japan Dashboard (Cabinet Office and Digital Agency 2025). The renewal responded to requests from a survey of 884 prefectural and municipal finance officials, such as the ability to compare data over time and to select multiple indicators. These are records of goals and redesigns, not measurements of how much decision quality improved.

Value is harder to create with unused screens and numbers without context

On dashboards inside organizations, the makers describe problems with how they are used. Barry McCardel, CEO of the analytics notebook company Hex, said that at every previous job, 80 to 90% of the team’s time went into maintaining dashboards while real analysis lived outside them (Berezovsky 2025). Duncan Clark, co-founder of Flourish, said that a dashboard is a form of visual analysis but does not always get used that way (Schwabish 2025). In BARC’s survey, 50% of data and analytics leaders said that BI usage had “increased a lot,” while the share of employees using it remained at 25% (BARC 2022). Those who use BI use it more, but the circle of users has not widened.

Numbers without context also lead to wrong conclusions. Max Roser said that even when the data are not wrong, people without context draw the wrong conclusions from them (Wiblin 2021). The UK statistics regulator cites a case in which the health agency’s statistics on COVID-19 infection rates by vaccination status were used to support anti-vaccine misinformation (Office for Statistics Regulation 2022). In 2025 guidance, the same regulator said that a dashboard can be a useful communication tool, but that using a dashboard as the sole method of publishing official statistics without a bulletin should happen only in exceptional circumstances and requires careful consideration (Office for Statistics Regulation 2025). Domo’s challenging renewals from customers with COVID-19 use cases also suggest that dashboards built for a particular event tend to lose value once the event passes (Domo 2025).

The value of visualization rests on the data beneath it. In BARC’s 2025 survey of 1,795 data professionals, the top priority was data security and privacy and the second was data quality management, while artificial intelligence and machine learning were gaining traction but did not rank at the top (BARC 2024a). No source in which a disinterested third party measured the quality of decisions or the return on investment with a stated method was found.

On the labor side, visualization is priced as part of analytical work. According to the US Bureau of Labor Statistics, there were 275,600 data scientists in 2025, with a median annual wage of $120,230 and projected growth of 35% through 2035, and their duties include using data visualization software to present findings (U.S. Bureau of Labor Statistics 2026a). No source showing, with a stated method, how much of a wage premium visualization skills alone carry could be obtained.

What is changing after generative AI

On the tool side, natural-language queries and notifications have been added. Copilot in Power BI lets business users chat with their data for on-the-fly analysis, but it requires a paid Fabric capacity (F2 or higher) or Power BI Premium (P1 or higher) (Microsoft 2026). Microsoft writes that without preparing data for AI, Copilot can misinterpret the data and return generic or inaccurate results. Tableau Pulse is described as helping users monitor their data in the flow of work, sending metric digests and alerts to tools such as Microsoft Teams (Tableau 2026).

The makers take a cautious view. In 2024, Benn Stancil, co-founder of Mode, said that an analytics tool that merely wraps a language model does not work in an actual business context at all (DataCamp 2024). Stancil also said that the analyst’s job shifts toward choosing the better of the convincing arguments that AI produces, that is, moderating the debate. The CEO of Rill said that AI agents are blind to dashboards but need access to the primitives behind them, such as metrics (Rill 2026). As of December 2025, Flourish had not launched any AI features, having been careful not to do it for the sake of it (Schwabish 2025).

Surveys and public bodies are also still at the stage of expectation. In BARC’s 2024 survey of 238 data leaders and practitioners, more than one third anticipated moderate improvements in their BI and analytics capabilities over the next 12 to 18 months thanks to generative AI (BARC 2024b). In March 2024, the UK statistics regulator wrote that it was not aware of any examples of AI being used to produce official statistical outputs (Office for Statistics Regulation 2024). In the Reuters Institute’s 2025 survey, 7% of respondents used AI chatbots for news each week (Newman et al. 2025). In the labor market, nearly 45% of Indeed’s data and analytics job postings contained AI-related terms (Indeed Hiring Lab 2026). The US Bureau of Labor Statistics projects that automated design tools such as AI will make graphic designers more productive and reduce the need for these workers (U.S. Bureau of Labor Statistics 2026b).

Where does the value move?

The sources above support three hypotheses. None of them has been measured directly; at this stage, several sources point in the same direction.

First, as charts become cheaper to produce, value moves upstream and downstream of the chart. Upstream lies the preparation of data and metrics. Copilot can return inaccurate answers without prepared data (Microsoft 2026), and what agents need are the metrics behind the screen (Rill 2026). Practitioners’ priorities also lie in data quality and governance (BARC 2024a). Downstream lies the editing of messages and context. Bostock’s standard that a graphic whose author cannot state its point should not be published (Schwabish 2020) and Roser’s observation that data without context lead to wrong conclusions (Wiblin 2021) point to work that remains even as the effort of making charts falls. This hypothesis would be wrong if AI-made charts kept changing organizational decisions without data preparation or editing.

Second, dashboards change from screens that people open to answers and notifications that reach them. While the share of employees using BI has not grown beyond 25% (BARC 2022), the tools have added ways for answers to return when people ask questions and for notifications to arrive when something changes (Microsoft 2026; Tableau 2026). This can be read as an attempt to overcome, without making people open a screen, the barrier that opening screens could not overcome. This hypothesis would be wrong if the share of employees using data did not rise even after natural-language queries spread.

Third, for charts aimed at the public, the value of showing sources and methods remains. Trust in the COVID-19 dashboard rose from less than half to over 90% within months (UK Health Security Agency 2021), and Burn-Murdoch said that showing the differences between measures and methods let readers judge for themselves (Gibson 2020). The statistics regulator is cautious about using AI for official statistics (Office for Statistics Regulation 2024) and about publishing through dashboards alone (Office for Statistics Regulation 2025). As charts become possible to make with AI, the ability to show who made them, from what data, and how, likely remains as the value of charts aimed at the public.

Where this account does not apply

No source in which a disinterested third party measured decision quality or return on investment with a stated method was found. The largest figures for value appeared during COVID-19, an exceptional event, and visualization in ordinary times cannot be assumed to create value on the same scale. The share of BI users is a figure from a vendor-sponsored survey of 214 leaders at the end of 2021 (BARC n.d.), and the definition of “user” differs among surveys. The definition of Microsoft’s user count is not disclosed. The struggles of stand-alone BI vendors rest on two companies, Domo and Alteryx, and do not establish an industry-wide trend.

For Japan, records of the goals and redesigns of dashboards exist, but no figures showing their scale of use or effect could be obtained. Sources with stated methods on the wage premium for visualization skills alone and on BI job postings since 2024 could not be reached either.

Unverified items

No unverified statement remains in the main claims. Sources that could not be reached (the originals of Gartner’s analytics platform market size and ABI usage rate, the methodology of Dresner’s survey, Forrester’s study of Tableau’s return on investment, the Harvard Business Review Analytic Services survey, the share of job postings and wage premium for visualization skills, and usage figures for Japanese dashboards) are recorded in the ledger with the routes attempted.

References

Footnotes

  1. The sources are sorted into three tiers: public bodies and statutory filings of listed companies (T1), the factual parts of surveys by research firms (T2), and statements by the makers of tools and news graphics (T3, developer voice). BARC’s survey of BI usage was sponsored by vendors including Tableau. Microsoft’s user figure is a statement by its executives, and its definition is not disclosed. Statements by makers who promote their own products were split into factual parts and the speakers’ own views. The ledger is at source/review/dataviz-business-value/industry.md. ↩

Author: Shuichiro Ogawa (Design Researcher / Consultant) About me →