The Rise of Alternative Metrics
The established global ranking systems — QS World University Rankings, Times Higher Education Rankings, ARWU — were designed in an era when bibliometric data and reputation surveys were the most accessible proxies for institutional quality. The digital age has made previously impossible data streams available: real-time employer hiring data, student satisfaction longitudinal tracking, graduate salary records, alumni career trajectory databases, and patent-to-product commercialisation rates. These richer data sources are increasingly informing alternative metrics that challenge the established methodological consensus.
Altmetrics — measures of research attention across social media, news coverage, policy documents, and public engagement rather than traditional citation counts — are being piloted by organisations like Altmetric.com and Plum Analytics. Research that generates significant public and media engagement, or that informs government policy, may not produce high citation counts in academic journals but creates genuine societal value that traditional metrics miss entirely. Some ranking publishers are experimenting with altmetric components, though none has yet given them significant weight in their primary rankings.
LinkedIn's Talent Insights platform, which tracks the career outcomes of millions of graduates at an institutional level, has been licensed by some ranking bodies for employment outcome calculations. As LinkedIn data becomes more comprehensive and better standardised, it offers the prospect of genuine large-scale graduate outcome tracking — something that would significantly improve the quality of employability rankings compared to current employer survey approaches.
AI and Machine Learning in Rankings
Artificial intelligence and machine learning are beginning to transform how ranking data is collected, processed, and validated. Current applications and near-future directions include:
- Automated data validation: Machine learning models can identify statistical anomalies in university-submitted data — implausibly large year-on-year improvements, inconsistencies between independently reported figures — that might indicate gaming or data errors. This could significantly improve data integrity without requiring expensive manual audits.
- Natural language processing for qualitative data: AI systems can process large volumes of student reviews, alumni testimonials, and employer feedback at a scale impossible for human analysts, potentially enabling rankings that incorporate genuine qualitative experience data alongside quantitative metrics.
- Research topic classification: Advanced NLP enables more precise classification of research papers by topic, method, and societal impact — allowing bibliometric analysis that goes beyond simple citation counts toward understanding the nature and direction of a university's research contributions.
- Predictive ranking models: Some researchers are developing models that predict future research impact from early-stage indicators — identifying rising institutions before their citation impact fully matures in traditional metrics, which could make rankings more forward-looking and less historically determined.
The integration of AI into rankings also raises concerns about transparency and explainability. Rankings are already criticised for opacity; machine learning models that produce rankings without clear causal pathways from inputs to outputs would further undermine public understanding and institutional accountability.
Student Satisfaction and Learning Outcomes
The most significant gap in current global rankings is the near-complete absence of direct measures of student experience and learning outcomes. Growing pressure from governments, student unions, and accreditation bodies is pushing ranking publishers toward incorporating these dimensions.
The UK's Teaching Excellence Framework (TEF), which rates universities on teaching quality using graduate employment outcomes, student continuation rates, and student satisfaction data from the National Student Survey, represents one model for how teaching quality might be systematically measured and incorporated into comparisons. TEF data is now integrated into some UK-specific guides, and THE has incorporated some TEF-derived indicators into its UK-specific rankings.
New Zealand's student satisfaction data from the Graduate Outcomes survey and Australia's QILT (Quality Indicators for Learning and Teaching) similarly provide national-level student experience data that could inform comparative ranking. The challenge is international standardisation: different countries collect student satisfaction data differently, with different questionnaire designs, sampling methods, and response rates, making direct comparison across countries technically demanding.
Quality Assurance agencies — which assess whether universities meet minimum standards for educational provision, governance, and student support — are another potential data source for ranking systems seeking to incorporate academic standards information. In several countries, quality assurance review outcomes are publicly available and could be systematically integrated into ranking methodologies.
Employer-Driven Rankings
As employers increasingly bypass traditional degree credentials in favour of skills-based hiring, there is growing interest in employer-compiled assessments of which universities produce work-ready graduates in specific technical domains. Tech companies, financial institutions, and consulting firms conduct detailed analyses of which universities their best-performing hires attended — analyses that often diverge significantly from public rankings.
Platforms like LinkedIn, Glassdoor, and specialised hiring analytics tools are beginning to publish or license proprietary graduate outcome data that effectively constitute implicit university rankings by field and employer type. A computer science employer in San Francisco may find that University of Waterloo graduates consistently outperform graduates from nominally higher-ranked institutions — information that, if publicly available at scale, could fundamentally reshape how universities in technology fields are evaluated.
Some employers — particularly in technology and consulting — have moved away from university filtering in hiring decisions altogether, relying instead on technical assessment platforms like HackerRank, Codility, and bespoke case studies. If this trend accelerates, the employer reputation components of university rankings may become less relevant as employers develop their own direct assessment pipelines.
Open Data and Transparency
A growing movement in academic publishing and research infrastructure advocates for open data — the principle that research outputs, institutional data, and scientific findings should be freely accessible rather than behind commercial paywalls. The implications for university rankings are significant:
- As more research is published open-access (required by major funders including the Gates Foundation, Wellcome Trust, and EU Horizon), the accuracy and completeness of bibliometric databases should improve — reducing current biases toward commercially indexed journals.
- Universities are under increasing pressure from governments and accreditation bodies to publish standardised institutional data — graduate employment rates, completion rates, research expenditure, staff demographics — that ranking publishers could incorporate as verified third-party data rather than relying on university self-reporting.
- Open research data requirements, which mandate that the underlying data supporting published research be publicly accessible, could enable more sophisticated citation and impact analysis than is currently possible with metadata-only bibliometric systems.
The European Universities initiative and various national open data mandates are accelerating the availability of standardised institutional data that could make rankings both more accurate and more resistant to gaming. However, the commercial interests of ranking publishers — who derive revenue from subscriptions, consulting services, and data licensing — are not always aligned with maximum methodological transparency.
Regional and Mission-Based Rankings
A significant countertrend to the globalisation of university rankings is the growing interest in mission-aligned assessment — evaluating universities against the specific goals they set for themselves rather than against a universal standard that may be irrelevant to their context.
Community colleges, teaching universities, applied universities, and institutions with social inclusion mandates all serve important societal functions that global research rankings entirely fail to capture. Evaluation frameworks like U-Multirank (European Commission), the Carnegie Classification system (US), and THE Impact Rankings (UN SDGs) represent attempts to assess universities against multiple dimensions simultaneously, allowing different institutional types to be recognised for what they actually do well.
Quality Assurance frameworks in many countries already take a mission-based approach — assessing whether universities are achieving the goals they have set for themselves rather than against a single universal standard. Incorporating quality assurance outcomes into ranking methodologies would provide a way to recognise the diversity of institutional missions while maintaining some basis for comparison.
Will Rankings Become Obsolete?
Periodic predictions that university rankings will become obsolete have so far consistently proven wrong. Each time a major criticism gains traction — gaming, bias, lack of teaching measures — ranking publishers have responded with methodological refinements that maintain their relevance and commercial viability. The incentives for rankings to exist are strong: students need decision support when navigating hundreds of institutions across dozens of countries; governments use rankings to monitor return on higher education investment; universities use them in marketing; and employers use them as screening proxies.
What seems more likely than obsolescence is fragmentation: a single global ranking becoming less dominant as users turn to a portfolio of specialist tools — subject rankings, employability databases, teaching quality frameworks, national QA assessments, personalised comparison platforms — each better calibrated to specific decision needs than any single composite number.
Technology is enabling more personalised comparison tools that could eventually replace or significantly supplement traditional rankings for student decision-making. Platforms that allow students to define their own criteria, weight them according to personal priorities, and receive customised institutional recommendations based on verified data represent the logical evolution of what rankings attempt to do — without forcing all students to accept a single, publisher-defined vision of what "best university" means.
For the foreseeable future, understanding how QS World University Rankings, Times Higher Education Rankings, and other established systems work will remain a valuable skill for students navigating higher education decisions. But the most sophisticated students will use these rankings as one input among many — combining them with subject rankings, employer reputation data, student satisfaction sources, cost information, and personal values alignment — rather than treating any single ranking position as a verdict on institutional quality.