Methodology

Our market insights data is the anonymous search behaviour data that has been captured from users on the Studyportals websites since 2013. Our dataset is uniquely large: currently, more than 52 million individual user records are added every 12 months, covering more than 220 countries and territories. These visitors generate over 60 million sessions and more than 200 million programme pageviews annually. This large amount of information has been cleaned, validated and organised.

As Studyportals is only operating one English-language platform and no country-specific websites, visitors are well-distributed across the entire globe, providing a reliable representation of an international pool of (potential) students interested in English-taught education.

Our database contains information on programme-level and currently comprises more than 250,000 individual courses, making Studyportals the largest platform for English-taught education on a global scale. According to our estimates of the total market size, this covers more than 90% of all English-taught programmes worldwide that are open to international students. To keep the database up to date, our more than 40 people strong data team is maintaining an elaborate updating schedule, proactively monitoring rankings and gathering information from university websites. It is our target to update each programme at least once a year, with more regular updates of regular changing information, such as tuition fees. To monitor the database quality, a data quality score is calculated taking into account various indicators such as data completeness and update frequency. As of June 2021, the Data Quality Score totalled 97.02 out of 100 possible points.

Students are predominantly searching for programmes based on general disciplines, rather than specific programme names. For this reason, we are maintaining a discipline list of 15 main and 203 sub-disciplines, based on commonly used categorisations as well as past user search behaviour. When a programme is added to our database, it is attached to up to three matching sub-disciplines. This allows for easy segmentation of our data. The visual below shows an overview of the 15 main disciplines in use.

In our analyses, pageviews on programme detail pages are used as an indicator for student interest, as this is the strongest and most granular indicator available.

The visual below shows the typical user journey on our websites. Typically, a user starts a search by selecting one of the more than 200 (sub-) disciplines and possibly further filters such as destination country or tuition fee. In the next step, the user browses through the search results. This is very different from a typical Google search, where users rarely go beyond the first search result. It is quite common that prospective students go through hundreds of results, selecting programmes they are interested in, based on the programme title, tuition fee, institution, location and programme duration. When a user decides to click through to one of these programmes, we can therefore assume that this user has a strong interest in this programme, based on the information provided. Finally, the user can decide to click through to the university website. This option is however only available for promoted programme listings and is therefore unfortunately not suitable for analysis.

While a single pageview has limited meaning, millions of aggregated pageviews linked to programmes, disciplines, country of origin etc. are a powerful indicator to benchmark the level of interest e.g., against the institutional, discipline or country average. Analysing the interest over time can give an indication of future growth prospects and the influence of current events and developments on student recruitment.

Besides its high granularity, another advantage of using this student interest indicator is its timeliness. In contrast to other indicators such as enrolment data, it is available (nearly) in real-time. Furthermore, prospective students typically start researching their programme options 6-24 months before enrolment. The indicator can therefore give insights about interest, only materialising in enrolments 1-2 years in the future.

The correlation with actual enrolment into university programmes has been verified with information coming from EP-NUFFIC, UNESCO, OECD, and HESA datasets. Data from national enrolment has proven erratic and exiguous, making comparisons across different systems extremely challenging. Nevertheless, Studyportals was able to validate most of the assumptions and is regularly performing research to discover further matches and investigate outliers.

Methodology Constraints

Due to its nature, the Studyportals dataset is subject to certain restrictions that should be taken into account in the analysis and decision-making process.

Offline recruitment

The Studyportals dataset is limited to providing insights of students researching their study options online on the Studyportals websites. This may therefore result in an incomplete picture for countries or subjects with a high amount of offline recruitment. However, an increasing number of students are using Studyportals, in addition, to e.g., local recruitment agents.

No reliable data about Chinese students

China is vastly underrepresented in the Studyportals dataset. Due to restricted internet access and a strong prevalence of Chinese language web offers in Chinese search engines, only a very little number of Chinese students are using the Studyportals websites. Furthermore, offline recruitment via agents is still the by far most popular option among Chinese students who are interested in studying abroad.

Studyportals-specific trends

The development of pageviews on the Studyportals websites is not only dependent on the development of the global student demand, but also on other internal and external factors such as website improvements, competition, Google Search algorithm updates etc. For this reason, most insights presented in our analyses are based on relative comparisons, rather than absolute numbers.

Difference between interest and enrolments

While our research suggests that interest can be a good indicator for future enrolments, there are several other factors that can have a strong influence. Difficulties to obtain a visa or appropriate financing can for example cause huge discrepancies between interest and realised enrolments. For a more complete picture, and to identify these bottlenecks, it is therefore strongly recommended to use the information on enrolment numbers in addition to this data set.

For this reason, all data provided should be always treated as indicative and verified by additional research and data sources.