A closer look at Creatives

Using job adverts to identify the skill needs of creative talent

Motivation

Skills development is a key focus of the UK Government The Government recently chose ‘developing skills’ to be one of ten key pillars in its Industrial Strategy. This was recognition that skills development is as important to growth as infrastructure, investment and trade.

And the Government has invited the creative industries to reach a sector deal As part of the Industrial Strategy, the creative industries are one of five sectors that have been invited to reach a sector deal with the Government. This deal will require the creative industries to show how they plan to harness each of the ten pillars to boost productivity and enhance competitiveness. In singling out the creative industries, the Government clearly recognises their importance to the UK economy. Between 2011 and 2015, employment in the creative industries grew by 19.5%, compared to growth of just 6.3% in the wider UK economy.

But the creative industries face an information gap on skills The plan put forward to the Government by the creative industries will need to show how the sector proposes to ‘boost skills’ and create high value, high productivity jobs. However, the sector faces a challenge: there is little granular evidence on the skills required by creative talent. Moreover, the creative industries employ individuals in an enormous range of creative occupations, from software developers to museum curators.

And Brexit only heightens the need to better understand skill requirements Understanding creative talent has become even more pressing in light of the UK’s decision to leave the European Union. Many creative sub-sectors rely on European migrants to fill skill shortages. In 2015, approximately 5.5% of workers in IT, software and computer services were from EU nations and 8.4% from outside of Europe. For Publishing, 9.5% of the workforce was from Europe, while 4% was made up of non-European talent. If access to non-UK talent is restricted, this may exacerbate skill shortages in these sub-sectors. In other creative sub-sectors, there are particular occupations that are reliant on migrant talent in a way that is hidden in industry statistics. The prospect of skill shortages only heightens the need for a granular picture of the skills needed by creative workers. This information will help to ensure that the UK talent pipeline can meet the needs of employers.

Data

Using online job advertisements to learn about skill needs Online job advertisements are a valuable source of information for learning about the skills required in different jobs. This research uses job adverts that were collected and processed by Burning Glass Ltd. The dataset contains millions of adverts for UK positions that were placed online between 2012 and early 2016 (inclusive). Duplicate adverts are deleted and the skills and software programs in each advert are harvested. There are thousands of unique terms, and the skills include types of knowledge, abilities, behaviours and work activities. Burning Glass places each advert into an occupation group, based on the title of the job advert. The taxonomy of occupation groups used in this analysis is the Standard Occupational Classification system (SOC codes at 4-digit level). For this research, basic skills have been excluded, as these skills are less precisely defined. For example, two employers asking for ‘communication skills’ may have very different expectations of what that means.

Method

Collecting skills in creative occupations This research aims to identify the skills used by workers in creative occupations. The focus is on workers in creative occupations, as opposed to workers in creative industries. The latter includes non-creative jobs (e.g. accountants working for museums). The former includes creatives who work in non-creative industries (e.g. graphic designers working for accountancy firms).

What is a ‘creative occupation’? The Department for Culture, Media and Sport (DCMS) defines ‘creative occupations’ as a set of 30 occupations from the Standard Occupational Classification system (SOC codes at 4-digit level). The occupations were chosen by DCMS informed by criteria developed by colleagues at Nesta and the occupations range from software developers to fashion designers, and from museum curators to journalists.

Collecting adverts for creative occupations The initial dataset contained all UK job adverts from 2012 to early 2016 that had been placed by Burning Glass into one of the 30 creative occupations. To limit the risk of jobs being incorrectly placed into the wrong occupations, a list of trusted job titles was assembled for each occupation, based on their frequency and manual checking. Only adverts with these job titles were selected for analysis. From this dataset a large, random sample of adverts was drawn for each occupation. The size of each sample was proportionate to employment in that occupation, where the latter was taken from DCMS’s creative economy employment statistics.

Clustering skills Thousands of skills were extracted from the adverts. These skills were grouped into clusters using a community detection algorithm. It works by grouping skills which tend to appear in the same adverts. The algorithm found five large skill clusters. Each cluster was given a name, based on the most frequently occurring skills that it contained. The naming process requires an element of judgement and invariably within each cluster there are examples of skills that are not entirely consistent with the title of the cluster.

The five key skill clusters are Support skills, Creating & design skills, Tech skills, Marketing skills and Teaching skills The table below shows the 30 skills in each cluster that appear most frequently in adverts for creative occupations.

Strengths & limitations

Job adverts provide skills information directly from employers Online job adverts can provide a highly granular picture of skill demands for individual occupations. This same picture cannot be painted using survey data, owing to small sample sizes. Moreover, job adverts allow employers to describe skill needs in their own terms - a bottom-up approach. In contrast most surveys take a top-down approach, by asking employers to select skills from a short and pre-determined list of terms.

But data from job adverts has definite drawbacks This data source is not a silver bullet, and there are a number of limitations to bear in mind. As a result, the findings should be seen as indicative and any percentages reported (for skills and skill-combinations) should be treated only as approximations.

• Not all jobs are advertised online Some occupation are less likely than others to advertise online. And several craft-related occupations are particularly affected by this issue. This is likely to be due to high levels of self-employment and the close nature of the community where recruitment is often conducted through social or professional networks. It was not possible to gather a sufficient number of job adverts to form an accurate representation of the skills required in Craft occupations. Accordingly, these occupations were excluded from the analysis.


• Adverts are not exhaustive The space in an online job advert is limited. As a consequence, an employer may not mention every skill that is required. This is particularly true for basic skills that employers may in any case assume all applicants will possess. Conversely, adverts may also include skills that are not essential, but would be desirable.


• Errors in collecting and processing adverts There are several ways in which errors can creep into the processed dataset which can cause the wrong skills to become associated with an occupation. For example, it can be difficult to distinguish where one advert ends and another starts, causing the adverts to become merged. It may also be difficult to distinguish skills mentioned in the advert from other words that appear on the webpage from which it was scraped. And finally, the English language itself can induce errors because one word can carry multiple meanings. For example, ‘Chef’ may refer to cooking but it may also refer to a programming language called Chef. A number of skills were necessarily excluded from the analysis for these reasons.

Developing solutions to these limitations is underway as part of Nesta’s research in the Economic Statistics Centre of Excellence (ESCoE).

Results

The DCMS places the 30 creative occupations into nine creative groups (see page 25) - each made up of between 2 and 5 SOC codes. The data visualisation below shows how each creative group relies on the five skill clusters, as well as providing information on employment, growth and salaries.

The creative groups vary enormously in size and growth In each panel the white bars on the left-hand side show the level of employment for the creative group in 2015, while the white dots on the right-hand side show the growth in employment between 2011 and 2015, as reported by DCMS. IT, software and computer services are the single largest creative group, accounting for approximately 28% of employment in creative occupations. They have also experienced the fastest growth between 2011 and 2015, expanding by 31%. The second largest creative group is Advertising & marketing and along with IT they account for over 50% of employment in creative occupations. The smallest creative groups are also broadly those which are growing most slowly. Employment levels in Publishing, Crafts and Museums, galleries & libraries all shrunk between 2011 and 2015.

There are large salary gaps between creative groups The white histograms at the bottom of each panel show the distribution of annual salaries collected from job adverts for a creative group. The black histogram in each panel shows the salary distribution for all creative groups combined. There is a gap of £16k between the median annual salaries of the highest and lowest paid creative groups. There are also differences in the distribution of salaries between groups. For example, the salaries offered in Film, TV, video, radio & photography are more skewed towards higher salaries than those for Museums, galleries & libraries which are more tightly bunched around lower salaries. The median salary for all creative occupations is approximately £33k per year. However one must be careful not to compare these median salary figures to official labour market statistics. The former are not weighted to account for variation in the tendency to advertise online, and adverts for better paid roles are less likely to contain a salary range.

Two clusters contain skills that are predominantly creative As mentioned above, the clustering algorithm finds that the skills used by creatives can be grouped into five main clusters: Support skills, Creating & design skills, Tech skills, Marketing skills and Teaching skills. Marketing skills and Creating & design skills could both be described as ‘creative skill clusters’. The two clusters include many skills that also appear in the names of the creative groups. For example, the Marketing cluster includes several skills that mention marketing which is itself a creative group (‘Advertising & marketing’). And the Creating & design cluster includes skills such as music, video production and design which all feature in the names of other creative groups. These two clusters arguably contain many of the defining skills required for creative jobs. These types of skills are fundamental when laying out the talent pipelines in a creative industries strategy.

But it’s not just about designing & marketing; complementary skills are key The three other clusters are Tech skills, Teaching skills and Support skills. The skills in these clusters may not spring to mind when one considers the skill needs of creative workers. However, the skills in each cluster are essential to the creative process:

1. The Tech skills cluster includes IT-related systems, programs, languages and libraries, as well as skills related to their use, such as technical support. These skills facilitate the smooth running of businesses, but also provide a platform for expressing creativity - be that a word processing program, a website or app.


2. The Teaching skills cluster includes a number of skills related to teaching and information systems. These skills are essential for sharing creative skills and capturing creative outputs. As a side note, this cluster includes a number of skills that look out of place, such as Biology and Teaching PE. These skills crop up because adverts for roles such as music teachers may mention other subjects, and also because a number of technical writing roles (for science) require specialist subject knowledge.


3. The Support skills cluster includes a number of skills related to management, selling and customer service. These skills are essential to ensuring that the creative process runs smoothly and remains viable.

It is important to remember that creatives need these complementary skills when designing policies to support creative workers. Whilst skills from the two creative clusters are clearly fundamental, supplementing these with skills from the three complementary skill clusters are crucial for the successful realisation of the creative process.

Digital tech skills are crucial to creative jobs The Tech skills cluster, and its presence in every creative group, show the inseparability of digital tech from creative skills. Moreover, several software programmes fall directly within the Creating & design cluster, including Adobe products, SketchUp and 3D Modelling/Design. Lastly, there are a number of basic software programs in the Support skills cluster, such as Microsoft products. All these results demonstrate how digital technology acts as a medium for making, facilitating and sharing creatives’ work.

There is no ‘one creative type’ Creative workers are often referred to as a collective group. But despite sharing a number of characteristics, the results above act as a reminder of the huge variety in creative types. The bar charts in each panel show that the creative groups draw on the five skill clusters in very different ways. For example, Advertising & marketing draws heavily on the Support and Marketing clusters, while Product, graphic and fashion design relies more heavily on skills from the Creating & design cluster. The results therefore show how important it is not to take a ‘cookie cutter approach’ to creative education. Such an approach would ignore the diversity of skills and skill-combinations, and potentially endanger skills provision for more niche creative roles.

There is some evidence of specialism Within each bar chart, the darker section shows the percentage of adverts that only require skills from that particular cluster. For the IT, software and services group around 30% of adverts only require skills from the Tech cluster, suggesting the demand for specialists. Similarly, around 40% of adverts for Architects only require skills from the Support cluster, as many roles rely heavily on project management skills and computer aided design which both fall into this cluster. Around 30% of adverts for the Music, performing & visual arts group only ask for skills in the Creating & design skills cluster. In the same group, around 15% of adverts only ask for Teaching skills. These two results suggest there are quite distinct occupations within this one creative group (such as composers and music teachers), who share similar mediums to express their creativity but have very different skill requirements.

But there is much more evidence for skill mixing Across the creative groups, most job adverts ask for skills from multiple clusters. In each panel, the half circles show the most common skill combinations. They show that skills from the Support and Marketing clusters are often required in combination with another cluster. For example, in Advertising & marketing, around 60% of adverts require both Support and Marketing skills. Jobs in Product, graphic & fashion design commonly mix Creating & design skills with other clusters. Meanwhile, jobs in the Museums, galleries & libraries group exhibit the broadest range of skill mixing. The evidence for skill mixing reinforces the case for developing individuals with a mixture of creative and complementary skills.

Are there hidden creative occupations?

The DCMS’s set of ‘creative occupations’ was never intended to be exhaustive and choices were driven in part by the Department’s mandate. Moreover, the identification of creative occupations is necessarily subjective. A natural question therefore is: are there other ‘creative occupations’ or ‘near creative occupations’ that perhaps risk being overlooked? This is an important question as gaining the label of a ‘creative occupation’ can alert policymakers to the economic significance of the occupation (see Creativity vs. Robots: the creative economy and the future of employment). Two different approaches are taken to answering this question.

---

Analysis and data visualisation by Dr. Cath Sleeman, quantitative research fellow at Nesta. Co-written with Dr. George Windsor, senior policy researcher at Nesta.