Labour market data is a powerful asset for any organisation. Whether you use it to plan your hiring strategy, reskill your workforce, benchmark your organisation against sector averages or, in the case of educational institutions, ensure the quality of your courses - its importance is indisputable.
But to gather relevant insights and make the right decisions, your data needs to be representative. In other words, it needs to give you an accurate and reliable picture of your labour market.
Lightcast collects job postings from 51,000 sources worldwide. Our software extracts key information from these postings, such as job title, employer, industry and advertised salary. All duplicate job postings are removed to provide a more accurate picture of the labour market, and this data is then analysed and incorporated into our tools.
The Lightcast team put together an analysis of the representativeness of our data in the APAC region, particularly focusing on New Zealand, Australia, and Singapore. For this analysis, we benchmarked our data against a variety of official national sources, which we explain in detail in each section. This blog will take you through our key findings across the three countries studied.
The sources Lightcast studied
We studied three labour market data sources in New Zealand: Jobs Online, maintained by the Ministry of Business, Innovation & Employment (MBIE), the Quarterly Employment Survey (QES), and the Household Labour Force Survey (HLFS).
For Australia, we analysed the Job Vacancy Survey (JVS)1 conducted by the Australian Bureau of Statistics (ABS) and the Internet Vacancy Index (IVI)2 conducted by the National Skills Commission (NSC). JVS publishes job postings data quarterly and only accounts for jobs that are available for immediate filling on the third Friday of the middle month of the quarter. To provide a fairer comparison, we analysed job postings only posted during a mid-quarter month (February, May, August and November).
The occupation classifications for both New Zealand and Australia are based on the ANZSCO Major Group classification standard, and the standard industry classification used in New Zealander and Australian government data is the Australian and New Zealand Standard Industrial Classification (ANZSIC).
Lastly, for Singapore, we analysed the quarterly Labour Market Survey (LMS) conducted by the Ministry of Manpower (MoM). Plus, the Singapore Standard Occupational Classification (SSOC) system is used to classify occupations, and the Singapore Standard Industrial Classification (SSIC) is used to define its industries. Because the government sources studied name and group industries differently, we combined our data in a similar manner for a more accurate benchmark.
Job postings - time series analysis
Lightcast data and the two Australian data series show a similar picture of the labour market trends since 2013. Unsurprisingly, there was a drop in job postings at the onset of Covid-19, followed by a steady rise since then.
We analysed both Lightcast job postings and Jobs Online data over a period of six years. Both Lightcast and New Zealand government sources detected the same drop at covid-19 peak time (early 2020), followed by a steady growth since then.
Lightcast and LMS data have had a correlation of 0.89, so both sources spotted similar job postings trends across time. The decline in the number of job postings at Covid-19 peak times was far less noticeable in Singapore than in New Zealand and Australia. Yet, we can see a significant increase in job postings since 2020, and Lightcast’s job postings index shows a much higher growth than the LMS index.
Various industries are significantly over-represented by Lightcast data, such as Healthcare and Social Assistance, Education and Training, Mining and Public Administration and Safety.
A few industries, however, are underrepresented by Lightcast data. This includes, for instance, Accommodation and Food Services, Construction and Manufacturing. Such industries would typically advertise for their open roles through offline job postings and word of mouth, or have perpetually open positions.
Based on the graphic below, we can see that Lightcast data has a particularly high representativeness in sectors like Healthcare and Social Assistance, or Public Administration and Safety.
However, for industries such as Construction and Manufacturing, Lightcast data isn’t as representative. This is mostly because word of mouth or offline job postings are generally more popular when recruiting for roles in these sectors, whereas in sectors like Healthcare and Social Assistance, online job postings are prevalent.
Lightcast data is over-representative in various industries, particularly Financial and Insurance Services, Professional Services and Accommodation and Food Services. Some of the industries where Lightcast data may not be as representative as MoM vacancy data source include those where recruitment is primarily conducted offline, such as Construction and Transportation and Storage.
The correlation between the Lightcast and LFS occupation distributions is 0.97, while the correlation between the Lightcast and IVI occupation distributions is 0.96. We can see an over-representation of Lightcast data in the “Professionals” occupation relative to LFS data, as well as a slight under-representation in a few categories such as Labourers.
The Professionals occupation category is high-skilled and tends to be posted online at higher rates than occupations based on manual work, such as Labourers. These tend to be recruited through offline channels.
Lightcast data is showing an over-representation for most occupations, including Clerical and Administrative Workers, Labourers, Professionals and Sales Workers. Such occupations tend to be highly advertised online, which enables us to gather more data. However, we see a slight under-representation in some categories, such as Managers, which are often promoted within and therefore do not have a 'public' job posting to collect.
The correlation between Lightcast data and MoM occupation distributions is 0.81, so they show very similar results across most of the occupations studied. Professionals is the occupation most over-represented by Lightcast, as recruitment tends to be online. On the other hand, Cleaners, Labourers and Related Workers are under-represented by Lightcast relative to MoM data, as recruitment for these occupations tends to be offline.
Drilling down into Australian regions, we compared our regional job postings data with the regional vacancy data from JVS and IVI. There is less than a 5% difference between Lightcast data and JVS and IVI sources in each Australian region we studied.
For most of the regions we recorded, our data is closely correlated with the HLFS data. Auckland and Wellington are the only two regions where Lightcast data shows a smaller proportion of job postings overall than HLFS data.
Getting representative labour market data
From this analysis, we can conclude that Lightcast data offers a strong representation of the labour market in New Zealand, Australia, and Singapore. Lightcast is actively benchmarking our data against a variety of third-party and official governmental sources to ensure that our data representativeness provides an accurate picture of the labour market.
Reach out to our team if you have any questions or would like to discuss your labour market challenges and how Lightcast can help.