Indeed Data Frequently Asked Questions
How many markets does Indeed cover?
Indeed has websites in over 60 markets and 28 languages. The full list of markets is here: https://www.indeed.com/worldwide. We have economists in Australia, Canada, France, Germany, Ireland, the UK, and the US.
What proprietary data does the Indeed Hiring Lab use for research?
Our analysis focuses on aggregated and anonymized data on job postings shown on Indeed sites around the world, as well as job seeker search behavior on these sites. We also use information from resumes that job seekers have uploaded to Indeed’s resume database to apply for jobs and to be found by employers searching the database for candidates.
It is important to note that we are not restricted to advertisers on Indeed. Instead, Indeed collects job postings anywhere on the internet and de-duplicates them so that when the same job is collected from multiple sources it is shown only once. Indeed is a generalist site in the sense that we focus on providing “all jobs,” not just a niche market.
Where does the job posting data come from?
The job posting data we use for research is from publicly available information shown on the relevant Indeed website. All jobs and websites are subject to review by Indeed’s Aggregation and Search Quality teams before being posted. Companies post jobs directly to Indeed. In addition, Indeed aggregates jobs from global websites, such as employer career sites, and other sources, such as applicant tracking systems, and allows anyone to access these positions from Indeed’s website.
Indeed allows any employer to list a job for free and also provides employers the option of paying for select placements. Employers find the paid option particularly attractive for hard-to-fill positions and jobs that need to be filled quickly. Employers pay Indeed to increase the visibility of a job posting or to reach a broader subset of candidates.
How is the job posting data structured?
Each job is given an ID code. We extract key information from the job posting, such as title, posting date, company, location, etc. that we can then use for research and analysis. Since job postings are collected from many different sources we do not have a standard schema for the data so the information is extracted from the text using state-of-the-art proprietary algorithms.
How do we count job postings?
For the stock of available job postings we often use daily postings visible on Indeed, averaged over a week or a month. For the flow of new postings we count job postings only when they are first visible. We also sometimes focus on unique job postings so we count each visible job only once.
How does Indeed job posting data relate to other measures of vacancies or job openings?
Government measures of job openings are typically based on surveys of a sample of employers. Indeed job postings are a complete count of all postings visible on the relevant Indeed website. The job postings on Indeed do not reflect a precise number of available jobs, as an opening may be listed in more than one place, may remain online for a period of time after being filled, or may not be advertised online at all. Additionally, employers can use a single job posting for multiple job openings. Indeed uses a deduplication process so that when the same job is collected from multiple sources we only show it once which limits double-counting. The deduplication process is focused on job seeker experience rather than on accurate counting of the number of available jobs.
How do we track job seeker activity?
We primarily measure job seekers by their search activity (entering terms in the what and/or where boxes on the relevant Indeed site) and by their clicks on specific job postings. Job seekers can search on Indeed via desktop or mobile device and we typically combine all activity together, but we can also analyze differences between the platforms.
Job searches are associated with the specific terms entered into the search boxes. These queries range from job titles to industries to job attributes (e.g. full time, part time, work from home) etc. – as well as the location of where the job seeker is looking to work.
Clicks are associated with the postings clicked on. From there we can infer information about the types of jobs the job seeker is looking for including job title, location, and company.
How is the job seeker data structured?
We work with anonymized job seeker data organized by browser cookie ID or by account ID. Job seekers do not need to create an account, but many do, and in that case we have information about the job seeker based on the information they provide in their account, including their resume if they upload one. We do not, however, ask any demographic questions including gender or race. Much of our analysis focuses on job seeker location, work history, and education. Job seeker location can be inferred from their IP address or from information provided in the account.
Why do we report normalized numbers?
The Hiring Lab focuses on analysis that applies to the overall labor market, not just the Indeed product. Reporting searches as shares per million and postings relative to a benchmark time period allows us to net out changes in the business focus on broader insights. The specific normalization used depends on the research question as described in the methodology section of the relevant research post.