What is demand?
Demand is the number of job postings that meet your search criteria. Depending on the tool you’re using, it is represented either as job postings currently online (in Recruit) or as jobs posted during a specified time period (in Plan). To calculate demand, we process, deduplicate, and count job postings from tens of thousands of websites across the internet.
What can I use this data for?
You can use demand data to understand how competitive the market is for a particular type of talent, or to investigate the types of talent that certain competitors are hiring for.
Acquire: Where do you get your demand data?
To calculate demand, we collect job postings from thousands of sources, including job boards, corporate sites, partner feeds, news sites, staffing websites, and applicant tracking systems. Every day, we process an average of 1.3 million job postings in 21 different languages.
Our complete list of sources is proprietary and confidential, but in some cases, we can verify whether we are harvesting from a source that is essential to your organization. To do so, please contact our Support team.
Organize: How do you prepare demand data for analysis?
Job postings are written in any format the employer sees fit. When data is unstructured like this, it must be cleaned and prepared before it can be analyzed. Cleaning and preparation can include translation, stripping punctuation and special characters, and removing extraneous text not related to the job.We rely on natural language processing to identify the existence and prevalence of skills, credentials, education and experience requirements, and other key attributes within the body of a job description.
Analyze: How do you calculate demand?
Demand is simply the sum of all job postings that match your search. Depending on the tool you’re using, demand may be the sum of job postings currently online (in Recruit), or job postings that were posted during a specified time range (in Plan).
Our database of more than three billion unique job postings over the past ten years represents a comprehensive view of the historical and current hiring demand universe, which allows us to identify and address any abnormalities that may appear in the data.
Deliver: How do you represent demand data?
Depending on the context of the data, we represent demand either as a count of job postings or a percentage of job postings that match your search criteria. When displaying trends is useful, we also show demand counts or percentages over different periods of time.
More about demand:
When was your demand data last updated?
How far back does your demand data go?
How do you identify duplicate job postings?
How do you identify and remove inactive job postings?
How do you identify key attributes within job postings?
What countries do you have demand data for?
What location types do you have demand data for?
What search filters impact demand?
What is an occupation? What is the SOC system?
How do you count a single job posting that will be used to hire multiple people?
How do you count multiple job postings in various locations that will be used to hire one person?
When was your demand data last updated?
Because our demand data comes from online job postings, we’re able to provide a near-real time view of the market. However, because it takes time to process job postings and ensure the accuracy of our data, there may be a delay of a few days between when a job appears on one of our sources and when it appears on our platform.
How far back does your demand data go?
Our searchable, user-facing database contains job postings from the past four years. However, we do maintain a larger internal database of job postings from the past 10 years, which allows our data scientists to identify and address any abnormalities that may appear in the data.
Our user-facing database is refreshed on a rolling basis, meaning that the oldest job postings are from the current date minus 48 months.
In Plan, demand is represented as jobs posted during a time period of your choosing. The earliest date you can choose is the current date minus 48 months.
In Recruit, demand is represented as job postings currently online, so the date of the oldest job posting will vary (but won’t be older than 48 months).
How do you identify duplicate job postings?
Sometimes, employers will post a job in multiple places – like their corporate website and a job board. To ensure that our calculations don’t include duplicate postings, we analyze each post’s requisition number, title, location, employer, and date and remove job postings that appear on multiple sources.
How do you identify and remove inactive job postings?
In TalentNeuron Recruit, demand is represented by job postings currently online. As such, it’s important to ensure that our database accurately reflects our sources. Whenever we visit a source, we compare its current job postings with the batch of postings we harvested from our previous visit and remove those that no longer appear on the source.
How do you identify key attributes within job postings?
We use natural language processing to identify the most important attributes of a job, including the occupation, location, employer, desired skills, and the language the job description is written in. Below are a few examples of the steps we take in identifying these attributes:
- Standard Occupational Classification (SOC) mapping: Mapping the role to one of 867 detailed occupations
- Location identification: For example, identifying that the location of a job is Paris, Texas and not Paris, France
- Employer identification: For example, identifying that the employer is Apple, Inc. (and not an apple orchard) or that Oracle is a skill (and not the company)
- Skill extraction: Identifying each unique skill referenced in the description
- Language identification: For example, identifying that the job description is written in Spanish as opposed to Portuguese
What countries do you have demand data for?
We have demand data for all of the countries available in Recruit and Plan:
- Argentina
- Australia
- Austria
- Belgium
- Brazil
- Canada
- China
- Costa Rica
- Czechia
- Denmark
- Finland
- France
- Germany
- Hungary
- India
- Ireland
- Italy
- Japan
- Luxembourg
- Malaysia
- Mexico
- Netherlands
- New Zealand
- Norway
- Philippines
- Poland
- Romania
- Saudi Arabia
- Singapore
- South Africa
- South Korea
- Spain
- Sweden
- Switzerland
- Thailand
- Turkey
- United Arab Emirates
- United Kingdom
- United States
What location types do you have demand data for?
We are able to provide demand data at the country, state, metropolitan area, and county levels.
What search filters impact demand?
We are able to provide demand data for all of the job attributes in our search experience: function, occupation, employer, skills, credentials, title, experience level, education level, employment type, and keywords. In other words, adding any of these filters to your search will change demand.
What is an occupation? What is the SOC system?
The Standard Occupational Classification (SOC) system is a federal statistical standard used by U.S. federal agencies to classify workers into occupational categories for the purpose of collecting, calculating, or disseminating data. All workers are classified into one of 867 detailed occupations based on similar job duties, skills, education, or training. These detailed occupations are grouped to form 459 broad occupations, 98 minor groups, and 23 major groups.
How do you count a single job posting that will be used to hire multiple people?
Typically, there’s no way to know how many positions an employer hopes to fill with a single job posting. As such, each unique posting is counted once in our demand calculations.
How do you count multiple job postings in various locations that will be used to hire one person?
Rather than counting 1 job ad for multiple locations and multiple job openings, TalentNeuron will display and aggregate this with a job posting count = 1
Example: This job ad lists 4 different hiring locations. Where previously we would count these as 4 different job postings, we will now aggregate them as 1 job posting in the platform.
