The Talent data model is the organised foundation behind the Talent dashboards in Humanforce Analytics. This article explains what a data model is and the types of analyses the Talent data model supports. It is written for Analysts, Recruiters and HR practitioners who want to understand what sits behind the numbers, rather than how to build a specific dashboard.
What is a data model?
A data model is a structured way of organising your recruitment data so it can be analysed consistently. It turns raw records from across the recruitment process into a set of related, clearly defined building blocks.
A data model is made up of three parts:
- Events record the things that happen during recruitment, such as an application being created, a candidate moving to a new stage, an offer being made, or a hire being confirmed.
- Descriptive information adds context to those events, such as the job, the candidate, the source of the application, the recruiter, or the date.
- Relationships connect the events and descriptive information together, so you can move between them in your analysis. For example, you can link a hire back to the vacancy it filled and the source the candidate came from.
Because the data is modelled this way, you can combine these building blocks to answer questions, without needing to write code or understand how the underlying systems store the data.
What the Talent data model describes
The Talent data model covers the end-to-end recruitment lifecycle. It brings together the following subject areas:
- Vacancies – the roles you are recruiting for, including job title, category, location, work type, status and the number of positions.
- Applications – each application a candidate makes to a specific vacancy, and the stage it has reached in the recruitment process.
- Candidate profiles – the people in your talent database, including attributes such as location, experience, availability and preferred work type.
- Offers – formal offers extended to candidates, including work type and remuneration details.
- Recruitment activity – the timeline of events across the process, from applications and stage changes through to hires.
- Sources – where applications and candidates came from, such as job boards, your careers website, referrals, agencies or your talent community.
- Talent pools and the talent community – groups of candidates you maintain for current and future hiring.
- Communications and invites – messages sent to candidates, their responses, and invitations to apply for a role or join your talent community.
- People – the recruiters and hiring managers involved in the process.
- Time – a calendar that lets you analyse trends by day, month or year.
A note on terminology: in the Talent data model, an application is counted as one "candidate" record. One person can submit several applications, so each is counted separately.
How the information connects
The subject areas above are linked through clear relationships, which let you move between them in your analysis:
- A vacancy can have many applications, and each application belongs to one vacancy.
- Each application belongs to one candidate profile, and a candidate profile can have many applications.
- Each vacancy has an assigned recruiter and a hiring manager.
- An application can be linked to an offer.
- Recruitment activity links events to the related vacancy, application, candidate, offer and recruiter.
- The calendar connects to activity dates, so any measure can be viewed over time.
What you can measure
The Talent data model includes a wide range of measures. They are grouped here by theme:
- Pipeline volume – applications created, candidates shortlisted, candidates interviewed, candidates hired, and the number currently in each stage of the pipeline.
- Speed – median time to hire, median days from application to hire, median days a vacancy is open, and the time applications spend in each stage.
- Funnel conversion – the number of applications needed at one stage to progress one application to the next, for each step from sourcing through to hire.
- Source and quality of hire – the share of hires from the talent community, talent pools, referrals and internal moves, along with rehires and the proportion of roles filled on time.
- Offers and invites – offers created and accepted, offer acceptance rate, and acceptance rates for job invites and talent community invites.
- Communications – messages sent, received and responded to, response rate, response channel, and average response time.
- Vacancies – vacancies created and opened, positions filled and unfilled, applications per job, and vacancies that are overdue.
- Talent pools – pool sizes, members ready to hire, target sizes and annual hire targets.
- Candidate demographics – measures such as the proportion of female hires and candidate availability.
How you can slice the data
You can group and filter these measures by a broad set of attributes, including:
- Job – title, category, status, work type, location, recruiter and hiring manager.
- Application – status, stage, source, source group, and whether it was successful.
- Candidate – location, years of experience, availability, relationship to your organisation, and preferred work type.
- Offer – status, recruitment method, work type and remuneration range.
- Communications – message template, response channel and whether a response was received.
- Talent pool – pool name, segment, business unit and owner.
- People – recruiter and hiring manager.
- Time – day, month or year.
What types of analyses are supported
By combining the measures and attributes above, you can answer questions such as:
- Sourcing effectiveness – Which sources produce the most applications, and which produce the most hires?
- Time to hire – What is the median time to hire by department, job category, location or recruiter?
- Pipeline health – Where in the funnel do applications progress or drop off, and where are the bottlenecks?
- Vacancy management – How many vacancies are open or overdue, and how many positions remain unfilled?
- Offer performance – What is the offer acceptance rate, and what remuneration ranges are being offered?
- Candidate engagement – What are the message response rates, and how do job and talent community invites perform?
- Talent pools – How large are your pools compared with their targets, and how many hires come from them?
- Diversity and demographics – How are candidates and hires distributed across attributes such as location and experience?
- Quality of hire – What share of hires are internal moves, rehires, or roles filled on time?
Things to keep in mind
A few characteristics of the data model affect how figures should be read:
- An application is counted as one "candidate". One person can have multiple applications, each counted separately.
- Historical coverage has a start date. Recruitment activity is available from a set historical point onward. Activity before that date is not reflected in trend measures.
- A hire is recorded when a candidate is moved to Offer Accepted, Hired or Onboarding, whichever happens first.
- Sources are grouped automatically. Sources that do not match a known group are grouped as Other.
- Remuneration figures exclude outliers. Values outside expected ranges are removed, so remuneration measures focus on plausible amounts.
- Filled-on-time measures need a target. A vacancy must have a target time to hire set for it to be included.
Related articles
- Feature Overview - Humanforce Analytics
- Talent Analytics Dashboard - Talent Community & Pools
- Overview: Permission Groups