Product

Retention you predict, then measure

Before the offer, Astris HR estimates how likely a candidate is to still be in the role at 30, 90, 180, and 365 days. After the offer, regular check-ins confirm what actually happened — and each result sharpens the next estimate.

astrishr.com/retention
RFSuccess forecast · Warehouse Associate
early estimate
30 days95%
90 days88%
180 days79%
365 days68%
Projected retention curve
Expected tenure ~ 14 months · confidence 72%
synthetic validation, real-cohort benchmark pending

How the estimate works today

Every match already comes with a retention estimate, so recruiters and caseworkers have a clear read from day one. As real hiring outcomes come in, those estimates keep getting more accurate.

An estimate on every match

Astris HR weighs practical factors — experience, language fit, support needs, commute, and schedule — to estimate how likely a candidate is to stay in the role.

It gets smarter with real results

As placements are made and outcomes are tracked, Astris HR learns from what actually happened. The estimates grow more accurate the more the platform is used.

It tells you how sure it is

Every estimate comes with a confidence level based on how much is known about the candidate. When Astris HR has little to go on, it says so plainly instead of showing false precision.

It shows what drives the number

The estimate isn't just a figure — it names the factors that most raised or lowered it. Recruiters and caseworkers can see what would change the outcome.

It improves your matches

Astris HR already uses how likely a candidate is to stay as a tiebreaker when ranking matches. As the estimates improve, the matches improve too.

Every estimate is on the record

Each estimate is saved along with the information it was based on. Questions are settled by looking at that record, not by argument.

The four horizons

One candidate, scored across the whole first year

Astris HR doesn't boil retention down to a single number — it gives an estimate at each checkpoint, so a recruiter can see where the risk actually sits. A candidate who looks strong at 30 days but shaky at 365 reads very differently from one who's borderline the whole way through.

30 days
95%
90 days
88%
180 days
79%
365 days
68%

Illustrative figures from synthetic validation — real-cohort benchmark pending.

Measuring what actually happens

An estimate is only as good as the real outcomes it's checked against. After every placement, Astris HR follows up with the candidate in their own language to find out how things are going.

Follow-up scheduleCheck-ins on day 1, 7, 30, 60, 90, 180, and 365, sent by text message in the candidate's preferred language.
What replies captureCandidates can reply in their own words. Astris HR sorts each reply into still employed, needs support, or no longer employed. Caseworkers can correct it and note why someone left.
Retention milestonesEach placement records whether the candidate reached 30, 90, 180, and 365 days. These real results are what Astris HR learns from.
Reasons for leavingWhen someone leaves, the reason is recorded — such as by choice, transportation, childcare, pay, schedule, fit, or training — and shown per program in analytics.
Closing the loopWhen a candidate stays, it reinforces what led to that match. When someone leaves early, it steers future matches away from what didn't work.

See Astris HR in your workflow

Run it against your candidate base or a sample JD. We'll set up a tenant in under a day and show you the matches your existing tools are missing.

Request a demo