Keep good candidates warm until the right role opens

An email cadence is short, urgent and aimed at getting a reply about a specific role. A nurture sequence is long, slow and aimed at not being forgotten. The two need different mechanics, which is why they will be built as separate objects rather than as one sequence tool with a longer delay setting.

A nurture sequence will run in weeks and months rather than days. It will not escalate. It will not chase a non-reply, because a non-reply to a nurture touch is a normal outcome and not a signal to try harder. It will pause automatically the moment a candidate enters a live process, and resume only when that process ends without a placement, so nobody receives a general market update on the morning of their second interview.

The reason nurture goes wrong is almost never the individual message. It is that three recruiters on three desks each thought their touch was the only one, and a candidate received four messages from the same agency in a fortnight. From the outside that reads as one organisation that does not know what it is doing.

Frequency caps will be enforced per candidate at the tenant level, not per sequence and not per recruiter. A cap will count every automated touch across email, WhatsApp and any future channel, with a separate and higher allowance for a recruiter reaching out by hand about a specific live role. When a nurture step would breach the cap, it will be skipped rather than queued, because a delayed pleasantry has no value and a stacked queue is how the breach happens later anyway.

A nurture sequence can run for a year. Consent checked once at enrolment is a check that was accurate a long time ago. Every step will re-read the candidate's consent basis and per-channel suppression state before it sends, and a candidate who opts out at month seven will not receive month eight.

Retention interacts with this directly. A record that has passed its retention window is flagged for review or erasure, and a flagged record will drop out of nurture rather than continuing to be messaged while the business decides whether it should still be held. The consent basis, its capture date and channel already live on the candidate record; nurture will read them rather than keeping a second list of its own.

This is not a new engine. The WhatsApp sequence runner is already in production, executing timed multi-step flows against candidate records with window awareness and suppression checks. The Wave 1 work extends that runner to email steps and recruiter task steps, and adds the pool-level enrolment and cap logic that nurture needs.

Building on something that already runs is the reason we are willing to put a date on it. It also sets the shape of the first release: nurture will support email, WhatsApp where WhatsApp is available in your market, and tasks. SMS and LinkedIn steps are not part of it, for the reasons set out on the multi-channel page.

Pools will be defined the way a desk already thinks: silver-medallists from a named client, contractors whose assignment ends in the next quarter, a skill pool for a role you fill repeatedly, or everyone who asked to be kept in mind and said so on the record. Membership will follow a saved search, so a pool stays current instead of being a snapshot somebody built in March.

Reporting will answer the only question that matters about nurture, which is whether it produced placements. Each placement already carries source attribution; a candidate who came back through a nurture pool will be attributed to that pool, so the programme is judged on hires rather than on open rates. Where the numbers are thin, the reporting will say so rather than presenting a confident figure computed from a handful of events.

There will be no lead scoring on candidates. Scoring people on how engaged they seem with your emails is a marketing idea that transfers badly to hiring, and it invites decisions being made about a person for reasons nobody can defend afterwards.

There will also be no content library or newsletter builder. If your nurture programme depends on producing a monthly publication, the product is not what is stopping you and adding a template gallery would not change that. What we will build is the pacing, the caps, the consent checks and the attribution, because those are the parts that agencies consistently get wrong by hand.