Refolk
ReferenceRecruiting and sourcing

The Move-Readiness Cue Reference: Message This Week or Hold

Look up any public move-readiness cue and state what it proves, how it lies, how many days it stays live, whether you may use it, and what it must stack with.

22 min readLast reviewed November 5, 2026Read as Markdown

Key takeaways

  • In Refolk's index, only 129 of 353,403 US Software Engineer profiles put "open to work" in the headline, about 1 in 2,740, so open-web sourcing misses most open-to-work flags.
  • LinkedIn's figure that framed profiles are 40% more likely to receive recruiter InMails counts messages received, not replies, so the flag's reply lift has never been published.
  • After a voluntary exit on a team, coworkers are 9.1% more likely to leave, and further quits are most likely between Day 45 and Day 135, peaking around Day 70 (Visier).
  • Team size changes how much a peer departure matters by more than 3x: 25.1% added quit probability on a team of two against 7 to 9% on large teams.
  • Under ICO guidance, inferring health, union membership or similar traits counts as special category processing even when the guess is wrong, so screen cue types before you score, not after.
  • No published source sets numeric stacking thresholds or decay curves for most cues, so the only way to calibrate cue weights is to log replies against the cue that triggered each message.

This reference is for ordering a slate of passive candidates you have already qualified: who gets a message this week, who waits, and how long they wait. It is written for in-house recruiters, sourcers, talent leaders and founders who do their own hiring. For each public move-readiness cue it gives what the cue proves, how it misleads, how many days it stays live, whether you may lawfully use it, and how much corroboration it needs before it earns a place in this week's batch.

What is a move-readiness cue?

A move-readiness cue is a dated, public event that suggests a person is open to leaving now, separate from whether they fit the role. Fit decides who is on the slate. Readiness decides only when you message them.

Most pages ranking for "signs a passive candidate is open to a new job" list five to seven signs with no measure of strength, or sell a black-box "likely to move" score. Neither helps you order a slate on a Monday morning. What does help is a short set of defined terms, used the same way every week.

  • Cue: one observable event tied to a person, such as a manager leaving their team or a headline change. A cue without a date is not a cue yet.
  • Event date: when the underlying event happened, not when you noticed it. All windows count from this date.
  • Window: the span of days during which a cue is believed to predict a move. Only a few windows have been measured.
  • Live or stale: a cue inside its window is live; a cue past it is stale and is dropped from scoring.
  • Weight: the points a live cue contributes to a person's score. The weights in this guide are my starting values, not published figures.
  • Stacking: adding up live cue weights inside one window to decide whether a person crosses the outreach threshold.
  • Lawful-use flag: whether a cue may be used at all, given that some cues reveal or infer protected characteristics.
A readiness cue decides when you call. It never decides whether the person belonged on the slate.

The cue reference: what each cue proves and how it lies

Each entry below gives one cue's evidence, its false-positive pattern, its shelf life in days where one is published, its lawful-use status, and its starting weight. Self-declared cues carry weight 3 and can stand alone; strong inferred cues carry 2 and need one corroborator; weak cues carry 1 and need two.

Where a shelf life says "not published", that is literal: no public source measures how fast that cue decays. Recheck it at every weekly batch and retire it when the event it rests on is superseded.

Open-to-work flag

  • What it is: LinkedIn's self-declared status. The "All LinkedIn members" mode adds a public frame and is visible to everyone, including colleagues at the current employer. The "Recruiters only" mode shows status only to LinkedIn Recruiter users and adds no frame.
  • What it proves: the person told the platform they are open. The public frame also proves they accept that their employer may see it.
  • How it lies: a laid-off or freelance person left it switched on. The published lift (framed profiles are 40% more likely to receive recruiter InMails, and 20% more likely to hear from their community) counts messages received, not replies.
  • Shelf life: not documented. Check when the flag was last updated and whether a current role is shown.
  • Lawful use: yes. It is a self-declaration.
  • Weight: 3 with a visible current role; 1 without one, until you confirm the flag is current.

Profile rewrite

  • What it is: a changed headline, about section, title or skills list on the candidate's own profile.
  • What it proves: effort spent on public self-presentation. In Felps et al. (2009), revising résumés was part of the job-search behaviour that carried turnover through teams, and patent US8352406B2 counts title or profile updates among its signals.
  • How it lies: it follows an internal promotion. If the title changed and the employer stayed the same, assume promotion until shown otherwise.
  • Shelf life: not published.
  • Lawful use: yes.
  • Weight: 1. A teammate's rewrite is also a weight-1 cue for the candidate, because peers' visible job searching is the channel contagion travels through.

Manager exit

  • What it is: the candidate's direct manager or unit head has left the company.
  • What it proves: the candidate's immediate situation has become uncertain. Pieper et al. (2023) studied 239 general manager departures at a large US retailer and found a potential delay before the disruption phase.
  • How it lies: someone was promoted internally into the role. Internal promotions may reduce the group turnover that follows a manager's departure.
  • Shelf life: opens after a delay; hold until Day 45. The close of the window is not published.
  • Lawful use: yes.
  • Weight: 0 before Day 45 or if the successor was promoted from inside the team; 1 after Day 45 if the seat is vacant or filled from outside.

Peer departures

  • What it is: one or more people on the candidate's team have left.
  • What it proves: measured contagion. Visier found employees are 9.1% more likely to leave after a voluntary exit on their team and 7.7% more likely after an involuntary one. The sample was 86 organizations with more than 1,000 employees, covering 17 million or more anonymized records from February 2019 to June 2022.
  • How it lies: the exits were a layoff, which shortens the window and shrinks the effect, or the team is large, where the uplift is only 7 to 9%.
  • Shelf life: 135 days after a voluntary exit, peaking around Day 70; 105 days after a layoff.
  • Lawful use: yes.
  • Weight: 2 for a voluntary exit from a team of ten or fewer, inside the window; 1 for a larger team or a layoff.

Reorg

  • What it is: an announced restructure affecting the candidate's organization.
  • What it proves: uncertainty, nothing more measured. No reply-rate lift or decay curve for reorgs has been published.
  • How it lies: a paper reorg renames reporting lines and leaves the team intact. Check whether anyone actually left.
  • Shelf life: not published. If the reorg included layoffs, use the 105-day layoff window.
  • Lawful use: yes.
  • Weight: 1.

Passed-over promotion

  • What it is: a role the candidate appeared to be in line for went to a peer or an outside hire.
  • What it proves: very little on its own. It is your inference about someone's ambitions, and no source measures it.
  • How it lies: the candidate is on a different track, never wanted the role, or you misread the hierarchy from job titles.
  • Shelf life: not published.
  • Lawful use: yes, provided the inference rests on role events only. If your reasoning touches on leave, age or health, the cue is excluded.
  • Weight: 1, and it never counts alone.

New-employer honeymoon

  • What it is: the candidate joined their current employer recently.
  • What it proves: nothing measured. I found no published first-year attrition rate or honeymoon window for recruiting outreach.
  • How it lies: tenure tracks age. US median tenure is 4.1 years overall and 6.1 years in management occupations, but 3.0 years for ages 25 to 34 against 9.6 years for ages 55 to 64.
  • Shelf life: not applicable.
  • Lawful use: only as a rule applied identically to everyone. Otherwise exclude it.
  • Weight: 0. At most, use it as a uniform hold rule, never as a score.

Portfolio or repo cleanup

  • What it is: on the public GitHub graph or a personal site, pinned repositories changed, READMEs rewritten, old projects archived, or a portfolio refreshed.
  • What it proves: the same family of behaviour as résumé revision. It has not been measured as a predictor on its own.
  • How it lies: routine maintenance, preparation for a talk, handing off a project, or an internal promotion.
  • Shelf life: not published.
  • Lawful use: yes, unless the content itself reveals special category data.
  • Weight: 1.

Community activity shift

  • What it is: a rise in posting, new connections, new recommendations, or posts using phrases like "looking for" or "seeking". The patent counts all of these as signals.
  • What it proves: explicit seeking language about the candidate's own next role is close to a self-declaration. A generic rise in activity proves little.
  • How it lies: "looking for" posts are often a hiring manager recruiting for their own team. Read the post, not the keyword.
  • Shelf life: not published.
  • Lawful use: a rise, yes. A drop in activity, no. Activity in union, religious, political or health communities, no.
  • Weight: 3 for explicit seeking language about themselves; 1 for a generic rise; never score a drop.

How long does each cue stay live?

Only peer-departure cues have a published shelf life: further quits are most likely between Day 45 and Day 135 after a voluntary exit, peaking around Day 70, and the window closes at Day 105 after a layoff. Every other cue's decay is unmeasured, so treat its shelf life as something you learn from your own reply log.

EventWindow opensPeakWindow closes
Voluntary peer exitDay 0Day 70Day 135
Layoff or involuntary exitDay 0Not publishedDay 105
Manager exit (Pieper)After a delayNot publishedNot published

The practical lesson is that readiness runs on a lag. People wait until the uncertainty about what happens next clears up, which is why Visier's peak sits at Day 70 and Pieper found a delay before disruption. A cue that is too fresh is a reason to hold someone with a date, not to drop them.

The contagion clock after a peer leaves

  1. Day 0
    Peer quits or is laid off; log the event date and hold
  2. Day 45
    Further quits become most likely; manager-exit holds come due
  3. Day 70
    Peak for further quits after a voluntary exit
  4. Day 105
    Window closes if the exits were a layoff
  5. Day 135
    Window closes after a voluntary exit; retire the cue
A peer exit is a reason to log a date on Day 0 and act between Day 45 and the window close, not to message the same week.

Why team size changes the weight

Team size moves the effect of a peer departure by more than 3x. Visier and its researcher Andrea Derler attribute this to interdependence: on a small team, one exit changes everyone's work.

Team sizeAdded quit probability
225.1%
3 to 512.1%
6 to 1014.6%
Larger7 to 9%

Note that 6 to 10 sits above 3 to 5 in this data, so I draw the weight line at ten, not at five. Don't over-read the gap between those two bands; the source reports the figures without the confidence ranges you would need to rank them.

25.1%
Added quit probability after a peer leaves a team of two
On teams larger than ten the same event adds only 7 to 9%, so the two must not carry the same weight.

Dating these events is the slow part: you need the exit date, the team size and whether it was voluntary, for every person on the slate. Refolk lets you ask for the event directly across LinkedIn, GitHub and the open web rather than reconstructing it profile by profile.

What does the open-to-work flag actually tell you?

The open-to-work flag tells you the person has declared openness to the platform, but from outside LinkedIn Recruiter you see only a tiny, self-selected share of those declarations. In Refolk's index of professional profiles, about 1 in 2,740 US Software Engineers puts "open to work" in the headline.

1 in 2,740
US Software Engineer profiles with "open to work" in the headline
In Refolk's index, 129 of 353,403 profiles titled Software Engineer, a share of 0.037%.
MeasureValueNote
US headline count129Software Engineer titles
US total titled353,403Refolk's index
US share0.037%About 1 in 2,740
UK headline count4One is not an engineer

The raw US-to-UK ratio is about 32x, but I did not query the UK total, so it is not normalized for market size. Don't read it as a difference in openness between markets.

Who declares openness in public

  1. US Software Engineer profiles
    353,403

    Refolk's index

  2. "Open to work" in headline
    129

    About 1 in 2,740

  3. Sampled headlines
    25

    Mostly employed, at Amazon/AWS (4), Google, Bank of America, Oracle, Uber

Of 353,403 US Software Engineer profiles in Refolk's index, 129 declare openness in the headline, and a 25-profile sample of those is mostly still employed.

Two things follow. First, the people who declare openness in their headline are mostly employed, not idle: the sample's current employers include large technology firms and a major bank. Second, the likely reason so few declare publicly is that employed people choose the Recruiters-only mode, which LinkedIn says it tries to hide from Recruiter users at the member's own company while adding that it "can't guarantee complete privacy". Open-web sourcing therefore misses most flags.

Which cues may you lawfully use?

You may use cues that rest on role events and self-declarations; you may not use cues that reveal or infer special category data or protected status, and the inference itself is the problem even when it is wrong. Under ICO guidance, profiling that infers health, beliefs, sexual orientation and similar traits is special category processing and needs an Article 9 condition, regardless of how confident you are that the inference is correct.

Article 9(1) lists racial or ethnic origin, political opinions, religious or philosophical beliefs, trade union membership, genetic data, biometric data, health data, and sex life or sexual orientation. Age and parental leave are not on that list; they fall under equality law instead. I have not verified the US position in this reference, so check with counsel or your DPO before relying on any cue in the table below.

CueStatusWhy
Activity drop or career gapExcludeMay infer parental or medical leave
Tenure or anniversaryExclude, or apply identicallyPartly tracks age
Union or advocacy activityExcludeTrade union membership is Article 9
Religious, political or health community activityExcludeInfers Article 9 traits

Some practitioners score first and screen for lawfulness afterwards. I screen first, because under the ICO's reasoning, collecting and weighing the cue is already the processing; removing it from the final score does not undo it.

How do you stack weak cues into this week's batch?

Add the weights of each person's live cues that fall inside one window, and message this week anyone who reaches the threshold; hold everyone else with a dated recheck. This count-or-weight method is the shape described in patent US8352406B2, which triggers on a minimum number of signals or a minimum aggregate signal weight within a set time period.

The patent's own numbers do not appear in the text I retrieved, and no other source publishes a threshold. So the rule below is my starting point, built to be replaced by your reply data:

  • Threshold: a score of 3 from live cues inside one window.
  • Self-declared cues (weight 3): an open-to-work flag with a current role, or explicit seeking language about themselves. These stand alone.
  • Strong inferred cues (weight 2): a voluntary peer exit from a team of ten or fewer, inside its window. These need one corroborator.
  • Weak cues (weight 1): profile rewrite, teammate rewrite, manager exit after Day 45, reorg, passed-over promotion, repo cleanup, generic activity rise, large-team or layoff peer exit. These need two corroborators.
  • Same-event rule: two cues caused by the same event count once. A reorg that removed the manager is one event, not two.

Message this week or hold

Stacked to thresholdSingle weak cue
Stacked but stale
Re-collect dates before sending anything
Live and stacked
This week's outreach batch
Stale and single
Drop the cue, keep the person on the slate
Live but single
Hold with a dated recheck
Stale or undated cuesLive, dated cues
Only people whose cues are both live and stacked to the threshold go into this week's batch; everyone else gets a dated recheck or a dropped cue, never a dropped name.

The batch record below is what the talent lead signs. It keeps holds visible so nobody falls off the slate because their cue was too fresh.

Weekly batch decision record
Slate:
Week of:
Threshold used: 3

THIS WEEK'S BATCH (score at or above threshold)
Name | Score | Live cues (type, event date, weight)

HELD (below threshold)
Name | Score | Recheck date | Cue it waits on (e.g. manager exit, Day 45)

DROPPED CUES (expired or excluded, person stays on slate)
Name | Cue type | Reason (stale / excluded type / same event)

Signed off by:
Date:

One record per slate per week; the recheck date and the cue it waits on are mandatory for every held name.

Step-by-step: from frozen slate to two lists

The procedure runs in seven steps, from a frozen slate to a signed batch list and a held list with recheck dates. No published source sets out this order end to end, and the durations are my estimates: under an hour to freeze the slate, about 15 minutes for the lawful-use screen, and 5 to 10 minutes per person for cue collection.

Ordering a qualified slate by move-readiness

  1. Freeze the qualified slate
    Fix the list of people already judged a fit and stop adding or removing names. Done means fit is decided and readiness work cannot change who is on the list.
  2. Screen cue types for lawful use
    Before collecting anything, strip out cue types that reveal or infer Article 9 data or protected status: health, union activity, activity gaps, leave and age-linked tenure rules. Done means a cue log template with no excluded types.
  3. Collect dated cues for each person
    Record the cue type, the event date, and whether it is visible only in LinkedIn Recruiter or on the open web. Done means every cue has a date.
  4. Apply windows
    Drop expired cues, such as a voluntary peer exit more than 135 days ago or a layoff more than 105 days ago, and flag cues inside the Day 45 to Day 70 peak. Done means every cue is marked live or stale.
  5. Stack live cues
    Add up the weights of each person's live cues inside one window, counting cues from the same event once. Done means one score per person.
  6. Batch and sign off
    People at or above the threshold go into this week's outreach; everyone else is held with a recheck date, such as Day 45 after a manager exit. Done means two lists signed off by the talent lead.
  7. Log replies against cues
    Record which cue triggered each message and whether the person replied. Done means a reply rate per cue type, which is the only way to calibrate weights.

Step 3 is where most of the time goes. When the cue you need is an event at the company level, such as a manager departure or a run of exits from one small team, asking Refolk for people at companies where that event happened gives you the date and the affected names in one pass.

How move-readiness cues go wrong

Every cue in this reference has a known false-positive pattern, and most bad outreach weeks come from one of the eight below. Each entry names the trap, what is really going on, and the check that catches it.

  1. The stale open-to-work flag. It looks like a ready candidate. Often it is a laid-off or freelance person who never switched the flag off, and the 40% lift you were relying on counted InMails received, not replies. Check: does the profile show a current role, and when was the flag last updated? How long flags typically stay on is not documented.
  2. The layoff read as resignations. Several exits from one team look like voluntary contagion. If they came from a reduction in force, the window is 105 days, not 135, and the effect is 7.7%, not 9.1%. Check: match the exit dates against public layoff news before weighting.
  3. The manager exit with an internal successor. It looks like disruption. If someone from the team was promoted into the seat, group turnover may actually fall. And because of Pieper's delay, a Day 1 message is too early either way. Check: who holds the role now, and has Day 45 passed?
  4. The large-team departure counted like a small-team one. One exit from a forty-person org gets the same weight as one from a team of two. The measured uplift is 7 to 9% against 25.1%. Check: estimate the team size from public profiles before assigning weight 2.
  5. The tenure or anniversary cue. "Two years in, ripe to move" looks neutral. Median tenure is 3.0 years at ages 25 to 34 and 9.6 years at 55 to 64, so the rule quietly sorts by age. Check: remove it, or apply one identical rule to every person on the slate.
  6. The activity drop or gap. Silence looks like disengagement. It may be parental or medical leave, and inferring either is protected or Article 9 processing whether you are right or not. Check: there is no safe check; exclude the cue type.
  7. The promotion read as a rewrite. A new headline and title look like someone polishing for the market. If the employer did not change, it is usually an internal promotion, which points the other way. Check: compare title change against employer change.
  8. The "looking for" post from a hiring manager. A seeking phrase triggers the keyword match. The person is recruiting for their own team. Check: read the post and confirm it is about the author's own next role.

Keeping the weights honest

Weights stay honest only if every message is logged against the cue that triggered it, because no reply-rate lift has been published for any cue in this reference. Within a few weekly batches your own reply rates per cue type will replace my starting weights, and they will reflect your market, your roles and your message rather than someone else's.

Run the log as a flat table, one row per cue per person, and never overwrite a row: when a cue goes stale, mark it stale. That history is what lets you see, for example, whether teammate rewrites predict replies in your market at all.

Cue log (one row per cue per person)
person | cue_type | event_date | visibility | lawful_check | window_close | status | weight | same_event_id | messaged_on | reply
| open_to_work_with_role | YYYY-MM-DD | recruiter_only / open_web | passed | not_published | live / stale | 3 | | YYYY-MM-DD | yes / no / none_yet
| peer_exit_voluntary_team_le10 | YYYY-MM-DD | open_web | passed | event_date + 135 | live / stale | 2 | E1 | |
| manager_exit_external_successor | YYYY-MM-DD | open_web | passed | not_published | hold_until_day_45 | 1 | E1 | |
| profile_rewrite_employer_changed_no | YYYY-MM-DD | open_web | passed | not_published | live / stale | 1 | | |

Paste into a spreadsheet; event_date drives every window, and reply is filled in only after outreach.

Revisit the thresholds and windows on a fixed schedule. The contagion figures here come from one large dataset; if a new study publishes reply-rate decay for reorgs or profile edits, add the window to the reference and date the change. And re-check LinkedIn's open-to-work visibility settings directly before relying on them, because platform privacy behaviour is the kind of detail that changes without notice.

Before you send this week's batch

  • The slate was frozen before any readiness cues were collected.
  • The cue log contains no activity drops, gaps, leave, union, health, religious or political cues.
  • Every cue in the log has an event date, not just a date noticed.
  • Every peer-exit cue is marked voluntary or layoff, with the matching 135 or 105 day window applied.
  • No manager-exit cue is live before Day 45, and every one names the current holder of the role.
  • Cues caused by the same event share a same_event_id and are counted once.
  • Every held person has a recheck date and the cue it waits on.
  • The talent lead has signed the batch decision record.
  • The cue that triggered each message will be recorded next to its reply.

Once the batch is out, the next job is the held list. Put the recheck dates in a calendar on the day you sign off, so that the person whose manager left on Day 10 reaches you on Day 45 with their cue already dated and ready to stack.

Questions practitioners ask

Does the LinkedIn open-to-work flag mean a candidate will reply?

No, not on its own evidence. LinkedIn's published figure says members with the public frame are 40% more likely to receive recruiter InMails, which measures recruiter behaviour, not candidate replies. A reply-rate lift that separates the public frame from the Recruiters-only mode has not been published. Treat the flag as a strong self-declaration, check that the profile shows a current role, and log your own reply rate against it.

How long after a manager leaves should I reach out to their team?

Hold until about Day 45, then check who holds the role. Research on 239 general manager departures found a delay before team disruption, and internal promotions may reduce the turnover that follows. Messaging on Day 1 is usually too early. If someone was promoted from inside the team, downgrade the cue; if the seat is vacant or went to an outside hire, keep it live and stack it.

Can I use time at current company as a sign someone is ready to move?

Only if you apply the exact same rule to everyone, and even then I would leave it out. US median tenure is 3.0 years for workers aged 25 to 34 and 9.6 years for ages 55 to 64, so any tenure or anniversary cue partly tracks age. No measured honeymoon window for recruiting outreach has been published, so the cue adds age risk without proven predictive value.

How many signals do I need before messaging a passive candidate?

No published threshold exists, so set one and calibrate it. The patent that describes this approach triggers on a minimum count or minimum total weight of signals within a set period, without public numbers. My starting rule: a self-declaration such as an open-to-work flag with a current role stands alone, a strong inferred cue needs one corroborator, and a weak cue needs two.

Why can't I find most open-to-work candidates with a web search?

Because most flags never reach the headline. In Refolk's index, only 129 of 353,403 US Software Engineer profiles write "open to work" in the headline. The Recruiters-only mode shows status only to LinkedIn Recruiter users and adds no public frame, and the 25-profile sample of headline declarers is mostly people who are still employed. Open-web sourcing sees a small, self-selected slice.

Is it legal to use a drop in someone's online activity as a move signal?

Exclude it. An activity drop or career gap may reflect parental or medical leave, and under ICO guidance inferring health or similar traits is special category processing whether or not the inference is right. Parental leave is a protected matter under equality law. I have not verified US guidance in this reference, so check with counsel before using any absence-based cue.

Try it on the search you came here for

Stop building boolean strings. Just describe the person.

Type one sentence. I plan the search, read GitHub, public LinkedIn and Crunchbase records, and the open web as it is right now, and hand back a ranked list with the reason next to every name.

  1. 01Describe them

    One plain sentence. Role, city, stack, stage, whatever matters to you.

  2. 02I read the web live

    GitHub, public LinkedIn and Crunchbase records, the open web. Not a database that went stale last quarter.

  3. 03You read the shortlist

    Ranked, with the reasoning under every name. Open a profile, ask a follow-up, narrow it down.

  • No boolean, no filters, no seat to buy. One box.
  • Read at search time, so a profile updated yesterday counts today.
  • Every step visible as it runs, every name with its reason.

500 free credits on sign-up. No card, no demo call. See real searches.

Read next