Signals in Outboundry. How does it work?

Three words you need

Everything in Signals is built from three nouns. Learn these, and the rest of the feature explains itself.

That last distinction matters for your bill. People you already own are never signal leads, however many times they appear.

Included monitoring: what runs for free

Before you buy anything, something is already running.

1,000 contacts from your lists are watched weekly for job changes, included with your base plan. This uses no watch capacity and no signal-lead allowance. A banner above your watches links through to list health in Leads.

Job-change alerts and account alerts on people and companies already in your lists are always free, on every tier. They are the highest-converting signal there is, so they are never rationed.

The six watch types

Every watch is one of six types. The type decides the input, the cadence, the capacity, and whether the watch consumes signal leads at all.

Cadence is a plan property, not a per-watch setting. The "How often" field in the creation sheet is read-only and states your plan's cadence for that type. Upgrading is how you check more often.

Post watches are the exception: a fixed 72-hour job regardless of tier. When a post watch completes, it hands its slot back, so one-off post harvesting does not permanently consume capacity.

Watch a profile is the flagship type. It is the only one that turns a single post into a permanent supply of warm leads: every future post by that page, every reactor and commenter, routed into a campaign.

Creating a watch

Two steps, never more. Step 1 is what do you want to watch Step 2 is what should happen.

You can start a new watch from the signal feed, the Watches tab, or the blank state.

Step 1: What do you want to watch?

Pick a template (the quickest path) or pick a watch type from the six cards, each showing its icon, name, promise and capacity. A subhead tells you how many of your watches are still free.

Templates prefill both the type and the form, so you skip straight to step 2:

Step 2: What should happen?

Here you set the watch name, the type's own input and options, where results go (campaign or lead list), and the tag to apply. The "how often" field shows your plan's cadence and cannot be changed.

Before you save, you get a 7-day preview and a plain-English summary with an estimated usage line.

Options by type

Watch people. One source selector covering lead lists, campaign leads and tags, plus a "paste profile URLs instead" escape that reveals a textarea.

Watch companies. A paste box for URLs or domains, then the pivotal toggle: "Also find me contacts at these companies." Without it, we alert you on the company only. With it on, a company signal delivers up to 5 matching people, and those count as signal leads.

Watch a search. Choose by radio between a saved Lead Finder search and a pasted Sales Navigator URL.

Watch keywords. Who to collect (authors only, or authors and engagers), which engagement counts, minimum reactions (5 / 20 / 50 / 100), and max posts per day (5 / 10 / 25 / 50). Authors are the high-intent party; engagers multiply volume fast, which is why the cap exists.

Watch a profile. Which engagement counts, minimum reactions before harvesting, an ICP filter, and "re-check each post for late engagement", which pulls new engagers at 24 and 72 hours, then stops.

Watch a post. Which engagement counts, plus the ICP filter. Nothing else; it is a 72-hour job.

Choosing your engagement type

Every harvesting type offers the same three-way choice, with the trade-off stated up front:

• Commenters only: fewest, warmest

• Reactors only: highest volume

• Reactors and commenters

Start with commenters on a broad page, and switch to reactors when you need to fill a campaign. Your choice changes the usage estimate in the summary, not just the result, so you can watch the projected number fall as you narrow.

ICP filtering

Harvesting a competitor's audience without a filter is how a campaign fills up with students and job seekers. ICP filtering is the answer, and it reuses your existing ICP scoring rather than making you learn a second targeting language.

Where it is available: profile watches, post watches, and company watches when "also find me contacts" is switched on. It is on by default.

How to set it:

1. Pick an ICP group from your ICP scoring settings. The selected group shows its colour dot, its description, how many of its attributes are required, and chips for the first four attributes with a "+N more" overflow.

2. Pick a minimum score to deliver:

   • 40+: loose, more volume

   • 60+: balanced (the default)

   • 80+: strict, best fit only

Filtering happens on delivery. Off-target people never land in a campaign, and they never count as signal leads.

No ICP groups yet? You get a warning block saying we will deliver everyone we find until a group exists, with a link to set one up. The toggle is never a dead end. We do not fail closed and deliver nothing.

On company watches, the ICP block is nested under the "also find me contacts" toggle, because a company-only watch has no people to filter.

The preview and the summary line

Two blocks stand between you and a watch that quietly eats your allowance.

The 7-day preview

Per type, this shows what the watch would have produced last week. For a profile watch, that reads something like: "This page posted 6 times in the last 30 days, averaging 190 reactions. Expect roughly 340 signal leads a month."

Keyword watches flip the block to a warning tone ("This watch would run hot") when the projection is heavy, and state it in your own units: "At this rate you would use your full monthly allowance in 3 days."

The summary line

One plain sentence composed from your form: "We'll check linkedin.com/company/example daily and add everyone who reacts or comments to Competitor switchers." Post watches say "track for 72 hours" instead.

Then the usage line, which either estimates the monthly draw against your plan allowance, or says the honest thing: "Estimated usage: no signal leads. Alerts on people already in your lists are free."

Destinations and the reason token

Every watch has a destination. There is no notify-only mode.

Choose a lead list when you want to review before enrolling anyone. Nothing is ever sent from a list.

Every watch also takes a tag, prefilled from the type or template (job-change, competitor-audience, intent). The tag is what makes signal-sourced leads separable later, in filters and in reporting.

The reason token

Signals' whole value is carried by one merge token.

{{signal_reason}} is filled from the signal's own reason text ("is now VP Sales at Acme Corp, was Director of Sales at Betaco"), so your first message can open with why now.

The enrol modal exposes it as a toggle, "Carry the reason into the sequence", on by default. The lead drawer names the token explicitly so you can write copy against it. Put the token in your first step and the opener writes itself.

The signal feed

The feed is a triage queue, not a dashboard. It opens on To action, and the tab carries a count of unactioned signals.

What a row shows

• Selection checkbox and a tinted icon: the person's initials when we have them; otherwise, the signal-kind glyph.

• Subject: a person, a company, or a group headline like "14 people engaged with Example's post".

• Kind pill: the event type in its own tint.

• Reason: the sentence that becomes {{signal_reason}}.

• Meta: the CRM context that decides whether this matters: "Was a customer at Betaco until March", "2 contacts in Pipeline, stage: Proposal Sent".

• From: the source watch with its type icon, and the age of the signal.

• Actions: Add to campaign, Dismiss, and a row menu with open lead record or see all N people, add to lead list, tag, view source watch, and mute this watch for 30 days.

Unactioned rows are visually unread. Once you action or dismiss a row, the action cluster is replaced by an Actioned or Dismissed pill, so nothing gets double-worked.

The eleven signal kinds

Grouped signals

A post with 31 engagers is one row, not 31. The row expands inline to the first five people with per-person Add to campaign, then hands off to the people drawer via "See all 31 people".

This keeps the feed usable: without grouping, one viral post would bury a whole day of job-change signals.

Triage tools

• Search across people, companies, reasons and watch names.

• Status tabs: To action, Done, All, with a live count in the toolbar.

• Watch filter, multi-select. Arriving from "View signals" on a watch card prefills the filter and switches to All, so you see that watch's whole history rather than an empty To-action list.

• Bulk bar on selection: add to campaign, add to list, tag, dismiss, deselect all.

• Pagination, 10 per page by default.

• An empty state that reads as success: "Nothing left to triage."

Signal drawers

Signal lead record

Opened from a single-person row. The head carries the avatar, subject, kind pill, title at company, source watch, detection time and LinkedIn URL. The body runs in the order a rep needs it:

• Why now: the reason and the CRM context, footed by a note that this is what fills {{signal_reason}}.

• Signal detail: the structured facts behind the event: new role, previous role, detected date; or roles posted, headcount movement, signal type.

• Contact or Company: title, company, location, headcount, industry, and email with a Verified or Guessed pill. When we have no email, it says "Not found yet" rather than showing a blank.

• History: a timeline mixing profile history with your own touches ("Replied to Q1 Outbound").

• Footer: View source watch, Add to lead list, Add to campaign.

Signal people drawer

Opened from a grouped signal. It lists every person behind the signal, with search across name, title and company, select-all (scoped to your current search when one is active), per-person Add to campaign, and a footer with bulk Add to lead list and Add to campaign.

The footer count switches between "Showing 14 of 14" and "6 selected", so you always know what the bulk buttons will act on.

Managing your watches

The Watches tab shows one card per watch: type icon and label, cadence, status, target line, destination, tag, and three stats: signals, signal leads, and last checked.

Watch statuses

The per-watch menu holds: view signals, edit, pause or resume, and delete. The toolbar has search, a multi-select type filter, New watch, and a count line showing active watches, watches left of your limit, and a link to usage and billing.

Editing a live watch

Edit reuses step 2 of the creation sheet, with the header reading "Editing a live watch" and the promise stated plainly: changes apply at the next check, and signals already delivered stay where they are.

Renaming a watch also renames it on its existing signals in the feed, so the From column never points at a name that no longer exists.

Deleting a watch

Monitoring stops immediately. Signals already delivered stay in the feed and in your lists, and the watch slot is returned. Nothing already handed to a campaign is retracted.

Analytics: which watches to keep

The Analytics tab exists to tell you what is working and what to cancel.

The funnel

Five KPIs over 90 days: signal leads delivered, enrolled (as a percentage of delivered), contacted (of enrolled), engaged, and replies (of contacted). The chain answers the only question that matters: did the leads we delivered get used, and did they answer.

Watch efficiency

Per watch, you get leads, enrolled percentage, reply rate, and a plain-English verdict:

• High value

• Healthy

• Free signals: typically your people watches, which cost no allowance and usually carry the best reply rate in the table

• Prune this: typically a broad keyword watch

Read it this way: enrolled percentage says whether the leads were usable, reply rate says whether they were right, and the verdict column names the ones to prune.

Pricing and metering

Signals is an add-on on top of your base plan. Two things are metered and nothing else: watches (the things you track) and signal leads (the people you get).

Tiers

How the meter works

• Top-ups: $10 per 1,000 signal leads, added to the current cycle. Unused top-up leads carry forward.

• Free and never counted: job-change alerts and account alerts on people and companies already in your lists. Volume there is capped by your plan's included monitoring, not by the add-on allowance.

• Included monitoring is a base-plan entitlement: 1,000 contacts, weekly job-change checks, no watch capacity used.

• Tier changes are prorated from the day of the switch, in either direction.

• A completed post watch frees its slot.

Where billing lives: cycle usage, top-ups, tier switching, the usage ledger and export are all in Settings › Billing › Signals. The Signals page itself shows only two usage facts inline: watches left, and a paused-for-overspend note on any watch running hot.

Troubleshooting

• A keyword I'm watching went viral.

  • The max-posts-per-day cap holds the line, and the 7-day preview warns you before the watch is ever created. If the projected draw would exhaust your cycle, we pause the watch ourselves and mark it as Needs attention, with the reason on the card.

• My monthly signal-lead allowance ran out.

  • Delivery stops, and your watches stay configured. The usage bar turns to a warning at 80%. Free job-change and account alerts keep flowing, because they were never metered. A top-up resumes delivery immediately.

• The same person was delivered by two watches.

  • They are deduplicated to one lead and counted once against your allowance. Both source watches are credited in the feed, so your efficiency numbers stay honest.

• A signal lead was already in my workspace.

  • No signal lead is consumed. The existing record is updated, and the event lands in the feed as a signal on a lead you already own.

• My profile watch has stopped producing anything.

  • If the page stopped posting, the watch stays active, costs nothing in signal leads, and the card shows the last check with a zero signal count. Analytics will eventually verdict it as prunable. We do not auto-delete it.

• My post URL was rejected.

  • A private, deleted or unreachable post fails validation at creation rather than silently producing nothing. If a post is deleted while a watch is already running, the watch completes early and frees its slot.

• People engaged with a post after the tracking window closed.

  • On profile watches, the re-check option pulls new engagers at 24 and 72 hours, then stops. Beyond your plan's post window, late engagers are not collected at all. This is a stated limit, not a bug.

• An engager didn't match any ICP group above my threshold.

  • They were filtered on delivery, so they never entered a campaign and never counted as a signal lead. Lowering the threshold to 40+ or switching groups changes future deliveries only.

• I turned on ICP filtering but have no ICP group.

  • The sheet says so inline and states the consequence: that we deliver everyone we find, with a link to create a group.

• Someone edited or deleted the ICP group my watch uses.

  • Edits apply at the next check. A deleted group leaves the watch unfiltered and marks it Needs attention, rather than quietly widening the net.

• A watched person's profile disappeared.

  • This is treated as "no longer lists a company" rather than a job change, which is how a departure reads in the feed anyway. The lead is flagged, and any campaign it is in is paused for that lead.

• My source list was rejected as too big.

  • Capacity is stated on the type card and again as a field hint: 250 people, 25 companies. An oversized source is rejected at creation with the overflow count named, never silently truncated. Split bigger targets across multiple watches.

• My destination campaign is paused, archived or deleted.

  • Delivered leads are held in a list instead of being dropped, and the watch is marked Needs attention. Nothing is enrolled into a campaign that is not running.

• I deleted a watch that had already delivered leads.

  • Monitoring stops immediately, but the signals and leads already delivered stay in your feed and lists. Consumed signal leads are not refunded.

• I downgraded and now have more watches than my new limit.

  • Nothing is auto-deleted. Watches beyond the new limit are paused, and you pick which to keep. The downgrade screen states the count before you commit.

• I downgraded mid-cycle after spending my allowance.

  • Proration applies to the price, not to leads already delivered. The lower allowance takes effect at your next renewal.

• I cancelled the add-on.

  • Watches stop at the end of the cycle and are retained, not deleted, so resubscribing restores them. Leads already delivered are yours and stay. Included monitoring keeps running on your base plan.

• I actioned some people in a grouped signal, but the row is still open.

  • That is intentional. Per-person enrolment does not mark the group row Done, because the remaining people are still work. The row stays actionable until you action or dismiss it.

• I muted a watch, but it's still delivering leads.

  • Muting suppresses the feed only, for 30 days. The watch keeps running and keeps delivering to its destination. Muting is a triage action, not a pause. Use Pause if you want it to stop.

• I edited a saved search in Lead Finder.

• The watch follows the search. The next check uses the new definition, and the watch card's target line updates to match, so the two can never silently diverge.

  ---

Frequently asked questions

• What exactly is a signal lead? A person a watch delivered who was not already in your workspace. People you already own are not signal leads, however many times they appear.

• So job changes are genuinely free? Yes. Alerts on people and companies already in your lists never count against your allowance.

• How fast is a signal? Your plan's cadence: weekly on Lite, daily on Signals, every 6 hours on Pro. Keyword watches run every 6 hours on the Signals tier. Post watches track continuously for 72 hours.

• Can I change how often a specific watch checks? No. Cadence is a plan property, shown read-only in the sheet. Upgrading is how you check more often.

• Do I need a LinkedIn account connected? Monitoring runs on our side. A connected LinkedIn account matters for what you do next (connecting, messaging), not for the watch itself.

• Why only 250 people or 25 companies per watch? That is the unit of work behind one check cycle. Bigger targets split across watches, which is what makes the capacity honest rather than best-effort.

• Should I pick reactors or commenters? Commenters are far fewer and far warmer; reactors are volume. Start with commenters on a broad page, and use reactors when you need to fill a campaign.

• Will a competitor watch fill my campaign with junk? Not with ICP filtering on, which is the default. Pick an ICP group and a minimum score, and off-target engagers are dropped on delivery, uncounted.

• Can a watch just notify me without adding anyone anywhere? Every watch has a destination, either a campaign or a list. Choose a list when you want to review before enrolling anyone; nothing is sent from a list.

• How does the first email know why I'm writing? The signal's reason fills {{signal_reason}} in at enrolment. Put the token in your first step, and the opener writes itself.

• Can I edit a watch without losing what it found? Yes. Changes apply at the next check, and everything already delivered stays where it is.

• What happens to a post watch when it finishes? It moves to Completed and hands its slot back. Its signals and leads stay in your feed and lists.

• Why did you pause my watch? Only for overspend: the projected draw would consume your cycle's allowance in days. The card says so, and resuming is one click once you have upgraded or topped up.

• Which watches should I keep? Read the efficiency table. Enrolled percentage says whether the leads were usable, reply rate says whether they were right, and the verdict column names the ones to prune.

• What happens if I cancel the add-on? Watches stop at the end of the cycle and are kept, so resubscribing restores them. The leads you already have are yours. Included monitoring keeps running on your base plan.

• Does Signals work outside LinkedIn? Not in v1. Every watch type reads LinkedIn today.

  ---

  What Signals does not do

So you know where the edges are:

• It is not a firmographic database. We monitor; we do not sell you a data warehouse.

• It is not real-time. The fastest cadence is every 6 hours, on the top tier. A watch is a subscription to change, not a webhook.

• It is not a scraper you configure. There is no proxy, session or rate-limit surface. Cadence and caps are set by your plan.

• It is not a sequence builder. Signals hand leads to campaigns; what happens next belongs to Campaigns.

• It is LinkedIn only in v1.