Skip to main content

Prospect Phone and Call Execution SOP

Overview​

What: verify a prospect's phone number, run a call block against verified prospects, and record the outcome of every dial in a form the cadence and any future automation can branch on.

When: after an audit report passes the quality checklist and before or in parallel with email discovery. For prospects generated by the prospecting agent, this SOP is the primary path, because those prospects reliably carry a business phone number and frequently carry no contact email at all.

Who: the Implementation Lead, or any team member who has completed the compliance briefing in the SC Outbound Calling Compliance Standard.

Time required: 3 to 5 minutes per prospect for verification. 45 minutes per call block, covering 20 to 30 dials.

This SOP does not cover: what to say on the call (see Prospecting Cold Call Scripts), finding an email address (see Prospect Email Discovery SOP), or verifying the audit itself (see GHL Prospecting Audit Accuracy SOP).


Prerequisites​

Access​

  • Symphony Core sub-account, Prospecting and Contacts modules
  • An assigned outbound calling number. Each caller dials from one number and keeps dialling from it; which number belongs to whom is in the SC Phone Number Register
  • Permission to create and edit opportunities in the Sales Pipeline

Required information​

  • A verified audit report for the prospect, passing the quality checklist
  • The prospect's business name, address, and niche
  • At least one candidate phone number

Upstream procedures​

Prospect generated by the prospecting agent
|
v
Audit report verified (Audit Accuracy SOP)
|
v
>>> Phone verification (this SOP) <<<
|
v
Call block executed, disposition logged (this SOP)
|
v
Cadence advances (Prospecting Pipeline SOP)

Before your first call block​

Read the SC Outbound Calling Compliance Standard Section 1 in full. The identification, new-voice re-disclosure, and stop-request rules are not optional and are not stylistic.


Procedure​

The block is a loop. Steps 1 to 4 happen once, at the top; steps 5 to 7 repeat per prospect.

Step 1: Verify the phone number​

Work the tiers in order. Stop at the first tier that yields a number that is clearly a published business number.

TierSourceHowConfidence
1Prospect cardThe number the prospecting agent returned on the prospect recordHigh. Sourced from Google Maps data
2Google Business ProfileSearch the business, read the number off the profileHigh
3Business websiteHeader, footer, or Contact pageHigh
4Industry directoryTrade association, licensing board, chamber listingMedium
5Third-party toolEnrichment providerLow. See the warning below

Warning on Tier 5. The business-to-business exemptions that make this calling lawful turn on whether the number reached is a business number. An appended or purchased mobile number that cannot be tied to a published listing weakens that position considerably, and much of this ICP is sole proprietors. Do not dial a number that cannot be traced to a published business listing.

Record the tier in the Contact Identified Via field. The field is shared with the Prospect Email Discovery SOP; keep the vocabulary identical across both.

Step 2: Validate before dialing​

Reject the number and return to Step 1 if any of these are true.

  • It is a personal mobile with no connection to a published business listing
  • It resolves to a different business than the one the audit covers
  • The business address is in Florida, Ohio, or Pennsylvania, all of which are closed. See the compliance standard Section 4
  • The contact already carries the suppression tag
  • rating or reviewCount is blank, which is a prospect sent the report without the offer (the Audit Only email, or the October report email). Fill both from the live Google profile, never the report, before the next line. The line applies to prospects already sent the offer as well, because the email rule does not test recency
  • It fails the pre-dial check for the Results-First offer, listed in the bucket taxonomy disqualifiers. Tick disqualified on its own in Prospect Bucket Options, note which line it failed, and do not dial. A record from Florida, Ohio or Pennsylvania is not a disqualification but a configuration fault: flag it

Step 3: Record the contact​

On the contact record in the sub-account:

FieldValue
First name / Last nameThe person's name, and only the person's name. If you have a first name and no surname, leave Last name blank
Business NameThe business, as it appears on the Google Business Profile. A business name never goes in a name field — it breaks every merge field that greets someone, and it is the one data error a reader sees in the email
PhoneVerified number in E.164 format
Native State and Contact Time ZoneThe prospect's state and IANA timezone. The excluded-state guard and the calling window read these
Call Attempt Count0. Not because you keep the count — the workflow C-021.Prospecting - increment call count does that on every dial — but because an empty field is invisible to the Prospecting - never called list. That filter has no is empty operator, so a record left blank is a prospect nobody can see and therefore nobody calls. Set 0 once and never touch it again

Add a verification note on the contact recording how the contact point was found — the tier from Step 1 (Prospect Card / GBP / Website / Directory / LinkedIn / GHL Enrichment / Third-Party Tool) — and the trade or practice type. This is the same tier vocabulary the email discovery SOP uses.

This applies to a record you create mid-call as much as to one handed to you. A gatekeeper who gives you their name and the business is two pieces of information going into two different fields, and a record with no phone number on it cannot be called back.

Apply the tag prospecting-contact-verified. This is the same tag the email discovery SOP applies, and it means the same thing: a human has confirmed this contact point is real and usable.

Step 4: Build the call block​

A call block is S2 follow-ups to prospects who have been emailed the report, and S1 cold dials to prospects who have not. See the Cold Call Scripts. Everything below -- windows, pacing, compliance, dispositions -- binds both.

Group prospects by vertical, because the two ICP groups are reachable at opposite times of day.

BlockWindow (recipient local)Vertical
Morning08:30 to 09:45Trades
Midday12:00 to 13:00Clinics
Afternoon15:00 to 16:00Clinics
Evening17:30 to 18:15Trades

Twenty to thirty dials per block. Two blocks a day is a sustainable pace for a founder with delivery responsibilities.

These windows are recipient-local, so which of them a caller can hold depends on where the caller is. The table is not a roster. A caller seven hours ahead of US Eastern reaches the evening trades block at half past midnight their time, which is agreed in week one and abandoned by week three — and the gap then reads as a discipline problem when it was a scheduling one. State each caller's covered blocks in their contractor profile, and drop the ones they cannot sustain rather than leaving them nominally assigned. Re-check at each US daylight-saving change, because most of the world moves on a different date or not at all.

The first dial is at 08:30, in every state (compliance standard §3). The 07:00 to 07:45 window that field practice favours for reaching contractors before they leave for job sites is not lawfully available. Do not use it.

Step 5: Run the block​

For each prospect:

  1. Open the audit report and tick the buckets the evidence supports (Reading the audit, immediately below this list). The ticks shape the email cadence. The call itself uses the one opener in the Cold Call Scripts and does not lead with a finding. What to do on each outcome is in the Call Outcomes Playbook.

  2. Read the tags. After the pre-dial gate in Step 2: prospecting-audit-sent or prospecting-offer-email-sent makes the call S2 (already emailed); neither makes it S1. The words are the same either way; the tag decides only the after-call step, which is to take the email and add prospecting-values-added on S1. See "Start Here" in Prospecting Cold Call Scripts. Record S1 or S2 in opener_used; no script test is running.

  3. Dial from your assigned number. Manually. Never use an autodialer, and never use an AI voice. Never dial from a number that is not yours — reputation attaches to the number, so a borrowed line spends someone else's.

  4. Identify yourself within 10 seconds — your own name, Symphony Core, and the reason for the call. The recording notice is played to the prospect by the number itself on answer, so say nothing about it; say the line yourself only if a new voice joins. The lines are in Prospecting Cold Call Scripts §7, Mandatory lines.

  5. If they object to being recorded, end the call. The recording is set on the number and cannot be stopped mid-call, so there is no continuing unrecorded. Log the disposition recording-declined.

  6. If they ask to be removed, acknowledge within 10 seconds, hang up, and apply the suppression tag before the next dial.

  7. Log the disposition from Step 6 on every dial, including no-answer.

  8. Do not touch Call Attempt Count — it increments itself on every outgoing dial. Voicemails have no counter: only two are scripted, voicemail 1 on call attempt 1 and voicemail 2 on call attempt 4, which is the cap of 2. Never leave a voicemail on any other dial.

  9. If a voicemail was left, send nothing extra. An S2 prospect already has the email the voicemail points to, and an S1 prospect has no address yet. The archived after-voicemail template (Template 22) is never sent.

  10. Never commit another person's time in writing until it is on a calendar. Not the Implementation Lead's, not a colleague's, not your own outside your block. Check the slot, book it, then write it down. A prospect told in an email that someone will call at a named time has been given a promise the business has not made. Send the booking link instead — symphonycore.com/meet-with-rohit, which is already in the snippet. The prospect picks a slot that is genuinely free, and nobody has to check a calendar to answer.

  11. Make sure the report has somewhere to go before the call ends. S1: the S1 close: write the address to the contact and send the follow-up email snippet before the next dial. Do not add prospecting-values-added as well; the snippet is their email. S2: they already have it. If they say it never arrived, resend the link with the snippet below. Never send the archived after-call templates (Templates 21 and 22) or Email 1A. Text only when the prospect asked for it and the number accepts texts — a missed-calls bucket may mean it does not.

    The snippet is C-021.Prospecting Send Marketing Audit. Subject and body are pre-filled and the audit URL is the merge field {{contact.audit_pdf_url}}, so it is always the right link. The only thing you type is the finding — one sentence with one number about their business, in place of the bracketed slot. Nothing about what Symphony Core sells goes in it.

    Two checks before you press send: the audit link is visible in the preview — it renders empty on a contact created during the call, because no audit has been run on that record — and the finding has replaced the bracket.

Reading the audit: what to tick and what to say​

Do this with the report open, before the line connects. About two minutes per prospect. It is the difference between a finding and an opinion, and it is what stops the call becoming improvisation.

Prospect Bucket Options (contact.prospect_bucket_options) is a checkbox field, so it holds more than one value. Tick every bucket the evidence supports. The ticks are the menu of things you can raise on the call, not a single verdict.

Tick thisWhat tells youWhat you sayWhat SC does about it
review-engineRating 4.0 or higher and under 15 reviews, counted on the live Google profile, never off the reportTheir rating and count, and how few new reviews in the last two months. Never a competitorA post-job review request that goes to every customer
review-response15 or more reviews, and fewer than 5 of the newest 10 carry an owner replyHow many recent reviews have no reply, and the date of the oldestReviews AI on a 24-hour delay, anything under three stars held for a person
missed-callsBusiness Critical Information → Text Enabled Business Number = No, or no after-hours path — and the trade is emergency-shapedWhat happens to a text sent to the number on their Google listingMissed-call text back. Never promise a date — A2P registration takes 7 to 10 business days
no-contact-pathBusiness Critical Information → Chat Widget = No, and no form found by hand anywhere on the siteThe only way to reach them is to pick up a phoneA form and chat widget routed into the CRM
no-bookingTechnology Stack → no scheduling tool detected, and the trade is appointment-shapedSomebody who wants to book has to call during business hoursA calendar with routing

The spoken wording for each is in Prospecting Cold Call Scripts §4, which gives the line to say for every row; what not to say is in §10. The never say for each bucket is in the Prospecting Bucket Taxonomy and it binds here: never imply review filtering, never quote the widget's response-rate percentage, never put a number on what a missing form or a dead text line is costing them.

Five rules that make the multi-select safe:

  1. More ticks is better, not worse. Each one is another thing you can talk about when the first does not land. A prospect who ticks three is not badly slotted.
  2. The lead is the highest ticked, reading the table top to bottom. That is the finding you open with, and the one that goes in Audit Finding Used. The order is deliberate: reviews lead because the gap is a number the prospect can verify on their phone while you are still talking.
  3. parked-no-match and disqualified are exclusive. Either is ticked on its own or not at all. Ticking one alongside a real bucket is a contradiction — it says both that there is nothing to sell and that there is.
  4. Never leave it blank after a dial. Blank does not mean no problems found; it means nobody looked. parked-no-match is the value for a prospect with no wedge worth leading with.
  5. Tick from evidence, not from impression. If you cannot point at the line in the report or the thing on the profile, do not tick it.

Two rows are gated on the trade, and that gate is the one people skip. missed-calls needs emergency-shaped work — plumbing, HVAC, electrical, septic, restoration, garage door, locksmith, storm tree work — because a caller with a burst pipe will not wait, and a dead text line costs the business the job. no-booking needs appointment-shaped work — landscaping, cleaning, flooring, painting, pest, remodeling, inspections, roofing, general construction. A trade that is genuinely both is emergency-shaped. The full table is in the taxonomy.

None of this needs the local heat map. The review buckets are hand-counted on the live Google profile and the three capture buckets come from Business Critical Information and Technology Stack. A report whose heat map failed to generate still supports every row in the table above, so a missing map is never a reason to improvise an opener.

Write Audit Finding Used as a short noun phrase carrying the real number, built from the lead bucket, so it reads inside a sentence. Not a sentence, and never a percentage.

Step 6: Log the disposition​

Every dial gets exactly one disposition. This taxonomy is the schema the cadence branches on and the schema any future automation will bind to, so the values are fixed. Do not invent new ones without amending this table.

appointment-set outranks connected-interested. Interest with nothing on the calendar is not a booking, and recording it as one is the single easiest way to make a week of calling look better than it was.

Where the value is recorded. The sub-account carries a native Custom Dispositions list (Settings → Phone System → Custom Dispositions). It is editable, it triggers workflows, and it holds a maximum of ten values — verified in the UI on 2026-10-29, where the Add button disables at ten and re-enables the moment a value is deleted. The list is set to these ten:

Appointment Set · Connected Interested · Connected Deferred · Connected Not Interested · Connected Do Not Call · Gatekeeper Routed · Gatekeeper Blocked · Voicemail Left · No Answer · Bad Number

The dropdown name converts to the canonical value mechanically: lowercase it and replace spaces with hyphens. "Connected Deferred" is connected-deferred, and that is the string the CALL note carries.

Thirteen behaviours, ten dropdown values. The cap does not reduce the model, because the three outcomes that lost their own value are each distinguished by a field that already exists:

BehaviourPick in the dropdownWhat distinguishes it
connected-disqualifiedConnected Not Interestedloss_reason = Poor Fit on the opportunity
wrong-businessBad Numberobjection:wrong-business in the CALL note, plus the audit flagged for refresh
recording-declinedConnected Not Interestedobjection:recording-declined in the CALL note

The CALL note still carries the precise canonical value in its disposition field, so the weekly count and any later analysis keep all thirteen. Only the dropdown is collapsed.

Nothing automated reads the disposition today. The contact custom field Last Disposition carries the same thirteen values, and the live prospecting workflows — C-021 01.generate audit, 02.send email and WF-3 — do not branch on it. The never-built C-024 design did, which is why earlier drafts of this SOP treated the custom field as load-bearing; it is not.

Set the native disposition. Last Disposition is optional and will be retired. It is kept for now only so the thirteen canonical names stay visible on the contact record while the ten-value dropdown beds in. When a workflow is built that needs to branch on an outcome, build it on the native disposition, which triggers workflows directly.

This table is the single source for what a disposition does to the opportunity. Every other document points here. Four columns, because three of them were being carried in people's heads: the disposition is what happened, the stage is where the opportunity ends up, the next action is what you do about it, and the last column exists because "move the opportunity" with no named owner is how an opportunity stays where it is.

DispositionMeaningOpportunity: stage and statusNext actionWho moves it
appointment-setA specific date and time was agreed with a decision maker and booked on a calendar→ Appointment Set, status stays openWrite booked:yes in the call note and add the row to the daily tracking sheet. The cadence stopsThe caller, by hand
connected-interestedSpoke with the decision maker, they want the report or a meeting, but nothing is on the calendarNo moveBook if you can. Tag prospecting-responded, send or resend the audit—
connected-deferredSpoke with the decision maker, who deferred and named a periodNo moveSet the re-engagement date first, then tag prospecting-not-ready-now. See the rule below—
connected-not-interestedSpoke with the decision maker, explicit no→ Closed Lost + status lostSend the audit once, then close. Set loss_reasonThe caller
connected-do-not-callAsked to be removed from the list→ Closed Lost + status lostSuppress every channel the same day — DND plus sc-suppressed, phone, email and textThe caller, same day
connected-disqualifiedSpoke, but they fail the ICP screen→ Closed Lost + status lostDo not re-engage. Dropdown: Connected Not InterestedThe caller
gatekeeper-routedReached a gatekeeper, obtained a name or a callable windowNo moveRetry inside the stated window. Does not count toward the attempt cap—
gatekeeper-blockedGatekeeper declined to routeNo moveRetry once at a different hour, then email only—
voicemail-leftVoicemail leftNo moveNothing to send (Step 5, item 9)—
no-answerRang out, or you ended the call before reaching a personNo moveRetry at a different hour, on a different day—
bad-numberDisconnected, wrong number, or faxNo move while another number existsReturn to Step 1. Only if no number can be verified, → Closed Lost + status abandonedThe caller, and only at the end
wrong-businessThe audit resolved to a different business→ Closed Lost + status abandonedFlag the audit for refresh. Dropdown: Bad NumberThe caller
recording-declinedObjected to being recordedNo moveEnd the call — the recording cannot be stopped mid-call. Do not redial without a decision. Dropdown: Connected Not Interested—

⚠️ Appointment Set requires a calendar event. A stage move without one is an exception, not a booking. A callable window from a gatekeeper is gatekeeper-routed, which moves nothing — even when a specific time was named, and even when it was put in writing to the prospect. The test is whether a decision maker agreed and whether it is on a calendar, not whether a time was mentioned.

Thirteen rows, and one case that is not in the table. Every row above assumes a dial happened. A lead rejected at review — Prospect Bucket Options ticked disqualified — closes → Closed Lost + status abandoned without any disposition, because no call was made. It is the only way an opportunity leaves the pipeline without a dial. The reasoning is in the next section.

⚠️ The sc-suppressed tag does not exist in the account. The connected-do-not-call row above asks for it and it cannot be applied — verified against the live tag list on 2026-10-30. Until it is created, suppress a do-not-call prospect with DND on every channel and record the request in the CALL note, and do not treat the tag as a control that exists. The same absence blocks the suppression check in the C-021 spec.

Stage and status are two fields, and both have to move​

A GHL opportunity carries a pipeline stage — where the card sits on the board — and a status, which is one of open, won, lost or abandoned. They are independent, and moving the stage does not change the status.

Moving the stage and leaving the status open is the failure mode. The card looks closed, and the opportunity keeps counting in every forecast, every open-pipeline filter and every pipeline-value figure. It is invisible precisely because the board looks right.

lost and abandoned are not interchangeable:

The discriminator is whether a conversation happened — not who decided. In connected-disqualified it is SC that says no rather than the prospect, and it is still lost, because the opportunity was worked and engaged before it ended.

StatusMeansDispositions
lostThe opportunity was worked and engaged, and ended negatively. Either party can have ended itconnected-not-interested, connected-do-not-call, connected-disqualified
abandonedNo conversation ever happened. Nobody was reached, the prospect was never reachable, or the lead was closed out before anyone calledwrong-business, bad-number with no alternative, attempt-cap exhaustion, and a prospect closed at review before any contact

A lead rejected at review is abandoned, not lost. When Prospect Bucket Options is ticked disqualified, the verdict came from a screen and not from a conversation, so there is no decline to record. This is the one row that fires without a dial, and it matters most on externally sourced leads, which arrive unscreened and in volume.

Collapsing the two destroys both of the numbers worth having, and they are different numbers:

  • lost measures the pitch. Of the people we actually spoke to, how many said no.
  • abandoned measures the list. How much of what we were given was never worth a conversation.

Pad the first with prospects who never answered the phone and the rate means nothing. Push rejected-at-review leads into lost and they drown every real loss, because on an externally sourced batch they can easily be the largest category of all.

Lost reason. A lost status takes a native Lost Reason, chosen from the account's list rather than typed. ⚠️ Only one reason exists — "Not interested in any marketing" — so every lost disposition collapses onto it or onto nothing, and the three dispositions that close an opportunity cannot be told apart afterwards. Live records confirm both outcomes occur: of the opportunities closed lost on 2026-10-30, one carries that reason and one carries none. Until more reasons exist, set it by hand on every lost close and treat a blank as an exception, and read the disposition automations before assuming the automation does it for you.

⚠️ The native list is not the loss_reason custom field the pipeline SOP describes. Two parallel systems; the native list is the one the status update actually accepts, and it is the one to standardise on.

Nine of the thirteen do not move anything. That is the normal case and it is deliberate: a prospect who did not answer has not changed their mind about anything. The four that move are the three ways a prospect leaves the pipeline and the one way they advance.

The same thing as a decision tree​

Four rules the table assumes​

1. A voicemail is not progress. voicemail-left moves nothing, and the cap of two is a lifetime cap across the whole cadence, not two per day. Three or more voicemails performs worse than none, so the second one is the last one that prospect will ever get from us.

2. Attempts are spread across days, and the cap is counted in attempts, not dials. The caps are 5 attempts and 1 dial per prospect per day. Three dials to the same prospect in an afternoon spends most of a lifetime budget before the prospect has had a chance to be available at a different hour.

Call Attempt Count counts dials, and it is set for you. Since 2026-11-02 the workflow C-021.Prospecting - increment call count adds one on every outgoing call, so the field is a reliable answer to has this prospect been called and a rough measure of how hard. It is not the cap. It cannot be, because it counts each number in a multi-number run separately and it counts busy and declined lines, neither of which is an attempt under the rule below. The cap of 5 is judged from the call log by the daily review, which collapses a day's dials to one attempt. Do not edit the field to make it agree with the cap; the two measure different things on purpose.

3. An inbound callback is the hottest event in the pipeline, and it now reaches you. A prospect returning a cold call has already decided to spend their attention. Treat it as live: thank them for calling back and go into the opener ("I'd called because most owners we talk to…"), not "how can I help you". Log it like any other call, with IN in the situation field:

CALL | 2026-10-28 | IN | n/a | connected-interested | 4:10 | primary:yes | booked:no | objection:none

IN is a third situation value alongside S1 and S2. The weekly count reads the CALL | prefix, so it keeps counting; only the per-variant breakdown excludes these, which is correct — an inbound callback is not a test of an opener. Your number rings you first and takes a voicemail if you miss it. So two things follow: an unanswered callback leaves you a message rather than vanishing, and clearing voicemail is part of closing a block — a message left overnight is a prospect who concluded you were not interested either. What to say on the greeting, on the callback and on the missed-call text is in Inside Sales Inbound and Voicemail Scripts.

4. A drop under ten seconds is not its own disposition. Two different events look identical in the call report, and only the recording tells them apart:

What happenedLog it asIn the objection: field
You ended the call on an IVR, menu or hold queue without reaching a personno-answerivr-abandoned
They answered, heard you, and hung up without speakingconnected-not-interestedhung-up-silent

The first consumes an attempt and nothing else. The second is a decision — a fast one — and it stops the cadence like any other no.

5. connected-deferred is not valid without a date. The value used to be called not now, which is what the prospect says rather than what happened. A deferral is an outcome and it has a date attached; "not now" is a mood. A deferral with no date set is indistinguishable from connected-interested and behaves worse — nothing schedules it, nobody chases it, and the prospect is never called again.

If they would not name a period, they are connected-not-interested, not a deferral. Press once — "is that a few weeks or after the new year?" — and take the answer you get.

Order of operations when the dial came from a task. Should a future cadence queue dials as Manual Call tasks: set the disposition first, then mark the task complete. Completing the task is what resumes the workflow, and the next step reads the disposition — a task completed before the disposition is set makes the workflow branch on the previous call's value. No live workflow works this way today — the design that did, C-024, was never built — but the habit is worth keeping so it is automatic if one is.

Then add one call note, on every dial. Last Disposition is overwritten by the next dial, so the per-dial history lives in a contact note in this exact one-line format:

CALL | <date> | <situation> | <variant> | <disposition> | <length m:ss> | primary:<yes|no|n/a> | booked:<yes|no> | objection:<short text or none>
CALL | 2026-10-08 | S2 | T3-B | connected-interested | 2:40 | primary:yes | booked:no | objection:none
  • Situation: S1 or S2 (older notes also carry S3, retired 2026-11-09). Variant: the value in opener_used.
  • Primary: whether the call hit its situation's job. S1: an email captured or the report requested. S2: reached the demo ask. Use n/a when nobody picked up.
  • One line per dial, starting with CALL |, nothing else in the note. The weekly count reads these lines, so a reworded line drops out of the count.
  • If the prospect asked for the report by text (S1), add text-requested to the objection field. That records their spoken request for the one text.

Step 7: Apply the stop rules​

RuleLimitWhy
Dial attempts per prospect598 percent of conversations that will ever happen have happened by the fifth attempt
Voicemails per prospect2Reply rate roughly doubles at one to two. Three or more performs worse than none
Dials per prospect per day1 reached lineSee the exception below — working several numbers for one business is one attempt, not several
After an explicit noStopSend the audit once, then close
After a do-not-call requestStop permanentlyCompliance obligation, not a courtesy

gatekeeper-routed does not count toward the attempt cap, because it produced information rather than consuming an attempt.

The five attempts fall on business days 1, 2, 4, 7 and 10, alternating early and late in the caller's block, so each prospect is tried at different hours across about two weeks. After attempt 5 with no conversation, the opportunity closes Closed Lost, status abandoned.

A live conversation that drops is called back once, straight away: "Sorry, I think we got cut off." It is the same attempt, not a second dial. A prospect who answered, heard the caller and hung up without a word is not called back; that is hung-up-silent (rule 4 above).

Working a prospect's several numbers in one sitting is one attempt. A business with a main line, a mobile and a dispatch number is one prospect; dialling all three back to back until one rings through is the procedure working, not three attempts against the cap. The cap exists to stop a prospect being pestered on the same line across a day.

Two conditions, and they are what separate this from pestering:

  • The numbers are different published numbers for the same business, taken from the contact's additional phone fields or from Step 1's tiers
  • They are dialled in one continuous run, not spread across the day

Log one disposition, for the line that produced the outcome. A busy or declined line that you immediately move on from is not an attempt — though the counter will have counted it, which is why the cap is read from the call log rather than from the field.

This rule was written after it flagged the right behaviour as a breach. On 2026-10-29 a caller dialled three numbers for one electrical contractor in 33 seconds — busy, busy, then a 2 minute 33 second conversation that was the longest call of the day. The daily review reported it as a stop-rule exception. The caller was correct and the rule was wrong.


Verification​

Each morning during an active calling period​

  • Check the saved Smart List "Prospecting - active with open task > 1 day" (contacts tagged prospecting-outreach-active with an open task older than one day). Anything on it is a stalled cadence: clear the task with a disposition, or escalate. This is the stall control. The automated daily sweep was specified in C-024, which was never built, so this check is not interim — it is the only one there is

Per call block​

  • Every dial has a disposition logged, including no-answer
  • Call Attempt Count moved on every contact dialled. If it did not, the increment workflow has stopped firing — tell the Implementation Lead, do not patch the field by hand
  • No voicemail was left outside call attempts 1 and 4
  • No archived after-call or after-voicemail template was sent
  • Any do-not-call request is suppressed, and the suppression was applied before the next dial
  • No prospect exceeded 5 attempts or 2 voicemails
  • No dial was placed outside 08:30 to 20:00 recipient local time
  • No dial was placed to Florida, Ohio, or Pennsylvania

Weekly spot check​

Pull the week's dials and confirm:

CheckHealthy rangeIf outside
Connect rate12 to 22 percentBelow 12, check number reputation at freecallerregistry.com and check the calling window
Average call durationRising over timeMeetings get booked around 3.7 minutes. Under a minute means the opener is not landing
Conversation to meeting5 to 10 percentBelow 5, the audit finding being led with is probably not compelling
Dials per booked meeting100 to 370Above 370, list quality is the likely cause, not the script
Suppression requestsUnder 2 percent of connectsAbove that, revisit targeting. It is also a reputation risk

Monthly​

  • Check every number used for outbound at freecallerregistry.com and file remediation if flagged. The numbers in use are listed in the SC Phone Number Register
  • Confirm the suppression list has been applied to every channel, not only voice
  • Confirm the audit reports being quoted are within the refresh window


Document History​

VersionDateChanges
1.262026-11-12Step 5 item 11: the S1 close sends the follow-up email snippet instead of adding prospecting-values-added, following the call scripts. From the 2026-11-12 review call with the inside-sales agent
1.252026-11-12First dial at 08:30, recipient local, in every state, per the compliance standard. The Implementation Lead set the first dial at 8:30 am on 2026-11-12. The morning trades block becomes 08:30 to 09:45
1.242026-11-12A caller is calling again, so the 2026-11-05 "No one is calling" banner goes. Step 5 item 1 keeps the bucket ticks for the email cadence but no longer says to lead the call with the highest one, which the one-opener scripts (v1.20) replaced. Item 8 drops "(S2 only)" from voicemail 2, which the scripts give every prospect. Rule 3 opens a callback with the opener, not the finding. primary for S2 is reaching the demo ask. Step 7 adds the attempt spacing (business days 1, 2, 4, 7 and 10) and one more same-day exception: a dropped live conversation is called back once, straight away. Links to the new Call Outcomes Playbook
1.232026-11-10Step 5 item 2: one script for every prospect; the tag decides only the after-call step. Reading the audit: the review-engine line is their own rating, count and recent reviews, never a competitor, matching the call scripts. Step 5 item 4 points to the call scripts' mandatory lines (§7), not Section 9, which is now Running a call block
1.222026-11-10Links follow the call scripts' renumbering (v1.18): the engage block is §3, mandatory lines §7, and the do-not-say lines are in §10
1.212026-11-09One call whatever the prospect was emailed. Step 2 fills blank rating and reviewCount from the live profile before the pre-dial check, and applies it to offer recipients too. Step 5 picks S1 or S2 from the two send tags; S3 and the caller's Email 1A send are retired, and the banner says a caller adds prospecting-values-added only in the S1 close. Voicemails sit on attempts 1 and 4 with nothing paired; the after-call item sends nothing archived and keeps the snippet for resends. Inbound callbacks are no longer called S3
1.202026-11-05Step 5 is marked paused: no one is calling, and the first email is sent automatically by the routing rule in the C-021 workflow spec
1.192026-11-04The caller now sends the first email. prospecting-audit-verified only generates the report; Step 5 item 2 adds the S3 send (fill rating, reviewCount, trade, add prospecting-values-added, dial 5 to 10 minutes later). Step 2 gains the Results-First pre-dial check, closed as disqualified without a dial. Rule 10 gains the explicit Email 1A exception: the caller sends it immediately before dialling, so it promises no one else's time. Item 11 sends Template 21 to prospects inside the offer, and item 9 pairs voicemails with Template 22. Step 4 drops the email-first cadence framing, and item 2 drops the A/B test, closed 2026-11-04
1.182026-11-02Step 4 says out loud that the block table is recipient-local, so which blocks a caller can hold depends on where the caller is. Written when the seat changed from a UTC+1 caller to a UTC+2 one, which puts the evening trades block at 00:30 the caller's time — a slot that gets agreed in week one and abandoned by week three, after which the gap reads as a discipline problem when it was a scheduling one. Each caller's covered blocks now belong in their contractor profile, and uncoverable blocks are dropped rather than left nominally assigned
1.172026-11-02Call Attempt Count becomes automatic, and stops being the cap. The workflow C-021.Prospecting - increment call count (Call Details trigger, direction Outgoing) now sets the field on every outgoing dial, so the two instructions to maintain it by hand are withdrawn — Step 5 item 8 and Step 7. The field had been set on zero of the 123 prospecting contacts since it was created on 2026-09-08, which is why the question which prospects have been called could not be answered from the CRM at all; 47 contacts were backfilled from the call log the same day. The field counts dials; the cap of 5 counts attempts, and they diverge wherever several published numbers for one business are worked in one run — exactly the case the 1.14 exception was written for. So the cap is now judged from the call log by the daily review and the field is explicitly not the control. Step 3 still sets 0 on creation, for a reason that had never been written down: the contact filter offers no is empty operator and a numeric comparison does not match an unset value, so a blank field drops the record out of the never-called list entirely. The 76 never-dialled prospects were seeded 0 the same day, and the saved lists Prospecting - never called (76) and Prospecting - called (47) now reconcile to the 123 tagged prospects
1.162026-11-02Step 3 gains a name-field rule and Step 5 names the snippet. Both come from the same live record: a caller created a contact for a gatekeeper mid-call, put the business name in Last name, and left Business Name and Phone empty — so the callback she had committed to had no number attached. Step 3 now says a business name never goes in a name field and that the rule applies to records created during a call, not only to records handed over. Step 5 item 11 names the snippet (C-021.Prospecting Send Marketing Audit), says the only thing the caller types is the finding, and warns that the audit merge field renders empty on a contact created during the call. Item 10 names the booking link as what to send instead of a time
1.152026-11-01Two rules added to Step 5 and one tightened in Step 6, all from a single live call. A caller finished a gatekeeper conversation, then emailed the prospect that the Implementation Lead would call at a named time — nothing was on a calendar, and the opportunity was moved to Appointment Set. So: never commit another person's time in writing until it is booked; send the audit link before the call ends, by snippet rather than composed prose, and text only when the number accepts texts. appointment-set now says "agreed with a decision maker and booked on a calendar", and a stage move with no calendar event is named as an exception — a callable window from a gatekeeper is gatekeeper-routed even when a time was named and put in writing
1.142026-10-30The one-dial-per-day stop rule gains the exception it always needed. Dialling several published numbers for one business in a single run is one attempt, not several — the cap exists to stop a prospect being pestered on the same line. Written because the rule flagged the right behaviour as a breach: on 2026-10-29 a caller worked three numbers for one contractor in 33 seconds and produced the longest conversation of the day, and the daily review reported it as an exception
1.132026-10-30Corrects the missed-calls row shipped hours earlier in 1.12, which dropped the emergency-shaped trade gate and would have had callers tick it on any business with a non-SMS number. Adds the trade-shape rule in full, since it gates two of the five rows and is the step people skip
1.122026-10-30Step 5.1 becomes a procedure instead of a sentence. "Have one specific finding ready" assumed the caller already knew how to get one; the new Reading the audit section gives the evidence, the spoken line and the SC fix for each of the five buckets, and the rules for the new multi-select Prospect Bucket Options field — tick everything the evidence supports, lead with the highest, and treat parked-no-match and disqualified as exclusive. Records that none of this depends on the local heat map, which removes the reason the day-1 caller improvised. Corrects the lost versus abandoned split shipped in 1.11, which was framed as reached-and-declined against never-reached and so covered no case where SC closed a lead before calling: the discriminator is whether a conversation happened, which keeps connected-disqualified in lost and puts a lead rejected at review in abandoned. Adds that case to the matrix as the only way an opportunity closes without a dial. Flags two live constraints verified the same day: the sc-suppressed tag the do-not-call row asks for does not exist, and only one native Lost Reason exists, so lost closes cannot be told apart afterwards
1.112026-10-30Status is added alongside stage, because they are two fields and only one was being named. Moving an opportunity to Closed Lost without setting its status leaves it counting as open in every forecast, invisibly, because the board looks right. Records the lost versus abandoned split — reached and declined against never reached — since collapsing them destroys the only rate that says whether the pitch works. Records that a lost status takes a native Lost Reason from the account list, which is not the loss_reason custom field canon describes elsewhere: two parallel systems, and the native one is what the status update accepts
1.102026-10-29Corrects three references to C-024 as though it were live. It was deprecated 2026-09-17 and never built; the live prospecting workflows are C-021 WF-1 and WF-3, and neither reads a disposition. So Last Disposition is not load-bearing and the dual-write instruction added hours earlier in 1.9 is withdrawn: set the native disposition, which can trigger workflows on its own
1.92026-10-29Step 6 becomes the single source for what a disposition does to the opportunity. The table gains the stage, the next action and the person who performs the move, because nine of the thirteen outcomes previously had no stated opportunity action and the first day of live calling moved nothing at all. The same logic is repeated as a decision tree, and four rules are written down that the table had been assuming: a voicemail moves nothing and the cap of two is a lifetime cap, attempts are spread across days with Call Attempt Count as the control, an inbound callback is logged as situation IN and must reach a person, and a sub-ten-second drop resolves to no-answer or connected-not-interested with the reason in the objection: field. Records that the native Custom Dispositions list now holds the ten values it is capped at, with the three remaining behaviours carried by loss_reason and the CALL note, and that Last Disposition is not load-bearing — the C-024 design that would have branched on it was never built, and the live C-021 workflows do not read it, so the native value is the one to set
1.82026-10-28Step 5's block loop is drawn as a diagram
1.72026-10-28Step 4 drops the spoken recording disclosure and the recorder-timing instruction — the outbound number announces the recording to the prospect on answer — and keeps the 10-second identification. Step 5 and the recording-declined row lose "continue unrecorded", which a platform recording makes impossible; the only response to an objection is to end the call
1.62026-10-27Thirteenth disposition, appointment-set, placed first because it is the outcome the calling role exists to produce. Separated from connected-interested, which is now explicitly interested but nothing on the calendar — the two were being carried by one value, which made interest and a booking indistinguishable in the data. Logging it also means moving the opportunity to the new Appointment Set stage and writing booked:yes in the CALL note, a field the format already carried and nothing had used
1.52026-10-06Step 5 picks the call situation (S1 cold, S2 follow-up, S3 send then call) and the assigned script variant before dialing. Step 6 adds a one-line CALL contact note on every dial, because Last Disposition is overwritten and per-variant results need per-dial history. Supports the call script A/B test with no new custom fields
1.32026-08-27Email-first cadence switch: call blocks re-scoped to optional warm follow-up dials (touch-3 script); cold blocks belong to the call-first variant. Windows, pacing, compliance, and dispositions unchanged and bind both
1.22026-08-27Field-model simplification (PRD v2.2): Step 3 records native State/Contact Time Zone and a verification note (tier + niche) instead of the Contact Identified Via/Prospect Niche/Audit Date fields; Voicemail Count retired — the 2-VM cap is structural (touches 1 and 4 only) with a matching per-block check
1.12026-08-22Added the disposition-before-task-complete order of operations for task-queued dials, and the morning stalled-cadence Smart List check (interim control until the C-024 automated sweep)
1.02026-08-18Initial version. Five-tier phone verification, pre-dial validation gate, vertical-matched call blocks, twelve-value call disposition taxonomy, stop rules, and verification checks
1.42026-09-25New York opened and Ohio and Pennsylvania closed, per the outbound calling compliance standard v1.1; recording disclosure now starts the recorder