⚠️ 16 May – 22 Jun undercounted — sits between the members_master import and the Linkserv live feed
🎯 Lead Center OPEN →
Leads all-time
10,048
1,482 in 30d
Verified conversion
5.4%
lead → paying member
Need follow-up
212
No branch set
3,368
unassignable — the gate
👥 Members Centre OPEN →
Real signups 30d
755
New MRR added 30d
R218k
from those signups
Top branch 30d
148
VBJ Town
Avg debit
R289
per new member
📞 Call Ops OPEN →
Numbers ever dialled
9,204
Glitch leads never called
91%
the leak
Best closers · 303/404
16.4%
vs sprayers 3.5–4.8%
📣 Marketing OPEN →
Spend · 30d
R21.4k
Cost per convo · best
R18
DHS Video (QG)
July campaign
R43
creative fatigue — fix now
💰 Finance OPEN →
Best run in 18 months
73.9%
Jun–Jul · R2.324M
THE WALL · Jan–May 27
R913k
/mo hitting 12-month anniversary
Chronic cohort
R452k
1,695 members · 18-month low
Intel · attribution with receiptssrc: linkserv_members · glitch_attribution_daily
The pipeline · never downgrades
Every lead hunted down.
Cold to Completed in six rounds. Forward or out — nobody hides in this pipeline.
Leads all-time
10,048
1,482 this month · +18% vs June
Pending follow-up
212
signup-sent >24h or engaged >48h
Verified conversion
5.4%
lead → paying member · phone-matched
Branch-blank gate
3,368
0 ever assigned · ≈R30k/mo on the table
Pipeline funnel — where leads sit
Leads generated — full Glitch history 7D · 1M · 3M · 6M · 1Y · ALL
🔥 Hot leads — live tape NEWEST FIRST
Lead
Branch
Stage
Interest
Age
Score
💸 Debit run — actuals & forecast score
R26.4k
August actual · 192 debited · 89 paid (46.4%)
Forecast scoring: V1 R29.4k (+11%) · V2 R52.9k (+101%) · V3 R43.0k (+63%) · V4 R35.4k (+34%) — actual R26,376. Rates were right, the roster was wrong. June R25.7k → August R26.4k (+3%). Ceiling: R123.7k.
the July wave debits 25 Aug / 1 Sep →
📦 Recovered leads — Keap export
4,261
verified-new · 94% carry a branch
Second Touch queue
9
pre-24h-window nudges · dry-run clean
Follow-up queue — oldest waiting
🤖 Glitch V4 — inline brain lanes REAL-TIME ROUTER ON EVERY MESSAGE
✍️ Signups — the real closed deals LINKSERV-VERIFIED SINCE FEB · PHONE-MATCHED TO GLITCH
Total signups
3,542
755 in the last 30d
🤖 Glitch closed online
32
zero human touch
🤝 Glitch assisted
128
convo first, signed at the gym
Machine share
5%
160 of 3,542 deals
Glitch-attributed MRR
R39.7k
recurring · +23.5% MoM
Avg new ticket
R289
vs book ARPU R242
Signups per month — who closed them
Attribution share — is the machine growing?
Latest signups LIVE FEED
Member
Branch
Debit
Channel
⚠️ 16 May – 22 Jun undercounted — sits between the members_master import and the Linkserv live feed
📊 Pipeline — where every lead sits STAGES NEVER DOWNGRADE
Cold → Completed · Glitch V4 validator active · drag disabled in prototype
Due now
478
signup-sent >24h or engaged >48h
Sent today
64
business-hours protected
Second Touch queue
9
pre-24h-window nudges · dry-run
Oldest waiting
14d
Sebokeng · never dialled
📤 Follow-up queue — oldest first FOLLOW-UP ENGINE · BH-PROTECTED
Lead
Branch
Stage
Last contact
Waiting
Action
📦 Recovered leads
5,105
verified-new from Keap exports
🏆 Best funnel
Webinar
11.2% → membership
📞 Call effect
2.4×
called leads convert 2.4× more
🕰 Freshness
94%
carry a branch · more files coming
Funnel → membership — how Keap converted
Contact vintage by year — converted vs open
Recovered lead register SAMPLE · FULL LIST IN PB keap_leads
Lead
Branch
Vintage
Tier
Status
🏋️ What one member costs to close
7.9
dials per deal · across the book
Speed-to-lead sweet spot
<1h
first-hour dials close 2.1× more
🚨 The leak
91%
hot leads nobody has dialled
🥊 Hardest-won deal
23
dials before the yes · Sasolburg
The journey funnel DIALS × DEALS
Speed-to-lead → close rate
Attempts vs close rate
✍️ The closer board WHO TURNS DIALS INTO DEALS
Agent
Style
Dials
Deals
Close %
Read
MY READThe 303/404 style — fewer numbers, more attempts each — closes 16–17%. Sprayers touch everything once and close 3.5–4.8%. Effort per lead beats volume across leads. Every time.
📋 All leads 10,048 ALL-TIME · SHOWING SAMPLE
Lead
Phone
Branch
Stage
Msgs
Last contact
🏢 Branch leaderboard LEADS · SIGNUPS · CALL RATE
Branch
Members
Leads MTD
Signups 30d
Call rate
Read
MY READThree Rivers ≠ TR Outdoor — separate entities, never fold. TR club runs 66% first-debit (#2 in the chain); TR Outdoor runs 35.7% (the worst). 3,156 leads still carry no branch at all — the assignment gate is the fix.
🎯 Assigned leads
1,204
across 9 active agents
Assignment rate · branched
70%
leads WITH a branch get assigned
Assignment rate · branchless
0%
0 of 3,368 EVER assigned
Unreachable revenue
R30k/mo
549 branchless with real convos
Agent assignment board
Agent
Branch
Assigned
Contacted
Conversion
MY READBRANCH-BLANK = HARD ASSIGNMENT GATE. Not one of the 3,368 branchless leads has ever been assigned, vs 70% of leads that carry a branch. Blank rate is 65–80% and worsening. Fix the gate, unlock ≈R30k/mo.
Lead Center · stages never downgradesrc: leads · conversations · second_touch_queue · keap_leads
The book · linkserv-verified
12,674 names on the book.
Every signup verified against the debit feed. Growth you can bank, not vibes.
Real signups · 30d
755
vs ~500/mo breakeven
New MRR added · 30d
R218k
avg ticket R289 vs book ARPU R242
Bot contribution · all-time
160
32 online + 128 assisted
Base health
73.9%
paying share of attempted value
Cumulative growth — 30 days
Signups by weekday THE WEEKEND HOLE
Branch leaderboard — signups 30d vs target
Membership mix
Latest signups MEMBERS.SLATER.FIT · LIVE FEED
Member
Branch
Debit
Channel
Whole base
12,674
debit-order members · 8 entities
Paying · jul run
9,597
73.9% of attempted value
In arrears
2,833
R754k no-funds book
Arrears total
R806k
outstanding balances → christine@
Whole base health MEMBER-LEVEL · JUN→JUL
Stable core (paid both)
8,349
Chronic no-funds (2+ runs)
1,695
Slipped (paid→fail)
906
Recovered (fail→paid)
824
Mandate pulled / closed
66
Recovery 30% · slippage 10.2% — the book churns in place
This month — booked vs banked
MY READBooked = contracts written at the desk. Banked = first debit actually paid. The gap is the first-debit coin flip — 46–55% for 18 months, never fixed. Every contract Welkom writes is a coin toss until the debit-date default changes.
Lead → deal reconciliation PHONE-JOIN · MEMBERS_MASTER × LEADS
Cohort
Count
Matched to a lead
Conversion
Full debit base
12,869
696
5.4% of leads convert
Calendar events marked SIGNED
1,551
52 this month
✅ AUTO-MARKED
Arrears members with live convos
438
—
BALANCE LANE
Marketing-signed · this month
128
phone-matched to Glitch convos
Bot contribution · all-time
160
32 online + 128 assisted
Glitch MRR
R39.7k
+23.5% MoM · cohort +R4k/mo
Cost per signup · best
R112
TR campaign · 2 verified off R224
Marketing-signed members — monthly
Glitch debit prediction RUN-FILE × COHORT MODEL
R26.4kAugust actual · 192 debited · 89 paid
46.4%debut pay rate · retention 89%
R123.7kceiling if the whole cohort paid
MY READEvery forecast overshot — the rates were right, the roster was wrong. Next model rebuilds on real first_debit_dates, not activation guesses. The July signup wave debits 25 Aug / 1 Sep — that's when this number jumps.
Members Centre · booked vs bankedsrc: linkserv_members · members_master
The phones · yeastar CDR · 5 branches wired
Dial or die.
91% of Glitch leads have never been called — that's the leak. The closers who dial 303/404 style close 16–17%. Sprayers close 3.5%.
Dials · TR · 17 days
6,573
Yeastar CDR · 6-hourly sync
Answer rate
51%
1.6 attempts per lead
Never-called leads
91%
the leak · itemised below
Closer close rate
16.4%
303/404 · vs sprayers 3.5–4.8%
Daily dials + connect rate
Hourly dial activity ANSWER-RATE HEAT
✍️ Agent scorecards CALLS.SLATER.FIT · THREE RIVERS · 17 DAYS
Agent
Dials
Talk time
Close %
Read
MY READExt 104 is a closer — 3,790 dials, 18h 42m of talk, 42% of talk share. Ext 102 is four dials old. The gap between the best desk and the worst desk is the whole revenue story of this page.
Call status split
Weekday performance
Talk duration distribution
Best time to dial — answer-rate heatmap MON–SUN × 07:00–18:00
Every campaign ranked by cost per WhatsApp conversation, hard-joined to signups. Winners get budget. Losers get cut.
The champ
R18/convo
DHS: Video (QG) · 680 convos
July cost/convo
R43
creative fatigue — fix now
Hard-joined leads
108
5 verified signups · TR camp 2 @ R112
Meta funnel · 30d
742
convos from 8,400 clicks · 412k impressions
Cost per WhatsApp conversation — monthly
Spend & conversations — monthly
Campaign league table MARKETING.SLATER.FIT · LIVE FROM POCKETBASE
Campaign
Status
Spend
Convos
R/convo
CTR%
Verdict
Ranked creatives GENOME · COST PER CONVO
#
Creative
Owner
Convos
R/convo
Score
Call
Meta funnel — impressions → signups
🔬 Live rankings — genome scores AD_CREATIVES × AD_INSIGHTS × HARD JOIN
MY READONE creative = 98% of tracked volume. DHS Video at R18–22/convo vs Dad's statics at R155–1,050. The Forward + Thumbnail aren't in the nightly — biggest untapped leverage in the account. Meta ad naming still blocks performance-on-creative for half the history.
🏆 Mikey's creatives — the bot funnel ad_name = "DHS: Video (QG)" IS MIKEY'S
Creative
Hook
Convos
R/convo
Call
DHS: Video (QG)
"R150 day one" hero
680
R18
SCALE
Phantom — R249 Ghost
outline ghost · no CTA
34
R22
SCALE
Forward — Group Chat
chat-pack forward format
18
R26
TEST
Margin — Handwritten
bold + handwritten pairing
9
R31
TEST
MY READThree proven hooks. The hero — "R150 day one" at R18/convo — feeds the Phantom factory. Every new template tests against this baseline or it doesn't ship.
📁 Bruce's creatives — landing-page funnel EVERYTHING NOT "DHS: VIDEO (QG)"
Creative
Funnel
Convos
R/convo
Call
Static — Landing A
landing page
7
R155
REWORK
Static — Landing B
landing page
2
R1,050
KILL
MY READDad's traffic is cheap and high-CTR but mostly untracked — the landing-page funnel has no hard join. Shared shortlist via PB ad_marks: purple = both starred. Account dark since 10 June on this funnel.
Marketing · winners get budgetsrc: ad_insights · ad_signups_hard · meta API v25
The factory · nightly drop · 23:00
Ad Factory.
ComfyUI + Slater LoRA on the AI PC. Manufactured overnight, CLIP-scored, dropped for review by morning.
Total ads shipped
—
connecting to the feed…
Latest drop
—
…
Nights on record
—
nightly 23:00 · CLIP-scored
Raw library
—
no-text statics · by segment
⚡ The drop —
CONNECTING TO THE FACTORY FEED…
🎨 Swap Studio SERVER-TRUE PREVIEW · THE WHEEL RECOLOURS IT LIVE · RENDER MATCHES EXACTLY
WARMING…
…
PICK COPY + A MODEL — THE PREVIEW RENDERS ITSELF
03 · Template
VOLT skin
04 · The wheel — OVERRIDES THE ACCENT
🗂 Raw statics — no text, straight from the factory
WARMING THE WALL…
🏆 Winners — what the ad spend actually buys GENOME-RANKED
R18DHS Video — hero: "R150 day one" · 680 convos · the engine
R22Phantom — R249 ghost static · gold standard template
R26The Forward — group-chat format · not in nightly yet ⚠️
R112TR campaign — cost per VERIFIED SIGNUP · the receipt
R1,050Dad's Static B — kill recommended
⭐ The shortlist SHARED · MIKEY ★ + BRUCE ★ = PURPLE
Monthly Linkserv run diagnostics. Upload each new run → full-book analysis lands here.
Collected · jul run
R2.32M
R13.4k vs jun · 9,597 paying
Collection rate · of value
73.9%
+1.3pp vs 72.6% jun · 18-month high
No funds · code 02
R754k
▼ R53k vs jun · 2,833 members
Chronic cohort · 2+ runs
R452k
1,695 members failed both runs
Run history LINKSERV · MANUAL MONTHLY IMPORT
Run
Attempted
Collected
Rate
No funds
Hard fail
Dec–Jan = seasonal low (69.2%) · Nov: +R1.03M annual maintenance batch · ~R16k/mo reverses post-run · next run ~12 Aug
Money by run — growth & decline
MY READGreen = banked, gold = bounced, red = mandates gone. The green has held near R2.3M for 18 months while the book underneath it shrinks — efficiency is doing the work that growth should be doing.
Collection rate trend
MY READThe most important line in the company. The last four runs are the first sustained climb in the whole dataset (73.9% high). Every 1pp = R377k a year.
Where the book moved JUN→JUL · MEMBER-LEVEL
Stable core (paid both)
R2.02M
Chronic no-funds
R452k
Slipped (paid→fail)
R218k
Recovered (fail→paid)
R196k
Mandate pulled
R33k
Recovery 30% · slippage 10.2% — they cancel out: the book churns in place
Fail rate by debit day JUL RUN
Day
Debits
Fail
60% of book debits 1st/30th · shift chronics → 23rd–25th ≈ R50–100k/mo
Failed debits by branch — 18-run trend
MY READWelkom — the only branch worse than a year ago (R133k → R174k). Sebokeng — the turnaround. SW5 — the benchmark: flat, tiny, clean.
Branch risk JUL RUN · TR & TR OUTDOOR SEPARATE
Branch
Billed
In arrears
Failed R
Fail rate
The maturity curve — pay rate by months on book 239K DEBITS · 18 RUNS
MY READMembers join paying ~52%, climb steadily, and by three years pay 92% — basically perfect, forever. The whole game is months 0–6. Fix the first debit and shepherd the first half-year.
First debit pay rate by branch BAR = LATEST · DASH = 17-RUN AVG
Above dash = improving (SW5, VBJ) · below = sliding (VER, TRO, SEB) · 90% of fails = no funds → fix debit date at signup
Pay rate by member vintage START YEAR
First 18 months = the kill zone · 24-month survivors pay 90%+ forever
Members by vintage — the halving JUL RUN
Each vintage ≈ half the next: 3,506 → 1,724 → 905
Vintage retention — % staying on book
Collections forecast — next 5 runs SEASONALITY + TREND MODEL
MY READDashed gold = my call: Aug ≈ R2.30M, small October peak, then the December floor. Rate gains and book shrinkage cancel out — forecast is flat. The August run grades this model — hold me to it.
Book health index 0–100 · RATE 50% + CORE 30% + FLOW 20%
MY READCrashed to 27 in the holidays, then climbed seven straight runs to 68 — the healthiest the book has ever measured. The December crash is seasonal: prep window is November.
Member flow — joiners vs leavers per run
MY READJanuary carries the entire year (1,400+ joiners vs ~600 in December), and the net line has been underwater 4 of the last 5 runs. The growth problem is a signup problem, not a cancellation spike.
Revenue quality — ARPU & stable core
MY READARPU went R223 → R242 (+8.3%) with zero churn reaction — proven pricing power. 85–87% of every month's money comes from members who paid last month: durable, fundable, sellable.
MY READJan–May 2027: 3,070 members worth R913k/month all hit their 12-month anniversary. At measured expiry behaviour (~69% gone within a month) that's ~R630k/mo — 27% of all collections — exposed in one window. A T-30 save-offer system live by December turns the wall into the biggest retention win in Slater history.
Intelligence board — the good & the bad 18 RUNS · 240K DEBITS
Finding
Number
Finance · the debit booksrc: linkserv runs · club manager API later
The corner team · polled every 60s
Engine Room.
The VPS, the workflows, the pipelines — everything that has to stay green for the machine to run.
Uptime 30d
99.98%
Kuma · 8/8 monitors green
Active workflows
42
n8n · scheduled + realtime
Executions 24h
1,912
0 failed overnight
Langfuse traces
428
V4 drafts traced · 24h
Workflow schedule — the daily heartbeat
Workflow
Cadence
Last run
Status
VPS — Hostinger KVM4 16GB · UBUNTU 24.04
CPU34%
RAM9.9 / 16 GB
DISK71%
DOCKER11 containers
Stack via Hostinger Docker Manager (hPanel) · PB password pinned to env on reboot