BSP Revenue Pipeline War Room | Pre-Evelyn Sprint

🔴 Action #1: Fix Offline Conversion Pipeline CRITICAL BLOCKER #1 PRIORITY ▼

🚨 WHY THIS IS #1: No new Google Ads campaigns should go live until this pipeline flows data. Smart Bidding without revenue signal = burning money. This unlocks 30x ROAS multiplier.

$0

Current Revenue Uploaded

$334K

Potential if Fixed

423

ST Jobs Available

0

GCLID Matches Found

🔧 Part A: Fix Systemd Timer Bug (5 minutes)

1

Open the broken service file

The timer calls --process which doesn't exist in argparse. Needs to be --days 7.

systemd/nexus-offline-conv.service

Nexus AI

2

Change `--process` to `--days 7`

Line 10: ExecStart=...nexus_offline_conversions.py --days 7

3

Deploy to VM and reload systemd

gcloud compute scp then sudo systemctl daemon-reload on VM

4

Verify: run the script manually

python nexus_offline_conversions.py --days 7 --dry-run should complete without error

🔧 Part B: Build GCLID Bridge (2-3 hours)

5

Create `nexus_gclid_bridge.py` (~300 lines)

Lightweight webhook receiver on VM port 8502. Receives WordPress form submissions, extracts GCLID, writes to ServiceTitan lead externalData.

Nexus AI

6

Create systemd service for the bridge

nexus-gclid-bridge.service: always-on, listening on port 8502, restart on failure.

systemd/nexus-gclid-bridge.service

7

Configure WordPress to POST form data

WPForms webhook or Zapier: on form submit, POST to https://morpheus.callbrightside.com:8502/gclid with GCLID + contact fields.

Robert

8

Test end-to-end: submit test form with GCLID

Fill out callbrightside.com form with test GCLID. Verify it appears in ServiceTitan lead's externalData.

9

Verify pipeline: run offline conversion script

python nexus_offline_conversions.py --days 7 --dry-run should now find GCLID matches and show upload candidates.

🔧 Part C: First Real Upload (After Bridge Works)

10

Run offline conversions with actual upload

python nexus_offline_conversions.py --days 30 (no --dry-run). Uploads matched GCLID+revenue to Google Ads UPLOAD_CLICKS actions.

11

Verify in Google Ads UI

Check account 7269555791 > Conversions > "Offline Revenue" action. Should show imported conversions with real dollar values.

Robert + Evelyn

💬 Executive Impact: Once this pipeline flows, Smart Bidding learns that "sewer repair Overland Park" drove an $8,500 job. It automatically bids more aggressively on high-value keywords and less on low-value ones. This is the 30x ROAS multiplier.

📝 Notes for Action #1:

📞 Action #2: Fix Phone-to-Campaign Mapping HIGH PRIORITY ▼

🚨 Problem: PHONE_TO_CAMPAIGN in nexus_attribution_engine.py line 109 is entirely commented out. Every phone-based attribution rule short-circuits. 110 abandoned calls in 6 months going untracked.

1

Identify BSP phone numbers from 3CX data

28,719 historical call records. Extract unique destination numbers and their call volumes.

nexus_attribution_engine.py:109-114

Nexus AI

2

⚠️ CONFIRM with Robert: phone number assignments

Proposed mapping needs verification:

Number	3CX Calls	Proposed Source	Confirm?
(913) 276-4920	44	PPC tracking number	❓ VERIFY
(913) 297-9941	62	Main line, Google organic	❓ VERIFY
(913) 963-1029	-	Official BSP, existing customer	❓ VERIFY
(913) 358-0252	-	Need identification	❌ UNKNOWN
(913) 358-0380	-	Need identification	❌ UNKNOWN

Robert

3

Uncomment and populate `PHONE_TO_CAMPAIGN` dict

After Robert confirms, update the mapping. Each number gets a campaign source string.

4

Verify: run attribution scan

python nexus_attribution_engine.py --scan --verbose --days 180. Check phone_stats section shows campaign mapping active.

🧠 Deeper Connection: Phone mapping + 3CX data + ServiceTitan = complete customer journey. When someone calls (913) 276-4920 from a Google Ad, we know: which keyword, which campaign, which agent answered, whether they booked, what service, what revenue. This feeds the Unified Attribution Engine (Phase 6 of the master plan).

📝 Notes for Action #2:

💰 Action #3: Revenue Resolver (Invoice-Based Truth) $97K GAP ▼

🚨 The Revenue Truth Gap: ST Jobs API job.total shows $227K for Jan-Feb. Kalen's verbal numbers are $334K. That's a $97K (30%) gap. Every ROI calculation is wrong because we're using quoted amounts, not invoiced amounts.

$227K

job.total (Quoted)

$334K

Kalen's Verbal

$97K

Gap (30%)

Invoice API

Source of Truth

1

Create `nexus_revenue_resolver.py` (~250 lines)

Shared module. Bulk-pulls invoices: GET /accounting/v2/tenant/{T}/invoices. Builds {job_id: invoice_total} lookup. SQLite cache (re-pull if > 24h). Exposes get_revenue(job_id) -> float.

nexus_revenue_resolver.py

Nexus AI

2

Update 4 scripts to use resolver

Replace job.get("total") with get_revenue(job_id) in:

Script	Line	Current
nexus_lead_roi_engine.py	318	`job.get("total")`
st_ticket_by_service.py	71	`job.get("total")`
nexus_attribution_engine.py	328	`job.get("total")`
nexus_cpl_monitor.py	345	`job.get("total")`

3

Update dashboard fallback values

Replace _FALLBACK_REVENUE_WEEKLY = 57_700 (line 102) and _ap_roi_map hardcoded values (line 4214) with invoice-based calculations. Every revenue panel shows "Source: ST Invoicing API" + staleness warning.

nexus_dashboard.py:102, 4214

4

Verify: compare totals

python nexus_revenue_resolver.py --test. Invoice totals should be > job.total. Output: revenue_truth_latest.json

🧠 Executive Panel Impact: The Monday Standup HTML and dashboard revenue panels currently show numbers derived from job.total. Once the resolver is live, EVERY revenue display auto-corrects to invoice truth. Stephanie and Kalen stop arguing about numbers; the system produces the definitive total.

📝 Notes for Action #3:

📍 Action #4: LSA Profile Updates QUICK WIN ▼

✅ Current LSA ROAS: 24.8x. Best-performing channel. These updates amplify what's already working.

➕ 4 New Service Categories to Add:

1

Log into LSA dashboard

Google Ads > Local Services Ads > Profile. Account 7269555791.

Robert

2

Add: Emergency Heaters

Seasonal (Nov-Feb), $1.5K avg ticket, Emergency Eric persona. Currently missing.

3

Add: Water Heater (expanded)

$2.5K avg ticket, 100% booking rate currently. Needs dedicated category.

4

Add: Leak Detection

Gateway service to larger jobs, $800 avg ticket. Research Robert persona.

5

Add: Sump Pumps

Seasonal (March-May), $1.2K avg, Emergency Eric persona. Spring storm season NOW.

📸 8 Priority LSA Photos (Audrey Assignment):

6

Send photo brief to Audrey

8 photos needed. LSA photo ranking is a massive signal since April 2025:

#	Photo	Priority
1	👨‍🔧 Team photo in branded shirts	HIGH
2	🚽 Before/After sewer repair	HIGH
3	📹 Tech with camera inspection	MED
4	🚚 Branded truck	MED
5	🌡️ Water heater install before/after	MED
6	🔍 Leak detection equipment	LOW
7	💧 Sump pump installation	LOW
8	🔥 Emergency heater repair	LOW

Audrey

📝 Notes for Action #4:

🚫 Action #5: Negative Keyword List from North Star NEXUS AI ▼

💬 Context: The Keyword North Star has 234 keywords in 28 clusters. From this data + competitor analysis, Nexus AI generates a comprehensive negative keyword list to prevent wasted ad spend.

1

Pull negative keywords from North Star system

Cross-reference 234 keywords. Identify: DIY terms, career terms, competitor brand terms, irrelevant geo terms, informational-only terms.

Nexus AI

2

Build negative keyword list by category

Category	Examples	Est. Count
🛠️ DIY / How-to	how to, DIY, tutorial, guide, yourself	~25
💼 Career / Jobs	jobs, salary, hiring, apprentice, career	~15
🏢 Competitors	roto-rooter, mr rooter, a1 sewer	~12
🌐 Wrong Geo	Kansas City MO (if targeting KS only), other states	~10
📚 Informational	what is, definition, wikipedia, reddit	~15
🏠 Non-Service	HVAC, electrical, roofing, backflow, appliance	~20

3

Export as Google Ads-ready CSV

Output: output/negative_keywords_google_ads.csv. Ready for bulk upload to all 5 new campaigns.

4

Give to Evelyn for campaign setup

Evelyn applies shared negative keyword list across all 5 new campaigns. Saves $100s/month on irrelevant clicks.

Evelyn

📝 Notes for Action #5:

✍️ Action #6: Ad Copy Drafts per Service CONTENT ENGINE ▼

🧠 Think Bigger: Most companies give their Google rep nothing and let them write generic ads. BSP gives Evelyn: persona-matched ad copy, keyword-researched headlines, field-intel-powered descriptions, and competitive differentiators. Zero competitors do this.

1

Generate 15 headlines + 4 descriptions per campaign

5 campaigns x (15 headlines + 4 descriptions) = 95 ad copy assets. Use content engine personas.

Nexus AI

2

5 messaging themes per campaign

Theme	Headline Style	Example
🚨 Emergency	Urgency + Availability	"KC Sewer Emergency? We're There in 60 Min"
🏆 Trust	Credentials + Heritage	"5th-Gen Master Plumber, 390+ 5-Star Reviews"
⭐ Social Proof	Reviews + Numbers	"KC's Highest-Rated Sewer Repair Team"
📍 Location	Hyperlocal	"Overland Park Sewer Experts, Not a Franchise"
💰 Value	Transparency + Pricing	"Free Sewer Camera Inspection, Upfront Pricing"

3

Export as Responsive Search Ad template

Output: output/google_ads_copy_drafts.json. Organized by campaign > ad group > headlines + descriptions. Ready for Evelyn to paste into campaign builder.

4

Include call-to-action extensions

Sitelinks, callout extensions, structured snippets per service. Pre-built so Evelyn can add immediately.

📝 Notes for Action #6:

📊 Executive Panel + HTML Updates Needed DASHBOARD ▼

💬 Think Bigger: Every action above feeds data into the dashboard and HTML deliverables. Here's what needs updating once each action completes.

Deliverable	What Changes	When
📊 Nexus Dashboard nexus_dashboard.py	• Phone Analytics page: replace "Enterprise Upgrade In Progress" with real 3CX data • Attribution Health page: populate 7-step pipeline from real data • Revenue panels: switch from job.total to invoice-based truth • Replace `_FALLBACK_REVENUE_WEEKLY` with calculated value	After Actions 1-3
📄 BSP_Google_Ads_Conversion_Fix_Plan.html	• Mark Phase 0 as COMPLETED (green) • Update Phase 1 timeline (GCLID bridge this week) • Add Phase 5 details from Evelyn call	NOW + Post-Evelyn
📄 BSP_Monday_Standup_March_2026.html	• Update revenue numbers to invoice-based truth (after Action 3) • Add Google Ads progress section (Phase 0 done, pipeline in progress) • Add 3CX call metrics when available	After Action 3
📄 BSP_LSA_Strategy.html	• Add 4 new service categories (Action 4) • Update with post-Phase 0 conversion value corrections • Add budget split recommendation from Evelyn call	After Action 4 + Evelyn
📄 BSP_Lead_Source_ROI_Report.html	• Recalculate all ROI using invoice totals (not job.total) • Revenue by source will shift significantly • Verdict badges may change (Google could improve from KEEP to GROW)	After Action 3
✨ NEW: BSP_Revenue_Pipeline_War_Room.html This document	Created tonight. Living action plan with saveable notes.	DONE
✨ NEW: evelyn_meeting_agenda_20260311.html	Created tonight. Meeting agenda with timer and saveable notes.	DONE

📝 Executive Panel Notes:

📅 Execution Timeline SPRINT ▼

When	Action	Owner	Impact
Tonight (Mar 10)	🔧 Fix systemd timer bug (5 min) 🔧 Build GCLID bridge script 🔧 Build revenue resolver script	Nexus AI	Unblocks pipeline
Tue Mar 11 AM	📞 Confirm phone mapping with Robert 📝 Generate negative keyword list ✍️ Generate ad copy drafts	Robert + Nexus AI	Assets ready for Evelyn
Wed Mar 11 10am	📞 Evelyn Call • Present pipeline status • Hand off keywords + ad copy • Agree on 5 new campaigns	Robert + Evelyn	Campaign plan locked
Wed-Thu (Mar 11-12)	🚀 Deploy GCLID bridge to VM 🔧 Configure WordPress webhook 🧪 End-to-end pipeline test	Robert + Nexus AI	Pipeline flowing
Fri Mar 14	📊 First offline conversion upload 📄 Update all HTML deliverables 📍 Add LSA service categories	All	Smart Bidding starts learning

📝 Master Notes AUTO-SAVES ▼

💾 All notes auto-save to localStorage every 2 seconds. They persist between sessions. Click Export to download as text file.

Not saved yet

Revenue Pipeline War Room

🔧 Part A: Fix Systemd Timer Bug (5 minutes)

Open the broken service file

Change --process to --days 7

Deploy to VM and reload systemd

Verify: run the script manually

🔧 Part B: Build GCLID Bridge (2-3 hours)

Create nexus_gclid_bridge.py (~300 lines)

Create systemd service for the bridge

Configure WordPress to POST form data

Test end-to-end: submit test form with GCLID

Verify pipeline: run offline conversion script

🔧 Part C: First Real Upload (After Bridge Works)

Run offline conversions with actual upload

Verify in Google Ads UI

Identify BSP phone numbers from 3CX data

⚠️ CONFIRM with Robert: phone number assignments

Uncomment and populate PHONE_TO_CAMPAIGN dict

Verify: run attribution scan

Create nexus_revenue_resolver.py (~250 lines)

Update 4 scripts to use resolver

Update dashboard fallback values

Verify: compare totals

➕ 4 New Service Categories to Add:

Log into LSA dashboard

Add: Emergency Heaters

Add: Water Heater (expanded)

Add: Leak Detection

Add: Sump Pumps

📸 8 Priority LSA Photos (Audrey Assignment):

Send photo brief to Audrey

Pull negative keywords from North Star system

Build negative keyword list by category

Export as Google Ads-ready CSV

Give to Evelyn for campaign setup

Generate 15 headlines + 4 descriptions per campaign

5 messaging themes per campaign

Export as Responsive Search Ad template

Include call-to-action extensions

🚨 UPDATE — $4.88M Open Estimates Found (Apr 14)

What This Means For the Pipe-Fix Effort

Immediate Implications For This Plan

Verification

🔧 POST-APR-14 — Sales-Close Pipe Found + Monday Dashboard Spec

🔧 The Pipe We Missed

🗺️ Roadmap Reorder

🛠️ Config Audit Scope (Decision #4 in the capstone)

📎 The Canonical Strategic View

Verification

Change `--process` to `--days 7`

Create `nexus_gclid_bridge.py` (~300 lines)

Uncomment and populate `PHONE_TO_CAMPAIGN` dict

Create `nexus_revenue_resolver.py` (~250 lines)