MERCURY
CASE STUDY · SALES COMMAND CENTRE

How one sales organization put $7.85M through a single command centre in nine months, and closed 30% more deals per campaign

An operational diagnostic found the real constraint wasn’t leads or AI. It was the twenty minutes of manual work between a customer saying “yes” and the money being collected. We removed it.

$7.85M

booked through the platform in its first nine months

2,367

deals processed end-to-end

$2.24M

booked in a single 10-day campaign window

30%

more deals closed per campaign, reported by the Head of Sales

№ 01EXECUTIVE SUMMARY
CLIENT

A personal-development education company, referred to here as the Client’s Sales Team. Launch-based revenue, programs from $1,997 to $29,997.

PROBLEM

Campaign sales operations spread across five systems and hand-built spreadsheets; every deal meant manual payment math, checkout links, contracts and tracking.

ENGAGEMENT

Operational diagnostic (August 2025) → custom sales command centre, in production three weeks later.

RESULT

$7.85M booked through the platform in nine months across three campaigns. The Head of Sales reports 30% more deals closed per campaign.

№ 02THE COMPANY

A quarter’s revenue, closed in days.

The client sells coaching and education programs through quarterly launch campaigns. A campaign compresses months of revenue into days: hundreds of sales conversations, deposits, scholarships, payment plans and contracts, all closing inside a short window. The sales organization, 20+ people under a Head of Sales, ran on GoHighLevel (CRM), Kajabi (course delivery), Stripe (payments) and Zoom (sales calls). Holding it all together: spreadsheets.

WHAT THE CLIENT ASKED FOR

An AI implementation roadmap. The expectation, reasonable in 2025, was that the next lever was AI: agents, content automation, more top of funnel.

№ 03WHAT THE DIAGNOSTIC FOUND

The constraint was not lead volume.

The audit’s blunt heading was “Extensive Manual Workload and ‘Spreadsheet Madness’”, and its blunt sentence was: “Most operations are run manually through different spreadsheets.”

It was operational drag at the exact moment of closing. During a campaign window, a deal that isn’t collected within hours goes cold, and every closed deal triggered five manual workflows across five systems: work out the payment math, create a checkout link, send a contract, update the CRM, update the tracker. Roughly twenty minutes of manual handling per closed deal, across payment calculation, checkout, contract and tracking. Sixty-plus hours of manual work per week, company-wide, concentrated where speed mattered most.

THE OLD TRACKER, MEASUREDRECONSTRUCTED · SYNTHETIC DATA
CONTACT
PAYMENT DATE
STATUS
COLLECTED
RATE
104
14/03
$1,997.00
#DIV/0!
105
march 3rd
closed?
$280.08
107.85%
106
2025-03-14
#REF!
#DIV/0!
107
3.14.25
$8,991.00
62%
NOTES COLUMN, ROW 105
Base Amount: $1,997.00 · Scholarship: −$978.53 · Interest (10%): +$101.85 · Number of Payments: 4 · Monthly Payment: $280.08
91

columns wide on one dashboard sheet: nine months of reporting pasted side by side, rebuilt by hand every month

$3.79M

of booked revenue administered this way in 2025, with roughly half the cash floating as untracked receivables

15

different free-text date formats in a single payment-date column

52%

of lead rows had no status at all; 13 contacts entered twice; refunds existed only as prose in a notes field

Contract status tracked on 8 of 309 rows. Payment-plan and scholarship math worked out by hand, per deal, in a notes column. The sheet’s own error cells on permanent display: 15× #DIV/0!, 2× #REF!, and one collection rate reading 107.85%.
№ 04WHAT WE ADVISED AGAINST

The cheapest work we did was the work we refused.

NO PLATFORM MIGRATION

A previous attempt to move off Kajabi had already cost the company $60K+ and failed; 65,000 customers of engagement history live there. We built around it.

NO CRM REPLACEMENT

GoHighLevel stayed. The command centre is connective tissue, not another platform to migrate to.

NO NEW PROCESSOR

Stripe stayed; we automated its orchestration.

AND WE REMOVED THINGS

Mid-engagement, the data showed the two-role commission model was fiction. In 596 of 600 deals the opener and closer were the same person. We deleted it.

№ 05WHAT WE BUILT

The Sales Command Centre.

One application between the CRM, the payment processor and the course platform, running the complete sales lifecycle on a single source of truth.

DEAL CONFIGURATION
AND CHECKOUT

Pay-in-full or 2–9 instalments with automatic interest math, scholarships (including reverse calculation from a target monthly payment), deposits, setup fees, custom multi-phase schedules and a live agreement preview. Submit the deal and a checkout link exists while the customer is still on the call; expired links regenerate themselves nightly.

PAYMENTS AND CONTRACTS

Card, wire (auto-generated PDF invoice) and crypto available on every deal. Contracts are generated and status-tracked automatically, on every deal rather than on 8 rows in 309.

REVENUE AND RECEIVABLES

Real-time totals, collection rates, per-rep leaderboards and per-instalment receivables in one view, with exports for accounting. Cash that used to float untracked is visible by instalment.

COMMISSIONS AND GOVERNANCE

Product- and campaign-specific rates computed to the cent, self-serve PDF invoices for closers, and a 7-day cooling period enforced by the system rather than by memory.

REPORTING AND INVOICING

Campaign and revenue reports assemble themselves from the live database, and a 23-stage funnel is tracked across 21,000+ contacts and 87,000+ events. Wire and commission invoices generate on demand as finished PDFs. The monthly rebuild by hand is gone.

AI-ASSISTED
CONTENT DELIVERY

A recorded call becomes a published lesson without a person in the middle: the recording is transcribed, cut and formatted, then uploaded to the course platform.

WHERE THE AI WORK SITS

The client asked for AI and received it, in the places where it pays. Models transcribe and structure sales calls, turn a raw recording into a formatted lesson ready for upload, and read the transaction record to draft the reports and invoices that people used to assemble by hand. Everything a model produces lands in the same database as the deals, so the numbers on the dashboard and the numbers in the report cannot drift apart.

ARCHITECTURE
GOHIGHLEVEL
CRM
KAJABI
COURSES
STRIPE
PAYMENTS
ZOOM
CALLS
ORCHESTRATION LAYER
WORKFLOW ENGINE · EDGE FUNCTIONS · SCHEDULED JOBS
SINGLE SOURCE OF TRUTH
SALES COMMAND CENTRE
CLOSER · MANAGER · ADMIN VIEWS
THREE WEEKS FROM DIAGNOSTIC TO PRODUCTION
AUG 22

Diagnostic delivered

SEP 11

First commit

SEP 13

Production

OCT 11

First live deal

OCTOBER

First full campaign ($1.6M in-app)

№ 06RESULTS

Three tiers of evidence, kept apart on purpose.

MEASUREDVERIFIED AGAINST THE PRODUCTION DATABASE · DATA THROUGH JULY 2026
$7.85M booked through the platform in its first nine months: 2,367 deals, equivalent to approximately $10M in annualized platform throughput.
Three campaigns run end-to-end, each larger than any month in the spreadsheet era: October 2025 ($1.6M in-app, $2.55M campaign total, 3.6× the June pre-platform peak), March 2026 ($2.72M, 821 deals), June 2026 ($2.22M, 741 deals).
$2.24M booked in ten days (Feb 28 – Mar 9, 2026; 532 deals), with a $406K peak day.
91.5% of booked revenue collected or on schedule, the remainder visible per-instalment, against roughly half the cash floating untracked before.
49% of deals on flexible payment plans, 21% with scholarships, every one machine-calculated; $308K of commissions computed automatically across 25 closers.
$2.24M

IN TEN DAYS

$406K

PEAK DAY · MAR 3, 2026

CAMPAIGN MONTHS, BEFORE AND AFTER
JUN 2025 · PRE-PLATFORM PEAK$702K
OCT 2025$1.59M in-app · $2.55M campaign total
MAR 2026$2.72M
JUN 2026$2.22M
RECEIVABLES VISIBILITY, BEFORE AND AFTER
SPREADSHEET ERAroughly half the cash floating untracked
ON THE PLATFORM91.5% collected or on schedule

The remaining 8.5% is visible per instalment, with a named deal and a due date behind every dollar. The before figure is the audit’s own estimate of untracked receivables.

REPORTED BY THE CLIENT

The Head of Sales attributes 30% more closed deals per campaign to the removal of process friction: deals that previously cooled during manual handling now close inside the call.

NOT YET CLAIMED

We don’t publish an hours-saved figure. The audit measured 60+ manual hours per week before; the after-state was never formally re-measured, so we claim the eliminated workflows, not a synthetic number.

Campaign revenue depends on offer, list and team. We claim the removal of friction; the 30% figure is the Head of Sales’ own assessment, published with the client’s approval.

“I’ve never seen a better sales app. Deals that previously cooled during manual handling now close while the customer is still on the call.”

HEAD OF SALES
CLIENT’S SALES TEAM

Facing a similar operational question?

Book a 30-minute diagnostic conversation. If we are not the right fit, we will say so.

Request free 30-minute analysis

The client is not named in this write-up at their request. Client contacts are available as a reference on request.

Figures marked as measured are drawn directly from the platform’s production database (data through July 24, 2026) and the client’s original spreadsheet exports. Client-reported figures are attributed statements by the client’s sales leadership, published with approval.