Home Services · Owner-led · Since 2025
Golden Rule
Privately held. Several trades. More than two decades of growth.
Where the drag was.
How to read the model
The parts in colour are where the drag was. Point at or tap a line to light that part.
Show the whole model
How Growth by Design was applied · Step 1 of 5 · The goal
What the owner wanted
Golden Rule is privately held, with several trades, and has grown for more than two decades on written processes and training.
The owner wanted more of the week for the work only an owner can do, and numbers they could trust, starting with which ad spend turned into paid jobs.
The whole engine: a first call on the left, a paid job and a returning customer on the right.
Step 2 of 5 · What we found
From call to paid job
We mapped the business from first call to paid job to find where growth stalled.
Then we traced ads to bookings and jobs, through the ad tracking, each campaign's phone number and the call records.
The base: the ad, phone and job systems that hold the records.
Step 3 of 5 · What was in the way
Three systems, three answers
The ad, phone and job records disagreed on what counted as a bookable call, a booking or an abandoned call, and spam was mixed in with real calls.
So Google learned from the wrong booking signal, and the team couldn't tell which ad spend worked. Fixes came one tool at a time, without the data to choose among them.
The brass gate and the lanes: the booking signal sent back to the ads was wrong.
Step 4 of 5 · What we did
Measure first, then fix
Diraj asked for measurement first, then the biggest fixes. With the team, we agreed one set of definitions: spam, a bookable call, an attempt, a booked job.
We audited ad tracking, repaired phone routing and technician availability issues, built service landing pages, and connected ad, call and job records so the team could see where leads dropped out. We compared booking under AI and human handling, and set a target to close the gap, with one person responsible.
We built AI skills and data connectors the team runs from its own accounts. They read records, flag exceptions and draft reports, and a person checks each result.
Diraj advised moving systems into the company's own accounts, naming owners for spend and live changes, and letting AI earn responsibility in stages, with people keeping the final say. The company's leaders decided what to act on.
The base and the brass gate: one set of definitions, connected records and named owners.
Step 5 of 5 · What they learned
Measurement comes before spend
The order mattered. Until the records agreed, no new ad dollar or tool could be judged, so demand work waited for sound measurement.
Fixing one step also tests the next: more booked calls put weight on the people who serve them.
The lanes and the base: ad spend can only be judged once the records underneath agree.
What a CEO can take from this.
Agree what a booking is
If your ad, phone and job systems count bookings differently, your ads learn from the wrong signal.
Measure before you optimise
Get records that agree before you choose which fix to make or where to spend.
Let AI earn responsibility
Give AI more to do in stages, and keep the final say on money and live changes with people.
If we started today
We would have an AI agent compare every call with job records from the first week, so the team could agree its definitions and fix each break before any new ad spend went live.
Growth by Design
This is Growth by Design at work. See the method behind it
Where to go next
Read about this kind of work
Other stories
- Read the ablemkr storyConstruction & energy staffing platform
- Read the TravelAI storyAI-Powered Travel Platform
Do your ad, phone and job records agree on which calls became booked jobs?
30 minutes. We'll tell you honestly if we're the right fit.

