Google Ads fails quietly. The account spends every day, the dashboard shows conversions, and the phone still does not ring more often.
The first leak is intent. Broad match keywords without a maintained negative list pull in searches that were never about the service: tutorials, job seekers, competitor brands, the wrong city. Each irrelevant click is paid for at the same price as a genuine enquiry.
The second leak is measurement. Page views, button clicks and time on site are set up as conversions because they are easy to configure. The bidding algorithm then optimises toward those signals and away from the calls and forms that actually become jobs. A campaign can report a healthy conversion rate while producing nothing a business owner would recognise as a customer.
The third leak is Performance Max run as a single black box. One campaign, every product or service in it, brand searches included, so the reported return is inflated by people who were already looking for the business by name. For a repair shop or a trade business, this is the difference between paid search that pays for itself and paid search that quietly subsidises Google.
The account is rebuilt from the conversion backwards, so every bid is placed against a real enquiry.
Calls, forms and chats are tracked as primary conversions with call duration thresholds. Where sales close later, offline conversion import connects the closed job back to the click that started it.
Search campaigns are split by service line so budget follows the work that is most profitable. Exact and phrase match carry the core terms; broad match is used only where a negative library already exists to protect it.
Every search term that spent money is reviewed each week. Irrelevant terms join a shared negative keyword library that grows for the life of the account, so the same waste is never paid for twice.
Asset groups are split by service, brand terms are excluded, and audience signals are seeded from converters. Reported return is then honest, and budget between Search and Performance Max is reallocated on cost per enquiry.
Phone Mechanic Parramatta shows this structure working beside Meta: Google Ads captured customers already searching for a repair while Meta created new demand, with both engines reported against the same counter. Remotie used the same discipline inside an agency's backend, managed remotely by Iftaykhar Mahmud.
Paid search return is measured in enquiries and revenue per dollar of spend, never in clicks.
Cost per enquiry is total spend divided by verified calls, forms and chats. Cost per sale is the same calculation on closed jobs, available once offline conversion import is in place. Revenue per dollar of ad spend is attributed revenue divided by spend, and the break-even point is one divided by gross margin on the service being advertised.
Because search captures existing demand, its return usually shows quickly and stabilises once automated bidding has enough conversion data. The financial question then becomes marginal: does the next dollar of budget still buy an enquiry below the target cost? Weekly reallocation answers that question with data instead of instinct.
Wasted spend removed is reported as its own line, because every irrelevant click cut is margin returned to the business. Client figures are shared on request, not published.
| Measure | How it is calculated | Reported |
|---|---|---|
| Cost per enquiry | Spend divided by verified calls, forms and chats | Weekly |
| Cost per sale | Spend divided by closed jobs from offline import | Monthly |
| Revenue per dollar of spend | Attributed revenue divided by spend | Weekly |
| Wasted spend removed | Cost of search terms added to the negative library | Weekly |
| Impression share on converting terms | Impressions won divided by eligible impressions, converting terms only | Monthly |
Google Ads is for businesses whose customers already search before they buy: repair stores, local services, agencies and stores with clear product intent. The work is capturing that intent at a cost the margin can carry, then proving it with conversion data rather than click reports.
The numbers reported every week, agreed in writing before anything is built.
Broad match keywords with no negative list. The fix is a weekly search-term audit and a maintained negative keyword library.
Page views or button clicks are tracked as conversions. The fix is to count calls, forms and sales only, with offline import where deals close later.
One campaign with every product in it. The fix is asset groups split by service line, with brand searches excluded so results are honest.
Phone Mechanic Parramatta is a mobile phone repair store in Westfield Parramatta, New South Wales, Australia. Iftaykhar Mahmud paired high-intent Google Ads with Meta demand generation so the store captured customers already searching for repairs and created new ones who were not: two acquisition engines feeding one counter. Delivered remotely from Dhaka, Bangladesh.
Remotie is a growth marketing agency headquartered in Melbourne, Australia, with delivery hubs in Dhaka and Manila. Iftaykhar Mahmud ran its full backend growth engine: email marketing, social media marketing, Google Ads, AI task automations and business development. The agency scaled client delivery without adding headcount, with the work delivered remotely from Dhaka, Bangladesh.
Campaign structure for Search, Performance Max and remarketing, keyword and negative keyword management, conversion tracking setup, bid strategy management and weekly reporting on cost per enquiry and return on ad spend.
Phone Mechanic Parramatta, a phone repair store in Parramatta, New South Wales, Australia, and Remotie, a growth marketing agency headquartered in Melbourne, Australia.
By cost per enquiry (calls and forms), cost per sale where sales data is available, and revenue divided by ad spend. Reporting uses verified conversion actions in Google Ads and Google Analytics 4, not click counts.
Search campaigns can generate enquiries in the first week because they capture people already searching. Automated bidding then needs a learning period, usually one to two weeks after a significant change, and enough conversion data before results stabilise. Performance is judged over 30 to 90 days, with search-term hygiene and budget reallocation happening every week in between.
Thirty minutes, no pitch deck. Bring your numbers and you will leave knowing whether this is the lever that moves them.
Free 30-minute discovery call. No pitch decks, just straight answers.
Start the conversation
Send the brief below or grab a 30-minute discovery slot. You will get a straight answer on whether I can move your numbers, and exactly how.
Every field helps me answer properly the first time.