Australia's oldest jeweller was spending $17,000+ to land one $70 sale. Here is how we found it, fixed it, and rebuilt the measurement underneath it.

Fairfax & Roberts is Australia's oldest jewellery house, trading in Sydney since 1858. It is not an ecommerce business. The average sale is worth $30,000+, and the real sale happens in a showroom consultation, not a checkout page. Every problem we found, and everything we fixed, comes back to one fact: the website and the ad account had been built and measured like a $50 impulse-buy store, not a $30,000 bespoke jeweller.

In the quarter before we took over, the account had spent $17,000+ and driven 39,000+ paid sessions to the site. The ecommerce data for that same period shows one online transaction. That is not a rounding error. It is what happens when every campaign is optimised for landing page views instead of appointments. Some ad sets had been shown to the same people ten times over, well past the point of any real return, and every ad in the account used the same "Shop Now" call to action. There was no appointment-booking event, no phone-click event, no lead event anywhere in the tracking. The business had no way of knowing what its marketing was actually doing.

The site had its own crisis running in parallel. A behavioural audit across 1,500+ unique URLs scored the site 4 out of 10 on conversion readiness. JavaScript errors were breaking nearly one in five sessions site-wide, and far worse on mobile and Instagram traffic specifically. Half the people who landed on the homepage left immediately. None of this was improving on its own, and some of it was getting worse.

Underneath all of it, the CRM told the same story. 200+ forms live, most of them unused, 20+ deals worth $230,000+ sitting in a pipeline stage nobody could account for, and a group of the business's highest-value contacts blocked from receiving any marketing at all.

We rebuilt the paid media plan around what the business actually sells: booked consultations, not clicks. We paused the campaigns that had been burning the same audience over and over, retired messaging that was attracting bargain hunters instead of the real buyer, and replaced every "Shop Now" button with a consultation-led call to action.

We found the account's best asset sitting untouched: a Signet Ring video with 140,000+ completed views and nothing built to capture a single one of them. That became a dedicated page and its own lead campaign.

On the site, we shipped five releases fixing the diamond search tool, rebuilding the main navigation, adding a booking tab and mobile sticky call to action, and shipping a proper homepage video and testimonials. We cleared 3,000+ orphaned pages, fixed 90+ missing H1 tags and 40+ broken internal links, and closed a live credibility gap where two different pages quoted two different founding dates for the same 168-year-old business.

We rebuilt the CRM from the ground up too: consolidating the form sprawl, identifying the $230,000+ stuck in limbo, and building a win-back sequence to recover the blocked contact list.

Underneath everything, we put in place the measurement that didn't exist on day one: a five-metric reporting framework, fortnightly behavioural reviews, and full event tracking across every stage of the consultation funnel.

This is an active engagement, not a finished story, and we're publishing it as one. Scroll depth has moved from a 31%+ baseline to averaging 50%+. The foundation, the part that actually matters for a business like this, is in place for the first time. The revenue-side proof (consultations booked, cost per qualified lead, recovered email revenue) is still building, and that's the next chapter, not this one.