Digital marketing for service businesses

The platform reported 59,270 conversions — form fills, calls, clicks — and roughly 5% of them ever became a shipment. I imported a hard offline conversion goal, “Order Picked Up,” straight from the client's ops data, ran each vehicle category as its own campaign with its own budget, and kept lead-volume bidding separate from the offline goal. Google learns at scale, the business reports on the real number.

Cost per lead sat flat at $10–12 whether the account spent $3K or $6K a month — the comfortable number to report. The number that pays the bills is cost per verified order. I optimised keyword by keyword against the client's own lead statuses in LeadByte instead of platform conversion counts, and cost per order fell roughly 4x while spend scaled.

The account arrived with high CPC, irrelevant traffic, low conversion volume and tracking nobody trusted. Full audit, then Search rebuilt from the ground up: intent-based ad groups, corrected geo segmentation, negative keyword lists, call tracking and UTM templates so every lead traces back to its source. One kitchen job covers the month's budget several times over.

Came in with policy disapprovals, limited delivery and spam calls eating the budget — a category most agencies avoid because of AV friction. I rebuilt Search into intent-based ad groups, cleaned search terms and negatives, rebuilt PMax assets, fixed call quality through CallRail, prepared the Advanced Verification documentation and appeals, and worked directly with Google support until delivery stabilised.

The problem wasn't spend, it was measurement. Google was optimising toward map clicks and direction requests; the real booking happened off-site in NexHealth and the algorithm never saw it. We built a dedicated implant landing page for paid traffic, wired tracking into the booking system so every call and form tied back to its ad, and reset bidding to optimise for booked consultations instead of soft signals.

Before me there was a website and nothing else — no ads, no tracking. Verified Google Ads account, Search + PMax + remarketing split by real buyer intent, negative lists so nothing leaked to DIY searchers and students, GTM and GA4 wired properly: forms, calls, WhatsApp, UTM on every source. First click in December, first signed contract in May. Five months that look dead on a dashboard — most accounts get killed in that window. This one survived because the tracking was solid enough to trust the pipeline.

Two languages with different search behaviour, an emissions niche where Google's policy gets touchy on wording, and a low-ticket walk-in model that had to run efficient from day one. Search split by service line and by language — RU and LV in separate ad groups to protect Quality Score — a branded and geo layer underneath, PMax on top, then post-launch cleanup stripping junk conversions so Smart Bidding learned from real leads.

Conversion tracking built from zero: Ringostat answered calls, missed calls and callback requests piped through GA4 as the primary signal alongside form fills. Search split by repair category and by marque, running next to PMax. Two weeks in the data showed PMax converting cheaper than general Search, so budget moved that week — PMax now drives about 69% of all conversions.

No history, no data, and eight campaign-platform combinations guessing at once. Before a single ad ran I built the full GTM and GA4 foundation so every lead and order traces back to its campaign and keyword. Bing launched on day one as a real second channel rather than a bolt-on test. The account closed its first month in the black.

On Meta, a core Advantage+ Shopping campaign carries the budget weight with a constant rotation of small-budget creative and audience tests underneath, finding winning angles before anything scales. On Google, branded Search defends the name against competitors while the growth budget runs through Performance Max split by product line and region, so each category bids on its own terms rather than being averaged together.

One blanket PMax campaign makes Google average its bidding across a $7 pillowcase and a $290 mattress, and the product that actually moves the business gets under-served. I built six dedicated campaigns, one per category, each with its own budget and bidding target, so the algorithm optimises mattresses like mattresses and accessories like accessories.