Web Development / Local SEO / Google Ads
Industry: Home Services
$4.07 AUD
Average CPC (Sydney market, 30 days)
In short
Arcetis designed, built, and operates a network of city- and service-specific microsites for a multi-market pest-control business in Sydney, Brisbane, and a third overseas market, with most sites sharing one templatized PHP/MySQL codebase and a single CRM integration, plus a dedicated Laravel 11 application for a regulated wildlife-removal service. The scope of work includes recurring Lighthouse and security/SEO audits, a 45-post and 29-page programmatic local SEO rebuild, and live Google Ads management that, in the Sydney market over a 30-day window, produced 9 conversions at a $157.31 AUD cost per conversion from $1,415.75 AUD in spend. Brisbane and a third overseas market have equivalent technical builds and campaign setup underway, with ad performance still being measured in those markets.
This client is a pest-control services business operating across multiple international markets, including Sydney and Brisbane. Rather than running a single national website, the business competes market by market, where localized service pages and local search visibility routinely outweigh generic, single-domain SEO. Arcetis was engaged to build and continuously operate the technical and marketing infrastructure behind that strategy: a network of four to five city- and service-specific microsites, a separate regulated wildlife-removal application, a shared CRM integration unifying leads across every site's booking forms, and the ongoing technical SEO, content, and Google Ads work required to keep each market's search presence competitive. The engagement spans initial builds, multiple audit and remediation rounds, a large programmatic content rebuild, and live paid-search management, with the Sydney market currently the most mature in terms of measured advertising performance.
Local SEO for a home-services business is inherently market by market: a pest-control search in one Sydney suburb has almost nothing in common with one in Brisbane's northern suburbs or the third overseas market, yet each market needs its own locally relevant pages, its own Google Business presence, and its own path from search click to booked job. Building four or five fully bespoke websites and an equal number of independent ad programs from scratch was not a realistic budget or timeline for the business, but a single generic national site would not rank for tightly local searches either. The business also had one service line, regulated wildlife removal covering possum and rodent control, that could not simply be folded into a generic pest-control template. This service carries its own compliance and messaging requirements in the Australian market, so it needed dedicated architecture rather than another cosmetic reskin of the shared site. On top of the split between templatized microsites and a bespoke application, every site needed to feed one downstream sales process. Leads from a Sydney booking form, a Brisbane booking form, and the wildlife-removal application all needed to land in the same CRM so the business could manage jobs, technicians, and follow-up from a single view, rather than staff checking multiple separate inboxes or admin panels across five sites. Finally, because the network had grown site by site over time, it carried real, ordinary technical debt: sites at different stages of maturity, inconsistent accessibility and performance baselines, legal and compliance gaps on some pages, and a Content-Security-Policy configuration that had not kept pace with the marketing tags, Google Tag Manager, Ads, and Analytics, actually running on the sites. None of this was catastrophic, but it needed to be found, prioritized, and fixed systematically through documented audits, rather than patched ad hoc, before the network could be trusted to carry meaningful Google Ads spend across multiple countries.
Arcetis is the sole technical and marketing partner responsible for this network end to end. Arcetis designed and built the shared PHP/MySQL microsite template, including its custom admin CMS, .htaccess routing, and sitemap generation, then forked and rebranded that codebase per city and service line rather than commissioning separate bespoke builds. Arcetis separately designed and built the Laravel 11 application for the regulated wildlife-removal service, integrating Blade, Vite-compiled Tailwind, Alpine.js, Resend, reCAPTCHA v3, and its own CRM and conversion-tracking connections. Arcetis conducted the production-readiness audit, the Lighthouse accessibility and performance pass, and the subsequent live audit round across the network, and Arcetis plans, builds, and manages the Google Ads campaigns, including the Sydney account with the performance data cited here. Brisbane and a third overseas market technical builds, campaign setup, and market research were also delivered by Arcetis, with paid-search performance measurement in progress in those markets.
Arcetis designed one PHP/MySQL codebase, with a custom-built admin CMS (authentication, blog editor, service editor, image uploads), custom .htaccess routing, hand-rolled XML sitemap generation, and schema.org JSON-LD markup, then forked and rebranded it per city and service line to produce a live microsite network across the Sydney, Brisbane, and a third overseas market markets, confirmed via shared git commit history across the sites. This let the business launch and maintain multiple locally relevant sites without the cost of separate from-scratch builds for each market.
One service line, possum, rodent, and other wildlife removal, was built separately as its own Laravel 11 application rather than forced into the shared template, using Blade templates, Tailwind compiled via Vite, Alpine.js for interactivity, Resend for transactional email, Google reCAPTCHA v3 for form protection, and its own CRM integration and Google Ads conversion tracking, reflecting the distinct compliance and messaging needs of a regulated service.
A "Premium rebuild" of the primary site added 45 blog posts and 29 suburb and location landing pages, along with comparison ("best-of") pages and centralized online booking. The 29 suburb pages render through a single dynamic, database-driven template rather than as 29 hand-built pages, letting the business scale locally relevant landing pages across an entire metro area from one maintainable codebase.
Arcetis ran a documented production-readiness audit covering security, SEO, accessibility, and performance, followed by a dedicated Lighthouse pass that fixed color-contrast issues (bringing accessibility close to a perfect score), improved back/forward-cache eligibility, added explicit font-display and image dimensions, and converted the logo to WebP. A later live audit round fixed legal pages, forced HTTP/2, rebuilt the Content-Security-Policy to properly allow Google Tag Manager, Ads, and Analytics, added an accessibility skip-link, and rebuilt the sitemap.
Every site in the network, including the standalone Laravel wildlife-removal application, pushes booking-form leads into the same third-party CRM integration, so leads from every city and service line land in one place for job management, dispatch, and follow-up, regardless of which microsite or domain generated the inquiry.
Arcetis manages live Google Ads campaigns market by market, with the Sydney account the most performance-mature to date, and has begun shipping llms.txt and llms-full.txt files on the two most current site builds, an emerging standard for structuring site content for AI answer engines such as ChatGPT, Perplexity, and Google AI Overviews, positioning the network for both traditional and AI-mediated search discovery.
The network's technical foundation is a single PHP and MySQL codebase, written without a framework, that Arcetis built once and then forked per market. Rather than maintaining several independent codebases with separate bugs, patches, and update cycles, the business runs multiple microsites that share the same admin CMS, covering authentication, a blog editor, a service-page editor, and image uploads, the same custom .htaccess routing rules, the same hand-rolled XML sitemap generator, and the same schema.org JSON-LD markup for structured data. This is confirmed at the code level: the sites share git commit history, meaning each city or niche variant is a genuine fork-and-rebrand of one original build rather than a from-scratch project. That is a deliberate trade-off Arcetis made and stands behind: it is faster to ship, cheaper to maintain, and easier to roll a fix or a new feature out across the whole network at once, at the cost of each individual site being somewhat less bespoke than a one-off build would be. The one clear exception is the wildlife-removal service line, which sits on its own Laravel 11 application rather than the shared template. Because regulated wildlife removal, possum and rodent control specifically, carries different compliance and messaging requirements than general pest control, Arcetis judged that reskinning the shared PHP template was not the right fit and built a dedicated application instead: Blade templates, Tailwind compiled through Vite, Alpine.js for lightweight interactivity, Resend for transactional email, Google reCAPTCHA v3 on forms, and its own CRM integration and Ads conversion tracking. Despite the split between the shared template and the standalone Laravel app, every site funnels leads into the same place. Each site's booking form, whether it runs on the shared PHP template or the standalone wildlife-removal application, integrates with a common third-party CRM. That means a Sydney lead, a Brisbane lead, and a wildlife-removal lead all arrive in one system, letting the business run job scheduling, technician dispatch, and follow-up from a single operational view instead of monitoring several separate inboxes or admin panels. The most concrete example of the network's local SEO approach is the "Premium rebuild" carried out on the primary site. Rather than hand-building dozens of individual location pages, Arcetis added 45 blog posts and 29 suburb and location landing pages, all rendered through one dynamic, database-driven template. Each suburb page pulls from the same underlying data model and template logic, varying only the location-specific content, which is what makes dozens of locally relevant landing pages achievable and maintainable rather than a permanent hand-coding burden every time the business wants to add a new suburb. The rebuild also added comparison ("best-of") pages, positioning the primary site to compete for higher-intent, comparison-style searches, and centralized online booking so that every acquisition channel, organic, paid, and referral, converges on one booking flow feeding the shared CRM. Two of the network's most recent builds go a step further and ship llms.txt and llms-full.txt files alongside the standard sitemap and robots.txt, a nascent but increasingly relevant standard for making site content legible to AI crawlers and answer engines, rather than relying solely on traditional search-engine indexing.
A dedicated Lighthouse audit fixed color-contrast failures across the network, bringing accessibility scores close to a perfect 100, improved back/forward-cache (bfcache) eligibility, added explicit font-display and image-dimension attributes to reduce layout shift, and converted the primary logo asset to WebP. A later live audit round added further fixes: legal-page corrections, forced HTTP/2, a rebuilt Content-Security-Policy properly allowing Google Tag Manager, Ads, and Analytics, a new accessibility skip-link, and a rebuilt XML sitemap.
The network's two most recently built sites ship llms.txt and llms-full.txt files, an emerging convention that structures a site's content specifically for large-language-model crawlers, ahead of most home-services competitors. As tools such as ChatGPT, Perplexity, and Google AI Overviews increasingly answer local-service queries directly, this positions those sites to be legible to AI-mediated discovery, not only traditional search indexing.
Not every site in the network is held to the same build standard, and Arcetis is candid about that rather than glossing over it: one of the newer sibling microsites currently loads Tailwind CSS via a CDN script rather than a properly compiled build step, unlike the Laravel application's Vite-compiled Tailwind. It is a real, minor piece of technical debt, flagged for remediation, and a useful marker of engineering self-awareness rather than a hidden flaw.
Rather than presenting the microsites as independent builds, Arcetis is transparent that most of the network is one original codebase, forked and rebranded per city or niche, a fact independently verifiable in each site's shared git commit history. The trade-off is intentional: centralized maintenance and faster rollout, in exchange for less per-site bespoke engineering than separate ground-up builds would provide.
Because the shared microsite template is vanilla PHP and MySQL with no framework, Arcetis built its own admin CMS from scratch, covering authentication, a blog editor, a service-page editor, and image uploads, alongside custom .htaccess routing, a hand-rolled XML sitemap generator, and schema.org JSON-LD structured-data markup on every page, without relying on an off-the-shelf CMS.
Average CPC (Sydney market, 30 days)
Conversions at $157.31 AUD cost per conversion
Technology used
Key learnings
Templatizing most of the network, one PHP/MySQL codebase forked and rebranded per city and service line, was the right call for this business, and Arcetis is candid about the trade-off rather than presenting the network as fully bespoke builds. Shared git commit history across the sites confirms the approach: it buys faster shipping, one place to patch security or SEO issues, and lower long-term maintenance cost, at the price of each individual site being less custom-engineered than a one-off build. The one place Arcetis chose not to templatize, the regulated wildlife-removal service, proves the model has limits: some service lines carry compliance and messaging requirements specific enough to warrant a dedicated Laravel application instead of a reskin. The audits surfaced a related lesson: a network built market by market over time accumulates uneven engineering rigor, and one sibling site still loads Tailwind via a CDN script rather than a compiled build, a known, flagged gap rather than a hidden one. Geographically, results are real but not yet uniform. The Sydney Google Ads account has a full 30-day performance record and a documented trend from zero conversions to a $121 cost per conversion in the most recent week measured. Brisbane and a third overseas market have real technical builds, campaign setup, and market research underway, but paid-search performance in those markets is still being measured, proof that a network can scale technically well ahead of when any one market's ad performance is provable.
Not necessarily a fully bespoke one. This network shows a middle path: one well-built template, forked and rebranded per city or service line, can deliver locally relevant local SEO pages and Google Ads landing experiences across multiple markets without the cost of several independent builds, while still reserving a fully custom build for the one service line, regulated wildlife removal, that genuinely needed different architecture.
Programmatic local SEO means generating many locally targeted pages from one dynamic, database-driven template rather than hand-coding each one. In this case, a "Premium rebuild" of the primary site added 45 blog posts and 29 suburb and location landing pages through a single template, plus comparison pages and centralized booking.
Only the Sydney market currently has a measured 30-day performance window: 12,255 impressions, 350 clicks, an average CPC of $4.07 AUD, $1,415.75 AUD in total spend, 9 conversions, a 2.59% conversion rate, and a $157.31 AUD cost per conversion, with a 13% impression share, ranked 5th among visible competitors. Brisbane and a third overseas market have live technical builds and campaign setup but no comparable ad-performance results yet.
llms.txt, and its longer counterpart llms-full.txt, is an emerging convention for structuring a website's content specifically for AI crawlers and answer engines like ChatGPT, Perplexity, and Google AI Overviews, similar in spirit to how sitemap.xml serves traditional search engines. The two most current sites in this network ship both files, positioning them to be legible to AI-mediated search discovery in addition to conventional organic search.