The argument changed this year
Last year the case for a custom website was mostly about speed, SEO and looking different from your competitors. All still true. All still worth the money.
But something shifted in 2026 that makes the template question more urgent than it used to be, and most local business owners in Toronto have not caught up to it yet.
Your website now has two audiences. One of them is human. The other is a language model deciding whether to say your name.
Your customer stopped scrolling
Here is what a customer search looked like two years ago. Type “garage door repair Vaughan” into Google. Get ten results. Open three. Compare. Call one.
Here is what a growing number of those searches look like now. Ask ChatGPT or Perplexity “who should I call for garage door repair in Vaughan.” Get two or three names. Call one.
Notice what disappeared. The comparison step. The list. The chance to be the seventh result that someone scrolled to because your photos looked better.
In the old version, a mediocre website still got seen. It just had to be findable. In the new version, you are either named or you are invisible, and there is no page two to climb toward.
Why templates lose this specific fight
A language model cannot see your website. It does not admire your hero image or notice that your colour palette is tasteful. It reads structure, text and machine-readable data.
That is precisely what a template gives you least of.
Templates produce generic structure. A stock services page says “Quality Service You Can Trust.” It does not say what you do, where you do it, or what it costs. A model reading that page learns almost nothing it can repeat back to a customer.
Templates duplicate content across thousands of sites. When your page shares its bones with ten thousand other businesses, nothing about it is distinctive enough to cite.
Templates rarely carry proper schema markup. LocalBusiness, Service and FAQ schema are how you hand a machine the facts in a format it can use. Most themes either skip this or generate it badly.
Templates bury the specifics. Service areas, licensing, hours, response times, pricing — the details that answer a real question — end up in a footer or a PDF instead of in structured, readable content.
None of this shows up in a design review. It shows up in whether you get recommended.
The three costs that were already there
The 2026 argument does not replace the older one. It sits on top of it.
Speed. Template sites load slowly because they ship code for every feature you might use, not the ones you actually use. Page builders compound this. Visitors leave before your page finishes rendering, and Core Web Vitals feed directly into how Google ranks you.
Local relevance. Ranking for “electrician in Scarborough” needs genuine Scarborough content — neighbourhoods, permits, local conditions. A template with the city name swapped into a heading reads as a doorway page, and Google has been penalizing those for years.
Memorability. Your competitor picked the same template. Same layout, same stock handshake photo. Customers do not consciously notice. They just fail to remember either of you afterward. You can fix slow. You can fix SEO. There is no plugin that fixes forgettable.
When a template is genuinely fine
It would be dishonest to say never.
If you are validating a brand new business and need something online this week, a template is the correct call. Spend the money proving demand first.
If your business does not depend on search — you run entirely on referrals, repeat customers or a contract book — your website is a brochure, and a brochure can be a template.
If you have a fixed budget under roughly two thousand dollars, a well-chosen template built carefully beats a cheap custom build done badly. A bad custom site is worse than a good template.
The line sits here: when search is how customers find you, the website stops being a brochure and becomes a lead channel. Lead channels earn investment. Brochures do not.
What a custom build actually gives you in 2026
A serious build for a local service business should include all of the following. If a quote is missing several, ask why.
- Clean semantic structure that both Google and language models can parse
- LocalBusiness, Service and FAQ schema implemented properly, not by a plugin guess
- Genuine location pages with unique local content, not a city name find-and-replace
- Content written to answer the questions customers actually type
- Consistent name, address and phone across every platform you appear on
- Core Web Vitals handled at build time rather than patched with a caching plugin
- Trust signals — licensing, insurance, warranty, response times — placed where they are visible before a visitor scrolls
- A mobile experience designed for someone standing outside with a problem
The part that is hard to retrofit
Speed can be fixed later. Content can be rewritten later. Schema can be added.
Structure is the expensive one. If your site was built on a foundation that buries your specifics inside generic components, making it legible to AI search usually means rebuilding rather than patching.
That is the real argument for doing it properly the first time, and it is the one that most quotes never mention.
Where to start
If you are not sure which side of the line your business sits on, the fastest answer is to ask an AI assistant what it says about your industry in your city, and see whether your name comes up.
If it does not, that is your answer.
We run free manual GEO audits for businesses across the Greater Toronto Area — a real review of how your site performs in AI-generated results and what specifically to fix. No automated report, no sales pressure. Get a free GEO audit →
Frequently Asked Questions
When customers find you through search, yes. A custom build controls the structure, speed and machine-readable data that determine whether you rank and whether AI tools recommend you. If your customers come from referrals rather than search, a template is usually sufficient.
It can, particularly in less competitive markets. The difficulty is that templates produce generic structure and often lack proper schema, which limits both traditional rankings and visibility in AI-generated answers.
Pricing depends on page count, integrations and how much location coverage you need. The meaningful comparison is not template versus custom – it is the build cost against the value of the leads the site produces over three to five years.
GEO – Generative Engine Optimization – is structuring your site so AI tools name your business when answering customer questions. It matters most for businesses where customers search before they call: trades, legal, medical, consulting and local services.
Related reading: What an AI-driven WordPress website actually means · Why page builders slow you down · WordPress for business: pros and cons · Website design and redesign services