What a website should cost, anchored to real numbers.

Nobody can give you a real number for what your website should cost without knowing what your website is supposed to do. The questions that decide the answer are not technical, they are commercial. Does your business actually run through the website. Are your buyers researching for weeks before they contact you, or deciding in thirty seconds. How many years should this site last before the next rebuild. What integrations does the site need to do its job. None of those are pricing questions. All of them are pricing inputs.

The other reason most pricing guides are useless is that they publish national averages stripped of buyer context. The mean cost of a small-business website rolled across DIY, freelancer, studio, and agency tiers is a number that describes nobody's actual decision. The right framing is by tier, with each tier's real range and a clear read on which buyer profile fits inside it. That is what this page does.

The honest baseline is that a service business website in 2026 should cost somewhere between zero and eighty thousand dollars to build, depending on the tier, and somewhere between two hundred and six thousand dollars per year to run. The five-figure span at the lower end and the six-figure span at the upper end exist because the work is genuinely different at each tier, not because anyone is inflating prices.

The market sorts into four tiers that each have a stable range. The boundaries are softer than the table below implies; a senior freelancer can deliver studio-tier work and a small studio can be priced like a freelancer. The framing still holds because the cost structure of each model is real and the failure modes of each model are predictable.

DIY templatedcovers Squarespace, Wix, WordPress.com, and the like. The setup is a weekend. The recurring platform fee runs two hundred to five hundred dollars per year for the business plan. The ranking ceiling is whatever the platform's template performance allows, which is usually the lower bound of what Google rewards. For a brochure with hours, an address, and a phone number, this tier is a defensible choice and I will tell you so.

Freelancer custom covers a single contractor or a two-person shop. Cost runs two thousand to fifteen thousand dollars for the build, depending on scope. Code ownership transfers if you contract for it. Time to launch runs two to eight weeks for tactical work. The predictable failure mode is post-launch: a freelancer is one person and people get sick, take other contracts, or stop returning emails. Studio engagements are usually stronger here because there is a published support arrangement.

Studio covers principal-architect studios like this one. Build cost runs four thousand five hundred to twenty-five thousand dollars and up, depending on the engagement tier. The ranking and performance ceiling are high because the build target is real engineering, not a template. Time to launch runs eight to fourteen weeks. The best fit is service businesses that run through the website and want the site to last five years. The wrong fit is a one-page brochure where the studio premium is overpaying for engineering you do not need.

Agency covers multi-headcount shops with coordinators, project managers, separate engineering and design teams, and procurement-friendly contracts. Build cost runs fifteen thousand to eighty thousand dollars and up, with enterprise engagements past one hundred thousand dollars. Time to launch runs twelve to twenty-four weeks and longer. The cost structure is real (real headcount, real overhead). The failure mode is paying agency rates for studio work, which I cover in the agency vs studio vs freelancer page.

Six factors push a website price up, in order of impact. Custom integrations beyond a contact form are first. A marketing site with a contact form is one cost. A site with a booking flow, a customer login, a payment processor, and a CRM sync is a different cost, because each integration is its own engineering surface to design, build, test, and maintain. Adding three integrations to a base scope can double the engagement.

Content depth is second. A ten-page site is roughly a third of the build effort of a forty-page site. Page count is not just a content-modeling problem; each page is a performance, accessibility, and SEO surface to engineer. Sites that need fifty location pages or one hundred service pages cost more, and the cost is mostly engineering, not copywriting.

Design specificity is third. A site that uses a clean design system applied across templates is one cost. A site where every page has bespoke art direction is a different cost, because each page is its own design and engineering pass. Most service businesses do not need bespoke art direction on every page; the ones that do (luxury, high-end professional services, hospitality) are paying the premium for a real reason.

Performance and accessibility floor is fourth. A site targeting a Lighthouse score in the high seventies is one cost. A site targeting a high-nineties score across mobile and desktop, in the field, at the seventy-fifth percentile, is a different cost. The performance work is real engineering, and it compounds with every additional page and every additional integration. WCAG 2.2 AA conformance is similarly real work, not a checkbox (W3C, 2023).

Content management ergonomics is fifth. A site where the owner edits content directly in code is one cost. A site where a non-technical content team needs to ship updates weekly through a CMS with previews, scheduling, and version history is a different cost. Headless CMS integrations (Sanity, Payload, Contentful) add real engineering scope and ongoing license fees.

Internal-tool surface is sixth and largest at the top. If the engagement is not just a marketing site but also a custom admin dashboard, a CRM, an ops console, or a customer portal, the cost spans into the high five and low six figures. That is a different category of work, and the right pricing reference is a custom-software engagement, not a website engagement.

Three things are worth paying for, almost always. Performance engineering, because the revenue impact is measurable and recurs forever. Accessibility, because the work shares root causes with performance and SEO and the legal exposure is real for any business serving the public. Code ownership, because what you cannot take with you is a recurring cost in disguise.

Three things are usually overpriced. The first is custom motion design beyond what the brand actually requires. Animations are inexpensive to add, expensive to do well, and easy to overspend on without measurable benefit. Second is bespoke art direction on every page, which is a real cost for the rare buyer who needs it and a wasted cost for the buyers who do not. Third is design rounds past the third or fourth iteration. Past that point you are usually buying decision avoidance, not better design.

Two things are commonly mispriced in both directions. SEO retainers at five hundred to one thousand dollars a month for a small service business rarely deliver proportional value at that revenue scale; a one-time foundation engagement plus maintenance is the right shape. CMS licenses are sometimes overpriced for small content teams (Contentful Enterprise is overkill for most service businesses) and sometimes underpriced (Sanity and Payload can deliver enterprise-grade ergonomics for under a thousand dollars per year).

The most useful pricing reference in any category is the one that names a real number. I publish four engagement tiers on this site so the math is auditable rather than obscured.

Fix Sprint is two thousand nine hundred ninety-five dollars, two-week fixed-price, post-Pathlight engagement. The three highest-impact issues from your scan, shipped to production, with a Lighthouse before-and-after and a fresh Pathlight re-scan. The engagement model lives at /pricing/fix-sprint; the methodology is documented at the performance audit page.

Starter at four thousand five hundred dollars covers a focused new-build for a service business with straightforward scope. Professional at nine thousand five hundred dollars covers a more substantial build with custom integrations. Enterprise starting at fifteen thousand dollars covers larger engagements where the scope is a per-engagement quote. Custom Next.js development for buyers who need a full rebuild is documented on the Next.js development page.

The reason I publish these prices is that opaque pricing is one of the most common failure modes in this category, and it costs everybody time. A buyer who knows my range up front can self-qualify in five minutes; a buyer who has to schedule a discovery call to learn whether I am even in their budget is being asked to pay for the privilege of finding out. The honest framing is the differentiator. The published number is the proof.