Website ROI in 2026: How to Calculate It — and Why the Math Just Changed
The formula fits in one line. The interesting part is that both of its inputs — what a website costs and what measuring it costs — collapsed in the last two years.
Most businesses can't answer a simple question: does your website make money? Not “does it get traffic” — does the money it brings in exceed the money it costs? The usual reason isn't laziness. Until recently, answering properly required either an analyst or a level of tooling small businesses couldn't justify.
That excuse expired. I measure both of my businesses' sites programmatically — data pulled through APIs into dashboards AI tooling built for me — and the same approach is available to anyone with an evening of curiosity. This guide covers the formula (short), the attribution problem (the real work), and the part most 2026 guides haven't caught up with: the ROI math itself changed, in your favor.
How do you calculate website ROI?
The formula is one line:
A site that generated $50,000 in attributable revenue against $15,000 in total costs returned 233% — every dollar in brought $2.33 back. Below 0% is a money-losing site; 100–400% is healthy; above 400% is excellent.
Everything difficult about ROI hides inside the word attributable. Traffic is not revenue. Leads are not revenue. Revenue is a closed deal you can trace back to the site — and tracing is the actual discipline this article is about.
Why the math changed: the denominator collapsed
The single biggest 2026 update to website ROI isn't about measurement — it's that costs dropped 5–10×. A credible business site with a lead pipeline used to be a $15,000–30,000 agency project plus monthly retainers. Built with AI coding tools — the approach I described in the vibe coding guide — the same capability costs a fraction of that, often built by one person in weeks.
Run the sensitivity: at $20,000 of cost, a site closing one $5,000 client per quarter returns 0%. At $3,000 of cost, the identical site returns 567%. Same site, same customers — the economics flipped purely because the denominator moved. If you priced a website project years ago and walked away, that decision deserves a rerun with 2026 numbers (I keep a current breakdown in how much a website costs).
Tracking revenue: the attribution discipline
E-commerce has it easy — the site is the register, analytics shows revenue per channel out of the box.
Service businesses have to build the chain: every lead captured with its source (which page, which search query, which form), every lead tracked through a pipeline to closed/won or lost, and a monthly reconciliation of closed deals back to first touch. This sounds like enterprise CRM territory. It isn't anymore.
On my own site, every inquiry lands in a pipeline with its source and status, and search data flows in through the Search Console API — so “which article did our last client read first?” is a lookup, not an archaeology project. The whole apparatus was built with an AI coding agent across a few evenings.
The point isn't my setup. The point is that the measurement layer — the thing that made ROI an unanswerable question for small businesses — became a weekend project. You can build it yourself if you have the time and desire, or hire someone who already runs one.
Counting true costs
Honest ROI counts everything: build cost (amortized over ~3 years), hosting and domains, content creation time (yours counts — price it at what your hour is worth), tools and subscriptions, and any marketing spend driving traffic to the site. Most businesses undercount their own time and overcount their traffic's value; do the opposite and your number becomes trustworthy.
How to improve website ROI (in order of leverage)
1. Answer leads faster. Response time is the cheapest multiplier on the whole list — an inquiry answered within the hour converts dramatically better, and AI can draft the personalized reply for you to approve (the pipeline mechanics are in the lead generation guide).
2. Push striking-distance pages. Your fastest new revenue is usually a page already ranking 5–20 for a commercial query — one improvement away from page one, free. Finding these programmatically is a core part of how I run SEO.
3. Cut costs that don't convert. With measurement in place, you'll find spend attached to pages and channels producing nothing. Kill it — cost cuts flow straight to the ROI numerator and denominator at once.
Frequently asked questions
How do you calculate website ROI?
ROI = (revenue attributable to the website − total website costs) ÷ total website costs × 100. The formula is trivial; the work is attribution — knowing which revenue the site actually caused. For a service business, that means tracking every lead's source through to closed deals, not guessing from traffic numbers.
What is a good ROI for a website?
Below 0% means the site is losing money; 100–400% is healthy; over 400% is excellent. But in 2026 the benchmark moved: AI-assisted development cut typical build costs by 5–10×, so the same revenue now clears a far lower cost bar. A site that would have broken even at 2020 prices is often strongly profitable at today's.
How did AI change the economics of websites?
Both sides of the ROI formula moved. Costs collapsed: a site that took an agency $15,000–30,000 to build can now be built with AI coding tools for a fraction of that, often by one person. And measurement got automated: analytics and Search Console data flow through APIs into dashboards, so attribution stops being a quarterly spreadsheet exercise.
How do I track which revenue comes from my website?
Capture the source on every lead (which page, which query, which form), keep leads in a pipeline with statuses through to closed/won, and reconcile monthly. If you can't answer “which article did our last customer read first,” you're not yet measuring — the tooling for this is now cheap enough that there's no excuse to guess.
Is a website still worth it for a small business in 2026?
More than before — precisely because the denominator shrank. When a credible site plus lead pipeline costs a few thousand dollars instead of twenty, even a handful of closed deals per year clears the bar. The honest exception: if your customers all come from one marketplace or referrals and you won't invest in content, a website is a business card, and business cards don't have ROI.
Not sure which type of problem you have?
That’s the most common answer. Let’s figure it out together.