Website ROI is hard to calculate honestly, because the cost side is visible and the return side is invisible. You know what you paid. You don't know who arrived, couldn't tell what you do, and left without leaving a trace. The way through it is to stop trying to value the new site and start pricing the current one: every visitor your site loses has a market price, because the ad market sets one every day. If a click on your capability costs $31, then a visitor who leaves in three seconds cost you $31, whether or not you were the one paying.
That reframe is the whole method. Here's how to run it.
Why the usual website ROI calculation doesn't work
The standard formula is fine on paper. Take the revenue attributable to the site, subtract what the site cost, divide by the cost. The trouble is the first term.
For an established B2B company, almost nothing arrives with a clean label on it. A buyer hears your name at a trade show, looks you up that night, decides you look credible, and calls three weeks later. Your CRM records a phone call. The website did the work and got no credit. Run the same story with a site that looks a decade old and the buyer never calls, and now nothing is recorded at all. Both outcomes look identical in your analytics, which is why the number people usually compute is the wrong number carefully.
So the honest version doesn't start with revenue. It starts with what a visitor costs. (What you pay is a separate question, and we answer it in what a B2B website actually costs.)
What is a website visitor actually worth?
You don't have to guess, because advertisers bid on this every day and Google publishes what they pay. We priced twenty capability terms a manufacturer would genuinely bid on, in the United States, in September 2026:
- cnc machining services, 8,100 searches a month, $31.34 a click
- cnc turning, 9,900 searches, $25.74
- injection molding services, 1,300 searches, $19.72
- swiss machining, 1,900 searches, $16.08
- sheet metal fabrication, 9,900 searches, $14.23
- precision machining, 5,400 searches, $12.05
The median across all twenty was $12.36, and the specific-process terms, the ones a buyer types when they already know what they need, sit at the top of the range.
Now apply that. Suppose 400 people land on your site in a month from search. At the median, that traffic has a replacement cost of roughly $4,900. You didn't pay it, which is exactly why it feels like it's worth nothing. It isn't. It's the price you would pay to buy the same attention tomorrow, and it's the correct denominator for everything below.
The three numbers that make website ROI computable
One: what the site currently converts. Count the quote requests, form fills and tracked calls that came through the site in the last twelve months. Divide your last build cost by that number. That's your realized cost per quote from the website, and for most established companies it lands somewhere between embarrassing and fine.
Two: what the site currently loses. This is the number hardly anyone computes, and the one that decides the answer. Across the 55,000+ US B2B websites we analyzed, 57% give a buyer no clear next step above the fold, no obvious way to start a quote and no visible way to reach a human. One in three keep the phone number in the footer only. Separately, our speed benchmarks score 56,005 B2B homepages on Google PageSpeed's mobile strategy: the median takes 3.7 seconds to show anything, and 70.5% land beyond the three-second mark.
Put those together. If 400 visitors arrive and the site behaves like the median one in our corpus, a large share of them leave before anything loads or without finding a next step. At $12.36 a visitor, that isn't a design problem. It's a line item.
Three: what a quote is worth to you. You already know this one, and it's the number the two above have to be measured against. An owner who knows their average order value and their close rate on inbound quotes can finish the calculation in about ten minutes. If you don't yet track the first two, our guide to the website KPIs worth tracking covers the handful that feed this and the ones that waste your time.
How to compare a website against your other channels
The comparison that settles most of these decisions isn't website versus nothing. It's website versus the thing you'd spend the money on instead.
Take one trade show. Add the booth, the freight, the travel, the staff time and the three days almost nobody is doing their normal job. Divide by the qualified conversations it produced. Then do the same arithmetic for a rebuild, using the $12.36 replacement cost per visitor and your actual traffic, spread across the three to five years the site will be in service.
The trade show wins on intensity and loses on duration. It happens once. The site runs every day, including the nights when a buyer is checking you out at eleven o'clock after a bad experience with their current supplier. That's not an argument against trade shows. It's an argument for noticing that one of them is amortized over three days and the other over three years, and that most companies scrutinize the cheaper one harder.
When the ROI case is weak, and you should hear that
Sometimes the honest answer is that a rebuild won't pay.
If every job you win comes from repeat customers and referrals, and hardly anyone who doesn't already know you is searching for what you do, then a new website is a credibility asset rather than a lead generator. That's a real reason to build one, and it's a different business case with a different bar. The test takes two minutes: open Google Search Console and look at whether you get impressions for anything other than your own company name. If the answer is no, buyers aren't looking for you yet, and the first problem is demand, not design.
The other weak case is a site that's already fast, already says what you do in the first line, and already has an obvious way to start a quote. If that's yours, the money belongs somewhere else this year, and ongoing search work is usually the better place for it. We'd rather say so.
