Most website reporting measures activity rather than outcome. Sessions are up, bounce rate is down, and nobody can say whether the site made any money. Here is what to track instead.
Measure website ROI by tracking qualified enquiries and revenue attributable to the site against total cost of ownership, including build, hosting, maintenance and content. The metrics that matter are conversion rate, cost per enquiry and enquiry-to-customer rate. Sessions and bounce rate are diagnostics, not outcomes.
Start with the cost side
- Build cost, amortised over the life of the site, typically three years.
- Hosting, domain and any platform subscriptions.
- Maintenance and support.
- Content production.
- Paid traffic, if you are sending any.
- Internal time, which is real even though nobody invoices for it.
Total these annually. Most businesses have never done this and are surprised by the number, usually because content and internal time were never counted.
Then the return side
| Metric | What it tells you |
|---|---|
| Qualified enquiries | Enquiries you could actually serve, not the total |
| Enquiry to customer rate | How many become clients |
| Average customer value | What one is worth |
| Attributed revenue | Enquiries times conversion times value |
| Cost per enquiry | Total cost divided by qualified enquiries |
Qualified matters. A site producing forty enquiries a month of which two are serviceable is performing worse than one producing eight that all are.
The metrics that mislead
- Sessions. Useful as a diagnostic, meaningless as a goal. Traffic that never converts costs money to serve.
- Bounce rate. A single-page visit where someone got your phone number and called is a success that reports as a failure.
- Time on page. Could be engagement, could be confusion.
- Keyword rankings. Position one for a term nobody searches with intent is not revenue.
- Impressions. Being seen is not being chosen.
None of these are useless. They are diagnostics that explain why the outcome numbers moved. They are not the outcome.
Setting it up so you can actually measure
- Define what counts as a conversion before you measure anything.
- Track form submissions, calls and any other genuine enquiry route. Untracked phone calls are the most common blind spot for service businesses.
- Tag traffic sources so you know what produced what.
- Record enquiry outcomes in your CRM so you can connect enquiries to revenue.
- Review quarterly rather than weekly. Monthly noise on small numbers leads to bad decisions.
A realistic benchmark
Conversion rates vary enormously by industry and traffic type, and anyone quoting a universal number is guessing. What matters more is your own trend and your cost per qualified enquiry against what that enquiry is worth.
If a customer is worth several thousand and your site produces qualified enquiries for a small fraction of that, the site is working regardless of what the sessions chart does.
Frequently asked questions
How do I calculate website ROI?
Total the annual cost of the site including build amortisation, hosting, maintenance and content. Then calculate revenue attributable to it: qualified enquiries times enquiry-to-customer rate times average customer value. ROI is return minus cost, over cost.
What is a good website conversion rate?
It varies far too much by industry and traffic source for a universal figure to be useful. Compare against your own baseline and your cost per qualified enquiry. A rate that looks low can be excellent if the customers are valuable.
Is bounce rate important?
As a diagnostic, sometimes. As a goal, no. A visitor who lands on your contact page, takes the phone number and calls registers as a bounce and is a success. Judge outcomes rather than proxies.
How do I track phone calls from my website?
Use call tracking, which assigns a unique number to web visitors so calls can be attributed. Without it, service businesses systematically undercount their website's contribution, sometimes by a wide margin.
How long before a new website shows a return?
Three to six months for the site to be indexed, ranked and generating a steady flow, longer in competitive markets. Judging a site at six weeks tells you about your launch, not about whether it works.