Most monthly website reports are long, full of charts and rarely change a decision. Sessions are up, pageviews are down, and nobody in the meeting can say whether the website did its job. Well-chosen website KPIs fix that by answering one question: is the site helping the business get what it needs?
This guide sorts website numbers into KPIs, diagnostics and vanity metrics, then shows how to choose three to five KPIs per goal, build a one-page monthly scorecard and read changes without over-reacting to noise. For the tracking set-up underneath, see how to track website enquiries in GA4.
The short answer
Website KPIs are the few numbers that show whether a website is meeting its business goal. For a site built to win enquiries, they are usually:
- Qualified enquiries: real prospects, counted in your CRM or inbox, not raw form submissions.
- Cost per qualified lead: what you spend to attract them, divided by how many arrived.
- Enquiry-to-sale rate: the share of qualified enquiries that become customers.
Diagnostic metrics such as conversion rate by page, engagement rate, non-brand organic clicks and Core Web Vitals pass rate explain why those KPIs moved. Raw pageviews and time on site are vanity metrics unless they answer a specific question.
Three kinds of website numbers
Unhelpful reports mix up three jobs. Business KPIs show whether the site is doing its job. Diagnostic metrics explain why a KPI moved and where to act. Vanity metrics feel informative but rarely change a decision.
One question sorts them: if this number fell by a fifth next month, what would we do differently? If there is no clear answer, it doesn’t belong in the report on its own.
Business KPIs: the top of the report
Qualified enquiries. Website enquiries minus spam, job applicants, suppliers pitching to you and customers asking for support. Agree with sales, in writing, what “qualified” means: the right type of project, a location you serve, a realistic timescale. Count it in your CRM or enquiry log, not analytics alone, which can undercount when visitors decline cookies and can’t tell a genuine prospect from spam.
Cost per qualified lead. Website marketing spend divided by qualified enquiries. Decide once what counts as spend (ads, content, SEO work, hosting and care) and keep it fixed, or the trend means nothing. Work it out per channel if you run paid campaigns.
Enquiry-to-sale rate. The quality check on everything above: a campaign that doubles enquiries and halves the close rate hasn’t helped. Sales take time to close, so measure a cohort, such as enquiries received three months ago. Add average deal value and you can calculate the ROI of your website.
Diagnostic metrics: what explains the KPIs
Conversion rate by page. A site-wide rate blends your blog, contact page and service pages into one unhelpful average. Look at the pages meant to produce enquiries; GA4’s landing page report shows sessions and key events for each entry page, so you can compare them. The conversion rate optimisation guide covers improving the weak ones.
Engagement rate. GA4 counts a session as engaged if it lasts longer than 10 seconds, includes a key event or has at least two page views; bounce rate is simply the reverse. Use it to compare pages and campaigns: a paid campaign far less engaged than organic traffic is probably reaching the wrong people.
Organic clicks. In Search Console’s Performance report, track clicks rather than impressions, and exclude branded searches with the branded queries filter where it is offered, or a “Queries not containing” filter for each spelling of your brand name. Non-brand clicks show whether you are found for what you sell. At the time of writing (June 2026), AI-generated answers on results pages can mean some searches end without a click, so read any fall alongside impressions and queries.
Core Web Vitals pass rate. The share of URLs that Search Console’s Core Web Vitals report rates as Good on mobile. Google’s thresholds are Largest Contentful Paint of 2.5 seconds or less, Interaction to Next Paint of 200 milliseconds or less and Cumulative Layout Shift of 0.1 or less, at the 75th percentile of real visits. Smaller sites may lack the data to appear. Core Web Vitals explained covers each metric and its fixes.
Vanity metrics: only with context
- Raw pageviews and sessions rise with bot traffic or an off-topic post that happens to rank, and can fall for good reasons, such as stopping a campaign whose visitors never enquired.
- Time on site (average engagement time in GA4) misleads both ways: a long visit to the contact page may mean confusion, and a short visit that ends in a phone call is a success.
- Bounce rate on its own can make a page that answers a quick question well look like a failure.
- Follower counts and authority scores say little about enquiries, and third-party authority scores are an SEO tool’s estimate, not a Google measure.
Each becomes a diagnostic the moment it answers a specific question, such as “did the new article series send visitors to our service pages?”
How to choose website KPIs for each goal
Start with the website’s job, which your website strategy should set out: who the site serves and what it must get them to do. Most sites have one main goal and perhaps a secondary one, each with its own short list of KPIs.
| Website goal | Business KPIs | Diagnostics to watch |
|---|---|---|
| Win enquiries | Qualified enquiries, cost per qualified lead, enquiry-to-sale rate | Conversion rate on service and landing pages, form completion, non-brand organic clicks |
| Take bookings | Completed bookings, cost per booking, share of bookings that attend | Drop-off at each booking step, mobile conversion rate |
| Recruit people | Qualified applications per role, hires who applied through the site | Application rate from job pages |
| Support the sales team | New customers who say they used the site before buying, branded search clicks | Views of case studies, pricing and proof pages |
If you sell online, the headline numbers are revenue, orders and conversion rate, and measuring e-commerce conversion rate has its own traps.
Balance volume, cost and quality
Pair a volume KPI (how many) with a cost KPI (at what price) and a quality KPI (how many became business) and the set is hard to game. Chase enquiries alone and you will celebrate a spam surge. Chase cost alone and you may cut the campaign that brought your best customers.
Write each KPI down
Many reporting arguments are really arguments about definitions. Give each KPI one definition, one source and one owner:
| KPI | Definition | Source | Owner |
|---|---|---|---|
| Qualified enquiries | Website enquiries sales accepted as real prospects | CRM, lead source = website | Sales |
| Cost per qualified lead | Monthly website marketing spend ÷ qualified enquiries | Finance and CRM | Marketing |
| Enquiry-to-sale rate | Customers won from enquiries received three months earlier | CRM | Sales |
The GA4 enquiry tracking guide shows how to check that analytics counts the enquiries your team really receives.
Build a one-page monthly website scorecard
A monthly website report should fit on one page and take two minutes to read. If a number needs a paragraph to justify being there, it belongs in a quarterly review.
- Headline KPIs. Three to five numbers, each shown for this month, the same month last year, the three-month average and the target, with a status: on track, watch or act.
- Diagnostics that moved. Only metrics that left their normal range, each with the likely reason. If nothing moved, say so.
- Change log. New pages, campaigns started or stopped, plugin or tracking changes, holidays, outages. Most “mysterious” movements are explained here.
- Actions. No more than three, each with an owner and a date, plus a one-line update on last month’s.
Open with one sentence on what the numbers mean, for example: “Qualified enquiries held steady, the close rate rose after the new pricing page went live, and a broken form is fixed.”
Before it goes out:
How to read changes without over-reacting
Monthly numbers wobble. The skill is telling the wobble from a real change.
Compare like with like
Month-on-month comparisons mix in seasonality, working days, holidays and campaign timing, so compare with the same month last year and the rolling three-month average. After a launch or redesign, treat the first three months as a fresh baseline, since the old site’s tracking probably counted different things.
Know your normal range
A rough statistical rule for counts such as enquiries: most months land within about two square roots of the average either side. For a hypothetical site averaging 25 qualified enquiries a month, the square root is 5, so anything from about 15 to 35 is ordinary variation, not a signal. Real business data usually swings a little more than that, so treat the range as a minimum. Act on a result outside it, or on three or more months moving the same way.
Check tracking before strategy
A sudden drop to almost nothing is rarely a collapse in demand. It is usually a form that stopped sending, a tag lost in an update or a cookie banner change. Send a test enquiry and confirm it arrives before anyone rewrites the homepage.
Walk down the chain
When a KPI moves, trace it back: traffic to the pages that produce enquiries, their conversion rate, form completion, then qualification. The first step that changed is where to look. Segment by device, channel, landing page and language, because averages hide problems that affect one group.
| What you see | Check first | Often means |
|---|---|---|
| Traffic up, enquiries flat | Where the new visitors landed and came from | Low-intent traffic, bots or a broad campaign |
| Enquiries up, close rate down | Lead sources and the enquiries themselves | Spam, or a campaign reaching the wrong audience |
| Non-brand organic clicks down | Queries and pages in Search Console | Lost rankings, or results pages answering the query themselves |
| Core Web Vitals pass rate down | Which URL group changed status | A new script, plugin or heavy images (fixes take weeks to show, as the report uses the last 28 days of data) |
If the same weaknesses recur month after month, that pattern is a finding in itself: check the signs your website needs a redesign rather than keep patching symptoms.
Frequently asked questions
How many KPIs should a website have?
Three to five per website goal, with one marked as the primary measure. Beyond that, the report becomes a list nobody acts on.
Is bounce rate still a useful website metric?
Only as a diagnostic. In GA4 it is the share of sessions that were not engaged, and a high figure on a page that answers a quick question can be perfectly healthy.
What should you benchmark your website’s conversion rate against?
There is no universal figure worth chasing, because rates depend on the offer, price, traffic source and what counts as a conversion. Your own history is the most reliable benchmark.
How often should website performance be reviewed?
Check enquiries and uptime weekly to catch a broken form early, report KPIs monthly, and revisit goals, targets and definitions quarterly.
What to do next
Pick the website’s main goal, choose three to five KPIs that measure it, write each one down with a source and an owner, and make sure tracking counts what your team really receives. Then act on what the numbers show: our guide to continuous improvement after launch sets out a routine.
For a clear starting point before setting targets, ask for a free website audit. We review what sits behind several of these diagnostics (speed, mobile experience, Google visibility, security and how easily visitors can get in touch); tell us which number worries you and we will look at it closely. Our Business and Growth website care plans add a monthly plain-language report on what was updated and checked, which makes your change log easier to keep.