It depends on what the website is for. If it exists to bring in inquiries, track six measures: inquiries, conversion rate by traffic source, the landing pages that lead to inquiries, engaged visits to commercial pages, cost per inquiry, and page speed. If people use your product on the site, swap the inquiry-specific measures for activation, drop-off, and return rates. Pageviews, impressions, and site-wide time on page can leave the monthly review.
Six isn’t a magic number. My rule for any business scorecard is four to seven metrics, depending on the business, and no vanity metrics: each one has to move the needle. Website analytics is where that rule gets broken most often, because the tools show you everything by default.
Why Most Default Analytics Reports Don’t Help
Analytics tools are built to serve every kind of website, from a local accounting firm to a national retailer with a marketing department. So they show everything they can collect: users, sessions, pageviews, events, devices, cities, browsers, screen resolutions. None of it is wrong. Most of it doesn’t help a small business decide anything.
The useful distinction is between two kinds of numbers:
- Vanity metrics look good in a report and require no action. Pageviews went up 12%. Great. What do you do differently on Monday?
- Commercial metrics track inquiries, customers, and the cost of getting them. When one moves, someone has a reason to act.
The cost of watching the wrong numbers is quiet. A business can spend months redesigning pages to raise time on site while inquiries slide, and nobody notices because the report everyone looks at went up. The test I use on any metric is whether it helps the business make more revenue, run more efficiently, or cut costs. If it does none of those, it comes off the scorecard.
If you haven’t looked at what your site already records, start there. Contact form submissions, booking confirmations, and phone logs usually exist before anyone opens an analytics tool. The business data you already have covers how to find them.
What I Track on My Own Two Sites
I run two sites that do different jobs, and I’ve only just started measuring both. What follows are the metrics I’m setting up, not results.
johnserra.com is meant to generate inquiries. It uses GA4. I’m tracking two things: the share of visitors who complete a high-intent action (an assessment or the contact form), and where qualified visitors come from, split by referral and search and broken down by page.
CareerTalkLab is a product. It’s a community whose members learn from and teach each other to advance in data and software careers, and I measure it with Umami. I’m tracking the share of visitors who start a lesson, completion and drop-off by module, and how many new learners come back on Day 7 and Day 30.
Both sites get one technical metric: how long the slowest page loads and server responses take, measured at the 95th percentile.
Those are target measures across two sites, not a combined scorecard or a claim that I already have results. The lists differ because the sites do different jobs. An inquiry site succeeds when a stranger reaches out. A product site succeeds when someone starts using it and comes back. Start with what your site is for, then pick four to seven measures for that site; a generic list of “top website KPIs” skips that step.
The Five Metrics for a Site That Brings in Inquiries
1. Key Conversion Actions
A conversion is an action that moves a stranger into your sales pipeline. Count those, not visits. What counts depends on the business:
- Professional services and consulting: completed contact forms, booked discovery calls, clicks on your email address.
- Local trades and service businesses: click-to-call taps, quote requests, requests for directions.
- Online stores and software: purchases, checkout starts, free trial signups.
In GA4, you mark these actions as key events (Google’s current name for what it used to call conversions). Other tools call them goals or conversions. If you track nothing else on your website, track the total number of these actions each month and compare it with a target.
2. Conversion Rate by Traffic Source
Your overall conversion rate blends every source together and hides where buyers come from. Split it by channel:
- Organic search: people who found you on Google or Bing, often while describing a specific problem.
- Direct: visits with no identifiable source. This can include typed addresses and bookmarks, but also links from apps or messages that pass no referrer.
- Referral: visitors from other websites, such as directories, associations, and partners.
- Social: visitors from platforms like X or LinkedIn. These often bring attention more than inquiries.
- Paid: ad clicks, if you run ads. These need the tightest tracking because you pay for every visit.
Then compare. Here is an example with made-up numbers. A source that sends 500 visits at a 4% conversion rate produces 20 inquiries. A source that sends 5,000 visits at 0.1% produces 5. The smaller source is worth four times as much, and it’s the one a traffic report makes look minor.
Source data from analytics tools is incomplete. Ad blockers, consent choices, and people who switch devices can break the trail. A “How did you hear about us?” field on your contact form fills some gaps, and tagging campaign links you control with UTM parameters helps identify those visits.
3. Top Converting Landing Pages
Visitors can arrive through your homepage, a service page, an article, or a guide. Your landing pages are the entry points; find which ones actually lead to inquiries rather than assuming the homepage does all the work.
Check two things each month:
- Which pages bring in the visitors who go on to convert?
- Does each high-traffic page give visitors a clear next step: a form, a phone number, or a link to the relevant service?
The common problem is a popular article with no conversions. It brings in the right readers and gives them nowhere to go. The fix is usually a clear next step near the top and bottom of the page, or a link to the service it relates to.
4. Engaged Visits to Commercial Pages
Raw traffic mixes useful visits with accidental clicks and visits that end quickly. Engagement is a helpful filter, but it cannot tell you by itself whether a visitor was a qualified buyer or even rule out automated traffic.
GA4 counts a session as engaged if it lasts longer than 10 seconds, includes a key event, or includes two or more page views. Engagement rate is the share of sessions that meet that bar. Privacy-first tools like Umami don’t use the same definition, so there the practical measure is unique visitors to your commercial pages: services, pricing, about, and contact.
A large traffic spike with almost no engagement is worth investigating. Check its sources and conversions before calling it a marketing win or deciding what caused it.
5. Cost per Inquiry
A website costs money and time. Measure what each inquiry costs you:
Cost per inquiry = (monthly spend on the site and its marketing + hours spent × your hourly rate) ÷ inquiries that month
Here’s an example with made-up numbers: a firm spends $200 a month on hosting, tools, and a small ad budget, and someone spends 6 hours a month on content at $50 an hour. That’s $500. With 20 inquiries, each costs $25. If the same firm spent $500 and got 2 inquiries, each would cost $250, and that’s a reason to look at whether the time would go further on direct outreach.
Count inquiries, not every form submission. Spam and job applicants through the contact form will make the site look cheaper than it is.
If Your Website Is the Product
For software, online courses, memberships, and tools, an inquiry isn’t usually the goal. Someone using the product is. Keep two measures from the inquiry scorecard—valuable conversion actions and traffic source—then replace the landing-page, engagement, and cost-per-inquiry measures with:
- Activation rate: the share of new visitors who take the first real product action, such as starting a lesson, creating a project, or running a first report.
- Drop-off by step: where people stop in a sequence, whether that’s a course module, an onboarding step, or a checkout page.
- Cohort return rate: of the people who signed up in a given week, the share who come back on Day 7 and Day 30.
With speed, that is a six-measure product-site starting scorecard. These are the kinds of measures I’m setting up for CareerTalkLab. They answer the question a product site actually has to answer: do people who arrive start using it, and do they keep using it?
The Metric Every Site Needs: Speed
A slow page loses visitors before any other metric has a chance to count them. Measure the slow end, not the average. The 95th percentile (P95) is the time within which 95% of page loads finish. An average of 1.5 seconds can hide a meaningful share of visitors waiting six.
You don’t need paid tools to start. Google’s PageSpeed Insights and the Core Web Vitals report in Search Console are free and show real-user field measurements when a page or site has enough data. Google reports those values at the 75th percentile, not P95, but for most small business sites that is a good enough place to start. If speed turns out to be a real problem, a real-user monitoring tool can report P95 directly.
The Cut List: Metrics to Stop Reviewing Every Month
These don’t need to be deleted from your analytics tool. They just don’t belong on the scorecard you review.
- Raw pageviews. Refreshes, back-button clicks, and multi-page wandering inflate them. More pageviews don’t mean more business.
- Bounce rate, by itself. Under the old Google Analytics definition, a visitor who read a whole page, found your phone number, and called still counted as a bounce. GA4 now defines bounce rate as the share of sessions that weren’t engaged, which is better, but engagement rate and conversions tell you the same thing more directly.
- Site-wide average time on site. Tabs left open and one long visit can skew it, and a longer visit isn’t better if the visitor couldn’t find what they needed.
- Social impressions. How many people saw a post on another platform says little about whether they visited your site, let alone contacted you.
- Keyword rankings in isolation. Ranking first for a phrase nobody searches, or one that attracts people who will never buy, produces nothing. Rankings matter only when they bring engaged visitors to pages that convert.
A 15-Minute Monthly Website Review
Once a month, with your scorecard open:
- Record conversions for the prior month against your target.
- Check conversion rate for your top three traffic sources. Note any that changed sharply.
- Find the top converting landing page and the page with the most engaged visits but the fewest conversions.
- Check speed on your two or three most important pages.
- Write down one action, with a name and a date. “Add a consultation link to the top article, Sam, by the 15th.” “Fix the phone link that doesn’t work on mobile.”
The last step is the one that makes the review worth doing. A monthly number nobody acts on is just another report. How to get your team to actually use your reports covers how to run that conversation so the action happens.
If you use GA4, its Traffic acquisition, Landing page, and event reports can help with this review. Check that the actions you count as inquiries are actually recorded.
Website Metrics Checklist
- I know what my website is for: inquiries, sales, or product use.
- I track the actions that matter as conversions or key events.
- I can see conversion rate by traffic source, not just overall.
- I know which landing pages bring in converting visitors.
- I review engaged visits, not raw traffic.
- I know roughly what each inquiry costs me.
- If the site is a product, I track activation, drop-off, and return rate.
- I check page speed at the slow end.
- My scorecard has four to seven metrics, and I review it monthly.
Frequently Asked Questions
How many website metrics should a small business track?
Four to seven for each site’s scorecard. An inquiry site can start with the five measures above plus speed. A product site can keep conversions and traffic source, replace the inquiry-specific measures with activation, drop-off, and return, and also watch speed. Cut or combine measures when they don’t lead to a decision. How many KPIs should a small business track explains the wider business-scorecard principle.
Is bounce rate still important?
Less than it used to be. GA4 redefined it as the opposite of engagement rate, so looking at both is redundant. Engagement rate and conversions tell you more.
Do I need Google Analytics?
No. GA4 is free and detailed, but it’s also complicated. Privacy-first tools like Umami or Plausible are simpler and cover page, referral, and event tracking. What matters is that you can see conversions, sources, and landing pages in whatever tool you use.
How often should I check website analytics?
Monthly for the scorecard. More often only when you’re testing something specific, such as a new landing page or a campaign, and you know in advance what number you’re waiting to see.
What’s a good conversion rate for a small business website?
It depends on the industry, the offer, and where the traffic comes from, so a generic benchmark won’t tell you much. Your own trailing three-month average is the more useful baseline. Improve against that.
If you want a second pair of eyes on your scorecard, or help deciding what belongs on it, get in touch.