Website Analytics for Beginners: The Five Numbers a Small Business Should Actually Watch

Quick answer

Most small businesses have analytics installed and never look at it, or look at it and see noise. The fix is ruthless reduction: a free three-tool stack (Google Analytics 4, Google Search Console, and optionally a behaviour tool like Microsoft Clarity), checked monthly, not daily, for exactly five numbers: visitors by source, enquiries/conversions, top landing pages, the searches you’re found for, and site speed. Everything else is detail behind those five. The purpose isn’t reporting — it’s answering one question: where should next month’s effort go?

These numbers are only as good as the site behind them; a professionally built WordPress website arrives with analytics wired in from launch.

The free stack, and what each tool answers

Setup belongs at launch (it’s on the launch checklist for a reason) — but the second-best time is today, because analytics only see forward. UK privacy note: GA4 and similar tools legally need consent via your cookie banner; a properly configured banner with Consent Mode keeps you compliant while preserving modelled data.

The five numbers, monthly

1. Visitors by source (GA4 → Acquisition): organic search vs direct vs social vs paid. The story is the trend: a growing organic line means content and SEO are compounding; a social-only chart means the rented-audience problem in data form.
2. Enquiries, the number that is the point (GA4 → Conversions): form sends, calls clicked, quote requests. If you track one thing, track this; if it isn’t configured, configure it this week, because “traffic” without enquiries is entertainment. Divide enquiries by visitors for a conversion rate: 1 to 5% is the normal small-business band; materially below that with decent traffic means the site, not the marketing, is the leak (the usual suspects).
3. Top landing pages (GA4 → Engagement): where people actually arrive, routinely a surprise, and it tells you which pages deserve the conversion attention, internal links and updates.
4. Queries and positions (Search Console → Performance): the searches you appear for, clicks, and average position. Two gold veins: queries at positions 8 to 15 (nearly page one — improve those pages first) and high-impression/low-click queries (title and description rewrites). This report is also next month’s content list, free.
5. Speed (PageSpeed Insights, quarterly is fine): slow pages tax every other number. Fix reds on the pages from #3 before anything exotic.

The vanity metrics to ignore

Bounce rate (a blog post read fully and left is a “bounce”, meaningless without context), raw pageviews (entertainment), session duration (skewed by tabs left open), and rankings checked daily (they wobble; the monthly trend is the signal, and SEO timelines are months regardless). The tell of a useful metric: you can name the action you’d take if it moved. If you can’t, stop watching it.

The monthly 20-minute ritual

Same time each month: the five numbers into a simple sheet (or just compare against last month on screen), then one decision: improve the striking-distance page, fix the leaking form, write the post that answers the rising query, double down on the source that’s growing. One decision, executed, monthly, beats a quarterly deep-dive that produces a document and no changes. When paid channels enter the picture the same discipline applies with money attached — cost per enquiry from each channel becomes the sixth number, and the arithmetic in your email and lead capture starts paying for itself in data as well as sales.

Setting up the one metric that matters

Conversion tracking is where most small-business analytics quietly fails: the tags are on, the traffic is counted, and nobody ever configured what counts as an enquiry. The practical sequence, in an afternoon:

1. Write down every way someone can contact you. Contact form, quote form, phone number tap, WhatsApp link, mailto link, booking widget, live chat. Most sites have five or six routes and track one.
2. Give the form a thank-you page. A dedicated URL like /thank-you/ is far more reliable to track than an on-page success message, and it gives you a clean page to measure. If your form can’t redirect, that’s worth fixing first.
3. Mark the events as key events in GA4. Phone taps and mailto clicks fire automatically as outbound or click events; the form submission needs the thank-you page view marked. Without this step, GA4 shows you activity, not outcomes.
4. Exclude yourself and your suppliers. Internal IP filtering, or at minimum the browser opt-out add-on. A five-person office refreshing the site daily can account for a fifth of a small site’s sessions and make every other number lie.
5. Test it properly. Submit a real enquiry from a phone on mobile data, then check it appears in GA4’s realtime report. A great many tracking setups have never once been tested end to end, which is why they fail silently for months.

Do this on the day the site launches, alongside the other pre-go-live checks, and you’ll have a full year of comparable data by the first anniversary.

The attribution gap for phone-led businesses

If most of your work arrives by phone, analytics is structurally blind: it sees the visit that preceded the call, not the call. Three fixes, in ascending cost. Ask every enquirer “how did you find us?” and log the answer in one column of a spreadsheet: unfashionable, imperfect, and still the most valuable dataset most trades businesses could start keeping this week. Second, use a distinct number on the website only, so calls to it are unambiguously web-sourced. Third, call tracking software (roughly £20 to £60/month) that swaps the displayed number per traffic source and reports channel by channel; worth it only once you’re spending real money on more than one channel and need to know which one to cut.

The reason to bother is that channel-level cost per enquiry is unknowable otherwise, and reallocating half a small business marketing budget on a hunch is a far bigger error than any tracking subtlety.

Reading small numbers without fooling yourself

A site with 400 visits a month and eight enquiries has a 2% conversion rate. Next month it does 400 visits and five enquiries. Nothing has broken. That is ordinary variation at low volumes, and treating it as a signal is how businesses redesign perfectly good pages. Some working rules:

• Below roughly 30 conversions a month, compare quarters rather than months.
• Judge a change against the same month last year as well as last month; seasonality is stronger than most owners expect, particularly in trades and hospitality.
• A single page’s ranking movement of a few positions is noise. A whole cluster of pages moving in the same direction is not.
• Beware the traffic spike with no enquiries: it is usually a bot wave, a referral from a scraper site, or one blog post going briefly viral with an audience that will never buy.

The setup errors that silently corrupt the data

Turning the numbers into next month’s work

The reports only earn their keep when each one implies an action. Search Console’s query list is the clearest example: terms you already appear for at positions eight to fifteen are the cheapest wins available, because the page exists and only needs strengthening. Terms with impressions but no clicks usually need a better title, not more content. Terms you’d expect to see and don’t are the gap that ordinary keyword research for beginners is designed to close, and doing it from your own data beats doing it from a tool’s suggestions.

On the site side, the landing-page report tells you where to spend improvement money. If your three most-visited pages load slowly on mobile, that is not a technical footnote; it is a direct tax on every enquiry you might have had, and the fixes worth doing are ranked in why slow sites lose UK customers. If those pages get traffic and no enquiries, the problem is the page’s offer, proof and call to action rather than the marketing feeding it.

Once the five numbers are stable and honest, the natural next step is asking what each channel costs per enquiry and whether the answer justifies the invoice, the point at which analytics stops being reporting and starts driving budget. That is also when engaging help on search engine optimisation becomes a measurable decision rather than an act of faith, because you already know your baseline and can hold any supplier to it.

Sources & Further Reading

• GA4 Setup Guidance, Google
• Search Console Getting Started, Google
• Cookies & Consent Guidance, Information Commissioner's Office