Straight answer: Google Analytics tells you what people did on your site, not why they did it, and not whether it made you money. It’s a behaviour log dressed up as a business dashboard, and most owners I meet are making decisions off numbers that are quietly wrong. Used it’s still one of the best free tools you’ll ever run your business through, but only once you know what it can’t see.
If you want to go deeper on this: When the Bank Says No: How Small Business Owners Actually Fund Growth.
What it shows you (the bits that are real)
Strip away the jargon and GA4 is doing three things well: counting visits, tracking the path someone took through your site, and flagging which pages people leave from. That’s it. Everything else is inference layered on top.
- Users and sessions, so you know roughly how many distinct people turned up and how many separate visits happened
- Traffic source, so you can see whether people arrived from Google, a social post, an email, or a direct type-in
- Landing page and exit page, showing where visits start and where they die
- Engagement time and events, tracking clicks, scrolls, form starts, video plays if you’ve set those up
- Conversions, but only for the specific actions you’ve told it to count
That last point trips up almost everyone. GA doesn’t know a conversion happened unless you built the tracking for it. I’ve sat with three separate clients this year who had GA4 installed for over a year and zero conversion events configured. They were reading traffic numbers and thinking they were reading sales numbers. They weren’t the same thing at all.
A real example: the client whose traffic doubled and revenue didn’t move
Last spring I worked with a small coaching business owner whose GA4 dashboard showed a 96% jump in users over three months, from around 1,400 to just under 2,750 a month. She was thrilled. She’d been posting consistently on LinkedIn and assumed that was paying off.
We dug in. Almost all of that increase sat in one channel: “Direct / (none)”, GA’s label for traffic it can’t identify a source for. Real direct traffic (someone typing your URL from memory) doesn’t usually double in a quarter for a business her size. When we checked her email platform’s own click reports and her Google Search Console data side by side, the truth came out: a chunk of it was bot and crawler traffic her hosting company had recently started routing differently, and another chunk was legitimate visits from a paid directory listing she’d forgotten she was still paying for, which GA was misclassifying because the referral tag had broken months earlier.
Net result: her real, human, likely-to-buy traffic had grown by about 11%, not 96%. Enquiries hadn’t moved because the growth people-facing traffic hadn’t really grown much either. That’s a full working example of why you cannot read a top-line traffic chart and call it business insight. You have to open the channel breakdown, and honestly you often need a second data source to sanity check it, which is exactly why I always tell people to run their Google Search Console data alongside their GA4 reports, because Search Console shows what Google itself thinks people searched and clicked, and it will contradict GA more often than you’d expect.
The uncomfortable bit nobody puts in these articles
Here’s the part that gets glossed over in most explainers: Google Analytics has a built-in blind spot that’s getting bigger every year, and it’s not a bug, it’s the direct result of privacy law. Under UK GDPR and the EU’s rules, if a visitor declines cookies through your consent banner, GA4 either doesn’t fire at all or fires in a stripped-down “consent mode” that estimates the missing data using modelling, not measurement. Google’s own documentation admits this modelling is an approximation. Depending on your audience and your banner design, industry estimates put consent decline rates anywhere from 15% to over 40% of visitors on some sites. That means a meaningful slice of your real visitors, sometimes closer to a third of them, simply never show up in your reports as themselves. GA fills the gap with statistical guesswork and presents it back to you looking exactly as solid and countable as everything else on the dashboard. Nobody tells you which numbers are real counts and which are modelled estimates, because GA4 doesn’t visually distinguish them for you. You’re reading a report that’s part fact, part educated guess, with no label on which is which. I don’t say this to scare anyone off using it. I say it because if you’re making a five-figure ad spend decision off a GA4 conversion number, you should know that number has some fiction baked into it by design.
What GA cannot tell you, full stop
Even with perfect setup and full consent, there are things Google Analytics was never built to answer, and I still see people trying to squeeze answers out of it that just aren’t there.
- Why someone bought or didn’t buy. GA shows the exit, not the reason. You need actual customer conversations or survey data for that.
- Whether a customer is happy after they buy. Retention and satisfaction live in your CRM, your support inbox, and your renewal or refund numbers, not in a web analytics tool.
- Phone call and in-person conversions, unless you’ve bolted on call tracking software, which most small businesses haven’t.
- Cross-device journeys reliably. Someone researching on their phone on the train and buying on a laptop that evening often shows up as two completely separate, unconnected users.
- True profitability of a channel. GA can tell you a channel drove 40 leads. It cannot tell you what those leads cost to acquire, close, and service unless you connect ad spend and sales data yourself.
If you want a rounded picture, GA needs to sit next to at least two other things: your Search Console data for what people searched to find you, and your actual sales or CRM records for what happened once they became a lead. On their own, page view counts tell you far less than people assume, which is the exact trap I’ve written about separately when it comes to what landing page views really tell you. A page getting 3,000 views a month feels impressive until you realise none of those visits are staying long enough to read past the headline.
A step-by-step for reading it this week
Rather than staring at the home screen and feeling vaguely reassured or vaguely worried, do this instead:
- Check your conversion events are set up. Go to Admin, then Events, and confirm you have at least one marked as a key event for a form submission, purchase, or booking. If there’s nothing there, everything else in this list is decoration.
- Look at Traffic Acquisition, not just the homepage overview. Sort by “Session default channel grouping” and check whether Direct or Unassigned is unusually large. If it’s over a quarter of your total, something’s broken in your tagging.
- Cross-check top landing pages against Search Console’s top queries. If GA says a page got 800 visits from organic search but Search Console shows almost no clicks for related terms, one of the two tools has a tracking issue and you need to find it before you trust either number.
- Segment by new versus returning users. A traffic spike made up mostly of new users who never come back tells a different story than steady growth in returning visitors, which usually signals real interest building.
- Pull the exit pages report and read it like a complaint form. Wherever people leave in high numbers on a page that’s meant to lead somewhere, that page has a problem worth fixing before you spend another penny driving more traffic to it.
Do that once a week for a month and you’ll notice patterns a single glance at the dashboard never shows you. It took me doing this monthly for years across client accounts before I stopped trusting the top-line numbers and started trusting the cross-checks instead.
Where the tool earns its keep despite all this
None of the above means ditch it. Free, reasonably detailed behavioural tracking on your own site is still a useful thing to have, and it sits well alongside other free tools worth running your business through, which is exactly the kind of stack I cover in 30 free business tools with no trials or upsells. GA is at its most useful for spotting trends over time (is engagement time on your key pages going up or down quarter on quarter), for catching sudden drop-offs that flag a broken link or a page that’s stopped loading, and for understanding which content types hold attention versus which get a click and an instant bounce. It’s also worth pairing with a proper look at how you’re generating traffic in the first place, because the tool only reflects the quality of what you’re feeding it. If your lead sources are weak, no amount of clever GA analysis fixes that; you need better sources, which is the whole point behind building out a wider set of channels the way I’ve laid out in 50 unconventional lead generation tactics for 2026.
The honest way to use it going forward
Treat every number in GA4 as a starting question, not a finished answer. “Traffic from organic search is up 30%” should immediately be followed by “up in which pages, from which queries, and did any of those visits convert.” Ask that every single time before you make a decision or report a win to a client or a boss, and you’ll catch the false positives before they cost you money. Skip that step, and you’ll end up like my coaching client, celebrating a number that had almost nothing to do with real people finding her business.
Frequently asked questions
Can Google Analytics tell me if I’m making money?
Not directly. GA4 can show you that a conversion event fired, like a purchase or a booking form, but it only knows the revenue figure if you’ve explicitly passed transaction values into it through ecommerce tracking. Without that setup, it counts actions, not pounds.
Why does GA4 show so much “Direct” traffic that doesn’t make sense?
Direct or unassigned traffic is GA’s catch-all for anything it can’t confidently attribute to a source, which includes broken UTM tags, some app-based clicks, cross-domain tracking gaps, and increasingly, visitors who declined cookie consent. A high Direct number is usually a sign your tracking setup needs auditing, not a genuine flood of loyal fans typing your URL from memory.
Is Google Analytics data accurate if visitors reject cookies?
Only partly. When a visitor declines consent, GA4 either doesn’t collect data on them at all or uses Google’s consent mode, which fills gaps with modelled estimates rather than actual measured events. Your reports can look complete while a real portion of that data is statistical guesswork rather than a true count.
Should a small business trust GA4 over Google Search Console?
Neither on its own gives the full picture, and they measure different things: Search Console shows what happened before someone clicked, GA4 shows what happened after. Running them side by side, and querying the gaps when the numbers disagree, tells you far more than trusting either report in isolation.
