Open Google Analytics for the first time and you will find several dozen reports, hundreds of possible charts, and no indication of which ones affect your bank account. Website analytics for business has a signal-to-noise problem, and the noise is winning: most owners either ignore analytics entirely or check one flattering number, feel vaguely good, and close the tab.
Both reactions are rational responses to a bad interface. The fix is not learning all of Google Analytics. It is deciding, in advance, the small set of questions you actually need answered, and ignoring everything else.
I will give you the five questions I set up for every client, where to find each answer, and the popular numbers I tell them to stop looking at.
First, Clear Out the Vanity Metrics
These numbers feel like progress and inform almost nothing:
- Pageviews. Ten thousand pageviews from people who leave in four seconds is worth less than two hundred visits from people ready to buy. Volume without intent is just server load.
- Total sessions, celebrated in isolation. Traffic going up is only good if the traffic is the right kind. A viral post from the wrong country can double your sessions and change nothing.
- Average time on page, sitewide. Averaged across your whole site, this blends bored visitors, interested visitors, and people who left the tab open during lunch. Meaningless as a single number.
- Bounce rate as a scoreboard. A high bounce rate on a blog post can be fine (they read, got the answer, left satisfied). A high bounce rate on your pricing page is a fire. The number needs context or it is noise.
None of these are useless in every context. All of them are useless as headlines. Money metrics look different.
The Five Questions That Are Actually About Money
1. How many visitors did what I wanted them to do?
This is conversion, and it is the only sitewide headline worth having. First, define the action: a purchase, a WhatsApp click, a quote form submitted, a call button tapped. Then make Google Analytics count it, using Goals (or Events if you have already moved to GA4).
Most Indonesian SME sites convert through WhatsApp, and here is the practical trick: track clicks on your WhatsApp link as a goal. If you skip this, your analytics will show traffic but zero outcomes, and every other report becomes uninterpretable. A service business getting 2,000 visits and 60 WhatsApp clicks a month has a 3 percent conversion rate and, finally, a number that connects the website to revenue.
2. Which traffic source sends buyers, not just visitors?
Look at Acquisition, then Source/Medium, but read it with conversions as the column that matters, not sessions. This report regularly reverses marketing decisions. A real pattern I keep encountering: Instagram sends 3,000 visits a month with a 0.4 percent conversion rate, while Google search sends 600 visits converting at 4 percent. The owner was about to increase Instagram spend because "it brings the most traffic." The search channel was quietly worth more than double, on a fifth of the volume.
Source quality, not source volume, is the decision-grade number.
3. Where exactly do people give up?
Every site has a path: landing page, product or service page, contact or checkout, done. The Behavior Flow report, or a simple funnel if you have goals configured, shows where the path leaks. The pattern to hunt is a single step with an outsized drop.
On an e-commerce project I reviewed, 60 percent of visitors who reached the cart abandoned on the shipping-cost step. That is not an analytics insight, that is a business instruction: show shipping costs earlier. Two changes later, checkout completion went from 41 to 63 percent, with zero additional traffic spend. Drop-off points are the cheapest revenue you will ever find, because the visitors are already there.
4. What is mobile doing to me?
Segment your conversion rate by device. In Indonesia, mobile is typically 70 to 85 percent of traffic, and it is routinely the worse-converting segment because the site was designed and tested on a desktop. If desktop converts at 4 percent and mobile at 1 percent, you do not have a marketing problem, you have a mobile usability problem, and it is silently taxing the majority of your visitors. This one segment check catches more expensive defects than any other, and those defects usually trace back to how the site was built in the first place, a pattern I described in The Cheap Website That Becomes Your Most Expensive Mistake.
5. Do people who found what they wanted come back?
For content sites and repeat-purchase businesses, check returning visitor share and, better, whether returning visitors convert at a higher rate. If nobody ever returns, your site is a billboard. Fine for some businesses, fatal for others. This number tells you whether investing in a newsletter, a promo broadcast list, or a retargeting audience has any foundation to build on.
A 20-Minute Monthly Routine
You do not need a dashboard subscription. Once the goals are configured, this is the whole practice:
- Open Analytics, set the date range to last month, compare to the previous month.
- Write down five numbers: conversions, conversion rate, top converting source, worst drop-off step, mobile vs desktop conversion.
- Ask one question: what is the single cheapest change suggested by these numbers?
- Make that change. Check the same five numbers next month.
That is it. Twenty minutes, once a month, with a bias toward one action rather than more charts. It slots naturally into a broader periodic review, the kind I outlined in Mid-Year Check: Is Your Digital Strategy Actually Working?.
One Housekeeping Warning: GA4 Is Coming
If you are on Universal Analytics, the version most sites installed years ago, take note: Google has announced it will stop processing data in Universal Analytics in July 2023 and is moving everyone to Google Analytics 4. GA4 is a different tool with a different data model, and crucially, it does not inherit your history.
You do not need to master GA4 today. You do need to install it in parallel now, alongside your existing setup, so it starts accumulating data. Every month you wait is a month of year-over-year comparison you will not have in 2023. It is a 30-minute task for whoever manages your site; put it on their list this month.
The Takeaway
Website analytics for business comes down to five questions: how many visitors convert, which sources send buyers, where people give up, how badly mobile underperforms, and whether anyone comes back. Configure a goal for your real conversion action this week, especially the WhatsApp click if that is how customers reach you, then run the 20-minute review monthly and act on one finding at a time.
Every number in this article is a proxy for the same underlying question: is the website earning its keep? If you have never been able to answer that, the problem was never that you lacked data. You were just looking at the wrong charts. And if configuring this properly is beyond whoever built your site, that is the kind of thing I help with.