Contents6
All three tools are free and installed almost everywhere. Yet they rarely produce decisions: people look at the curves, note that traffic is going up or down, and move on. The problem is not the tool, it is the lack of a method between the number and the action. This guide complements our guide to marketing automation for SMEs, on the measurement side.
In short
Tag Manager sets up the tracking, GA4 shows behaviour, Search Console shows arrivals from Google.
Check the tracking before interpreting it: a broken tag manufactures false trends.
A useful finding cross-checks at least two sources.
Every recommendation carries an action, an owner and its measure.
Who measures what
What each tool sees
Tag Manager: which tags fire, on which action
GA4: visits, pages, key events (form sent, call, quote)
Search Console: queries, impressions, clicks and position on Google
What none of them sees alone
Whether the contact became a customer (that is the CRM)
Why a visitor left
Visitors who refused cookies, depending on your consent mode settings
That last point changes how the numbers read. It is explained in Consent Mode: conversion tracking explained without jargon.
The four-step method
- Check.
Are key events coming through? A test in Tag Manager preview mode takes ten minutes and saves you from commenting on a wrong number.
- Cross-check.
A gap in GA4 is often explained by Search Console, by a site update or by a campaign.
- Decide.
One action per finding, with an owner.
- Measure.
The date of the change, and the same metric four weeks later.
Three typical recommendations
Each example follows the same format: the finding, the likely cause, the action, the measure.
Traffic goes up, quote requests go down
- Finding: in GA4, visits to the contact page increase, the “quote request” event decreases. It is often the
generate_leadevent, which Google recommends for a form submission. - Likely cause: in Tag Manager preview mode, the tag no longer fires since the form was changed during a site update. Requests still arrive by email, but they are no longer counted.
- Action: fix the trigger, and add a check to every site update.
- Measure: the number of requests counted in GA4 matches the number received by email or in the CRM.
A page seen a lot on Google, rarely clicked
- Finding: in Search Console, a service page has many impressions and a low click-through rate, on the first page.
- Likely cause: its title and description do not repeat the main query.
- Action: rewrite the title and description, and put the answer at the top of the page.
- Measure: the page’s click-through rate on the same query, four weeks before and after.
A page that attracts, but does not convert
- Finding: in GA4, a blog article receives a large share of visits and good engagement time, but no request comes from it.
- Likely cause: the article offers the reader no next step.
- Action: a link to the matching offer at the end of a section, and downloadable content related to the topic.
- Measure: clicks to the offer and requests coming from the page.
Key takeaway
The first recommendation of a tracking audit is often the same: fix the tracking. As long as it is wrong, all the others rest on nothing.
Frequently asked questions
GA4 and Search Console give different numbers: which one to believe?
Both, because they do not measure the same thing. Search Console counts clicks from Google, before the visitor reaches the site. GA4 counts visits, after the page loads and depending on the consent given. A stable gap is not a problem; a gap that changes suddenly is one.
Do you still need Google Tag Manager?
It remains the simplest way to manage several tags without touching the site’s code each time. On a small site with a single tag, it is not essential. As soon as there is advertising and several conversions to track, it saves time.
How many metrics should a monthly report contain?
Few: the ones that lead to a decision. Requests received, their source, the pages that produce them and two or three warning signals are often enough. The rest can be looked up when a question comes up. A report of thirty charts is never read to the end, and the decision gets lost in it.
From monthly report to decisions
A useful report fits on one page: what changed, why, what we are doing, what we measured from the previous month. The rest can be consulted on demand. Putting it together can be automated, as we describe in Automated marketing reporting: tracking what pays; the judgement stays human. To trace back to the signed sale, you need to connect these tools to the CRM: Marketing attribution: which channel really sold.
What we can take on
- Setting up or fixing Tag Manager, Google Analytics and Search Console, with consent: AI Growth Foundation.
- A single monthly report, from search to sale, with its recommendations: digital marketing.
- Rank tracking on Google and in AI answers: rank tracking.
Sources
- Google Analytics, recommended events (new tab)support.google.com
Jim · founder of Brain Plus AI
Also in digital marketing
Practical guide8 min
Automated marketing reporting: track what pays off
A useful marketing report does not copy out figures: every week it says what each channel earned, straight from the tools and unedited. The figures to track, how to connect Analytics, Search Console and your CRM, what AI can comment on, and the data pitfalls to know before you start.
Understand8 min
Consent Mode: conversion tracking explained without jargon
When a visitor declines cookies, your campaigns lose part of what they measure. Google's Consent Mode, and its Microsoft equivalent, pass that choice on to the tags and estimate some of the lost conversions: what it does, what it does not, and what its estimates are worth.
Understand7 min
Marketing attribution: which channel really sold
A B2B sale goes through several channels over several weeks, and every tool claims it. Marketing attribution shares the credit between them; in B2B, it holds when each contact's source is recorded in the CRM and tied to the signed sale.