Indie Dev SEO Handbook - Intermediate (4): Growth Data Compass

Preface: Do You Really “Know” How to Read Data?

The SEO Handbook series has been on hiatus for a while, but I still want to finish this series completely. I’ve been pretty busy lately and haven’t updated much, but luckily I had the whole SEO Handbook outline ready early. So, let’s begin.

First, a sharp question: You open GA4 every day, staring at that rising and falling total traffic curve — besides feeling anxious or pleased, do you know what you should actually do next?

If you can’t answer, that’s normal. Because only watching “total traffic” is typical rookie behavior — it’s a “vanity metric” that can’t guide action.

Truly valuable data insights lurk in GSC (Google Search Console) and GA4’s detailed reports — they watch “process metrics” that directly guide action.

This article is my custom-made “data compass” user guide for indie developers. We’ll learn how to shift from “watching for fun” to “watching for insight.”

Chapter 1: The “Two Faces” of the Data Compass: GSC vs. GA4

Before analyzing, we must clarify: a complete SEO data compass has “two faces” — you must use GSC and GA4 together, otherwise you’re “blind men touching an elephant.”

GSC (Opportunity Compass):

Tells you what happened “before” the click. It reveals Google’s “view” of you (ranking, impressions) and your “market opportunities” (what users are searching).

GA4 (Experience Compass):

Tells you what users did “after” the click. It reveals whether your pages truly satisfied user “intent” (dwell time, engagement rate).

In GSC, you can pull data with landing page as the core dimension, including impressions, clicks, which keywords users discovered and clicked into your page, whether this matches your target goals, and which keywords outside your plan still led users to find your page. These keyword signals can be Key Points for continued page optimization.

GA4 shows more granular page data, and through custom events, you can understand exactly what users did on your page, plus various UX metrics. This is extremely helpful for CRO work and targeted page experience optimization and improvement, whether it’s a tool page, e-commerce product page, or blog article page.

My core view: GSC helps you “bring people in,” GA4 tells you “if guests are satisfied.” Healthy growth must address both.

Chapter 2: GSC Metrics Deep Dive - Mining Your “Growth Gold”

GSC is our “opportunity compass.” We’ll focus on its four most important metrics: Impressions, CTR (Click-Through Rate), Position (Average Ranking), and Queries (Keywords).

We don’t look at them in isolation — we look at their combinations, which tell us different “scenarios”:

Scenario 1: High Impressions + Low CTR

Diagnosis: Your “storefront” lost. Google considers your content relevant to the topic (giving you lots of impressions), but your Title and Meta Description on SERP have zero appeal — users clicked your competitors.

Action Map: Immediately rewrite your Title and Meta Description. Study top-ranking competitor titles, think about how to write more pain-point-hitting and differentiated copy. You can reference competitor situations — doesn’t mean copy exactly, but analyze what makes theirs more effective. Also, due to AI Overview’s high coverage now, this actually hugely impacts SEO clicks. If you find impressions growing but clicks declining, consider this factor too, and find ways to get your page into AI Overview as well — this can somewhat improve your GEO effectiveness.

Scenario 2: High CTR + Low Ranking (Positions 5-20)

Diagnosis: Your “product” is treasure! This (average position 5-20) is SEO’s most important “opportunity zone.” Data shows users really like your title (high CTR), but Google only puts you on page 2 or page 1 bottom because you lack authority (backlinks) or content depth.

Action Map: This is your “growth gold mine.” Immediately list these pages as highest priority, allocate resources: 1. Add quality internal links; 2. Launch link building campaign (guest blogs, PR, etc.); 3. Update articles, add more “Information Gain.” 4. Definitely compare with competitor pages ranking above you, understand “where exactly you lost” rather than blindly executing optimization or stacking backlinks.

Scenario 3: Discovering “Content Gaps” (High Impression Queries)

Diagnosis: Analyzing a page’s Queries, you discover it gets massive impressions for “how to connect AI tool A to B,” but your article body doesn’t answer this question in detail at all.

Action Map: This is content inspiration users are “feeding” you through GSC. Immediately add an H2 section to the original answering this “content gap,” or write a new standalone article to address this need.

Chapter 3: GA4 Metrics Decoded - Insight into Google’s “Satisfaction” Algorithm

GSC helped us “bring people in,” now it’s GA4 “experience compass’s” turn, telling us “if guests are satisfied.” Google’s “Helpful Content Update” core is rewarding pages that satisfy users.

We focus on three metrics:

Avg. Engagement Time:

This is the gold standard for judging “Information Gain.” Are users spending enough time on your page to read? If your 3000-word masterpiece has 10-second average engagement, your content simply isn’t retaining people. Especially important is first and second screen content, also considering mobile vs. desktop experience. Mobile has high volume, deeply affecting your SEO; desktop usually has higher traffic value for most products.

Engagement Rate (Engaged Sessions Percentage):

This is GA4’s replacement for “bounce rate.” An “engaged session” is one lasting over 10 seconds, or having conversion, or having 2+ page views. Higher percentage means your page “hooks” users better.

Bounce Rate:

Though hidden, it still has value. But you must view it dialectically: A “what is GEO” definition page where users read for 30 seconds then leave might be “satisfied bounce” — they found their answer. A landing page where users leave after 3 seconds — that’s “disaster bounce.”

Action Map: When you find a page with great GSC data but poor GA4 experience metrics (e.g., extremely short engagement time), you must immediately alert. This means your content quality can’t hold the traffic — Google’s algorithm will eventually discover and demote you.

Chapter 4: “Relativity” — Why Your Data Has No “Absolute Standard”

When analyzing data, the most common rookie mistake is asking everywhere: “Mr. Guo, my bounce rate is 60%, is that high?” This is an unanswerable question.

You must establish a data “relativity” mindset: No absolutely good data, only relatively good data.

Context Determines Standard:

Your “blog article page” with 5-minute average engagement and 80% bounce rate might be perfectly healthy data. Your “product pricing page” with 20-second average engagement and 80% bounce rate might be a disaster. You must establish different evaluation standards for different page types.

Relative Benchmarks:

How do you know if your “blog article” 5-minute engagement is good or bad? You need a benchmark. Here you can use tools like Similarweb or Semrush to check your vertical competitors’ “average time on site” and “bounce rate” data.

Action Map: Understanding that Similarweb and similar tools’ data isn’t “perfectly accurate,” use it for “directional” reference. If your competitor’s average time on site is 3 minutes and yours is only 1 minute, you know your content depth or readability definitely has big problems.

Conclusion & Next Preview: From “Manual Analysis” to “AI Empowerment”

In this article, we established a “dual compass” thinking framework. You should now deeply understand how GSC’s “opportunity metrics” and GA4’s “experience metrics” work together to point toward growth direction.

But when your site has hundreds of pages and thousands of keywords, manually executing GSC+GA4 “joint diagnosis” becomes a new bottleneck.

Therefore, in the next article, we’ll officially enter “Vibe Coding” practice: learning how to use Claude Code or Codex to replace tedious Excel, “conversing with data” to automatically mine all growth insights from GSC and GA4 raw data. I’ve always felt Codex and Claude Code are more like “general-purpose agents” — after all, even cleaning computer storage can be done with them.