And what Medium analytics revealed about content discovery in the GEO era

For a while, I thought the problem was simple.

My Medium articles weren’t getting enough views.

Like many writers, the first instinct was to question the obvious things. Maybe the topics weren’t interesting enough. Maybe the writing wasn’t engaging. Maybe the articles were too technical, or too long.

But when I started digging into Medium’s analytics more carefully, something surprising emerged.

The problem wasn’t happening after readers clicked.

The problem was happening before anyone even saw the article.

Our articles weren’t failing at engagement.

They were failing at discovery.

The Metric That Changed the Way I Looked at Content

Most people who look at Medium analytics focus on views, reads, and engagement.

But there is another metric that turned out to be far more revealing: presentations.

A presentation simply means that Medium showed your article to a reader somewhere on the platform — in the feed, recommendations, or topic pages.

In other words, presentations measure whether the platform decided to surface your content at all.

And when I started comparing multiple articles side by side, a pattern became obvious.

Some articles had thousands of presentations.

Others barely had a few hundred.

Once that happened, the rest of the outcome was almost predetermined.

Low presentations meant low views.
Low views naturally meant lower reads.

The bottleneck wasn’t engagement.

The bottleneck was getting discovered in the first place.

The Real Problem Wasn’t CTR or Read Ratio

Interestingly, once readers actually opened the articles, the engagement numbers were quite healthy.

Click-through rates were reasonable.
Read ratios were solid.
Readers who landed on the article generally stayed.

This meant the content itself was not the core issue.

The problem occurred earlier in the funnel.

If an article received very few presentations, it never had the opportunity to generate meaningful views — regardless of how good the content was.

In other words:

Good engagement cannot compensate for poor discovery.

This realization shifted the entire way I started thinking about content performance.

The Discovery Funnel Hidden in Medium Analytics

Once you look at Medium metrics through this lens, they begin to resemble a simple discovery funnel.

Stage 1 — Selection (Presentations)
The platform decides whether your article should be surfaced to readers.

Stage 2 — Intent Match (CTR)
Readers decide whether the article looks relevant enough to click.

Stage 3 — Answer Quality (Reads / Read Ratio)
Readers decide whether the article actually delivers value.

When we mapped our articles across these stages, the pattern became clear.

Most of the variation in performance was happening at Stage 1: Selection.

The articles simply weren’t being surfaced frequently enough.

Which meant they never had the chance to accumulate views.

Why Some Articles Were Being Discovered

When we compared higher-performing articles with lower-performing ones, the difference wasn’t writing quality.

Instead, the strongest articles shared a few structural traits.

They communicated the problem clearly.
Their titles immediately suggested the outcome a reader would achieve.
The opening paragraphs quickly established why the article mattered.

In short, both the platform and the reader could quickly understand what question the article was answering.

This clarity appeared to increase the platform’s confidence in distributing the article.

More presentations led to more opportunities for readers to discover it.

And that translated directly into higher views.

Why Some Articles Struggled to Get Discovered

Other articles told a different story.

Some began with abstract concepts rather than clear problems.

Some titles sounded interesting but did not immediately reveal what specific issue the article was solving.

Others required readers to read deeper before understanding the value of the content.

For a human reader, this may not feel like a major barrier.

But for discovery systems — which must decide quickly whether content is relevant — these signals matter a lot.

If the platform cannot confidently categorize the article, it is less likely to surface it widely.

And when presentations drop, everything else downstream drops with it.

This is how an article with strong insights can still struggle to gain visibility.

The Discovery Problem Is Bigger Than Medium

While this analysis started with Medium analytics, the pattern reflects a broader shift happening across the internet.

Search engines and generative AI systems increasingly sit between users and content.

Instead of browsing links, users often receive synthesized answers directly.

The discovery journey is evolving from:

Search → Links → Article

to something closer to:

Question → System selects content → Answer → Optional click

In this environment, content competes not just for attention — but for selection.

This shift is often described as Generative Engine Optimization (GEO).

What GEO Means for Content Discovery

In a GEO-driven discovery system, platforms and AI models need to quickly determine:

• What question a piece of content answers

• Whether it can confidently satisfy user intent

• Whether it contains structured, extractable information

Content that signals these things clearly has a much higher chance of being surfaced.

Content that hides the problem behind abstract framing often struggles to be discovered — even if the insights are strong.

In other words, discovery systems reward clarity of intent.

The GEO Content Pyramid

One helpful way to think about GEO-ready content is through four layers.

Query Hook

The title and opening clearly signal the question being answered.

Instant Answer

The reader quickly understands the value of the article.

Modular Knowledge

Information is organized into sections that can stand alone as explanations.

Trust Layer

Real-world examples and practical insights demonstrate credibility.

When these layers align, content becomes easier for systems to understand and distribute.

What This Means for Content Writers

The key lesson from this analysis was surprisingly simple.

Our articles were not failing because readers didn’t like them.

They were failing because readers never saw them in the first place.

Views are often treated as the primary success metric for content.

But views are actually a downstream outcome.

Before an article can earn views, it must first pass the discovery filter.

It must be selected, surfaced, and shown to readers.

Only then does engagement begin to matter.

Final Thought

For a long time, content teams have focused on writing better articles.

But in the GEO era, writing quality alone is not enough.

Articles must also make it immediately obvious:

what problem they solve.

Because in modern discovery systems, content does not only compete for attention.

It competes for something even more fundamental.

The chance to be discovered at all.