Tag Archive for: AI

In SEO, we often operate under the assumption that more is better: more content, more pages, more keywords, more traffic. 

The expectation is that more of these elements will lead to more revenue.

With generative AI, this belief has only grown stronger, enabling content production at an unprecedented scale. But what if more isn’t the answer?

The pursuit of maximalism has led to significant challenges.

Google now penalizes excessive, bloated content, and what once seemed like a straightforward strategy is proving detrimental.

This isn’t just an algorithmic issue – users are overwhelmed, conversions are declining, and trust in brands is eroding. 

We’re creating content faster than ever, yet the results are diminishing. In many cases, we’re doing far more than necessary. The solution doesn’t lie in adding more but in doing less.

This brings us to an ancient principle, via negativa (the “negative way” or “way of negation”), rooted in Neo-Platonic philosophy and medieval theology. 

Nassim Taleb highlights that what we know can be disproven, but negative knowledge – knowing what doesn’t work – is enduring.

To truly understand what works, we must first identify what doesn’t. This process of inversion is critical.

This article will challenge the notion that more is always better in SEO. We’ll explore:

• Why via negativa works.
• The pitfalls of excess.
• How to regain control if your content strategy has gone too far. 

Real-world examples will illustrate how less can be more.

Subtraction in action: Why it works (and why we resist it)

Brands have progressively minimized their logos to leave only what is essential:

This resembles Picasso’s “Le Taureau,” a famous simplification of a bull’s drawing:

“Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away.”

– Antoine de Saint-Exupéry

Digital technology has made our lives more abundant, much like Picasso adding detail to enrich the image of a bull. 

But at some point, there’s too much, leading to distraction from what truly matters.

Consider cooking as another example. Japanese cuisine is known for its simplicity, removing unnecessary elements to highlight original flavors.

When reducing a sauce, the goal is balance – neither too thick nor too thin, but just right.

What about SEO and content marketing?

• Great information architecture and navigation are simple and intuitive, aligning with the principle of “Don’t make me think.”
• Great content is as long as necessary and as short as possible to (increase information density and) emphasize each part instead of diluting it.
• Successful domains have only as many URLs as needed to attract the right traffic, not the most.

So why do we still believe more is better and struggle to do less?

Fortunately, scientific studies answer this question. Here are the core ideas to explain our shortcomings:

• We add more because it’s easier to identify what’s visible and ignore the rest (unlike mountains of paper in an office – an example from Gerry McGovern’s “World Wide Waste”).
• We don’t think of subtraction easily, as it requires more cognitive effort.
• We default to instinctive solutions, creating a self-reinforcing cycle because the “idea to add” becomes more cognitively accessible if more people discuss it.

Neglecting something beneficial in the long run can cause long-term damage. 

In SEO, this damage manifests in distinct ways that compound over time if unchecked.

The problems of addition over subtraction in SEO and how to solve them

Maximalism in SEO presents three core problems:

• An insatiable appetite for more is a red flag.
• Google no longer tolerates mediocre content.
• Overemphasis on SEO gains leads to imbalance.

Let’s examine each of these in turn.

Controlling our appetite 

A great example of our hunger for more being a problem is programmatic SEO and AI content. 

By AI content, I mean pushing out articles like McDonald’s does to burgers – average at best and not suited for a gourmet.

Google’s John Mueller said regarding programmatic SEO:

• “I love fire, but also programmatic SEO is often a fancy banner for spam.”

We often see Google as a mountain to climb, with generative AI or programmatic SEO acting as a performance-enhancing drug. 

However, Google is not just a mountain; it’s a volcano. 

By focusing too much on reaching the “summit,” we overlook warning signs and take unnecessary risks, which ultimately lead to our downfall.

What are the risks?

• An unnatural surge of new pages within a short time frame.
• Too many new pages relative to the rest of your site.
• Ignoring advice from Google spokespeople.

I know this is controversial, but there are indications that rapid, excessive actions may not align with Google’s interests (right now). 

For instance, the Search Quality Rater Guidelines suggest that perceived effort plays a role in assessing content quality.

We often have an insatiable appetite. Being too hungry and not knowing when enough is enough creates problems.

“’Enough’ is realizing that the opposite – an insatiable appetite for more – will push you to the point of regret. The only way to know how much food you can eat is to eat until you’re sick. Few try this because vomiting hurts more than any meal is good.”

– Morgan Housel

• Takeaway: Anyone can quickly create low-quality content, but few can create extraordinary content. Focus on creating content that is 10 times better, not 10 times more.

Dig deeper: 5 SEO mistakes sacrificing quantity and quality (and how to fix them)

Serving Google our best meal since they don’t want our junk food

Three changes in the last two years dramatically changed our playing field:

• Thresholds for brand signals increased.
• Indexing became harder.
• Google developed technical “allergies.”

There’s no direct proof for the first point (yet), however, there are two pieces of evidence:

• Tom Capper theorizes that “helpful content” is a smoke screen. Google increased the threshold for brand signals. High domain authority, but low brand authority is a red flag (and helps to sort the cesspool of primarily SEO-supported domains).
• Mark Williams-Cook points out that site quality and predicted site quality must have been all over the place with the influx of AI content, so Google changed this.

Some domains experienced short-term success but were filled with poor-quality content, lacking oversight and nutritional value.

If we break the system or find loopholes, Google behaves like the government. 

There may be short-term arbitrage opportunities. (Emphasis on short-term as neither the government nor Google appreciate attempts to deceive them.)

By ignoring good advice for too long, we tend to fly too close to the sun and eventually burn. Here’s a great example of a domain that didn’t listen:

We often hear impressive stories like, “I created 16,824,973 articles with just a few clicks.”

It sounds great on paper, but we rarely hear about the failures.

It’s nicer to talk about how well we did versus how we miserably failed and destroyed our brand reputation.

This is known as survivorship bias:

We only hear about the planes that survived the war. 

The solution we often propose is to fix the most-hit spots. What we miss is that the planes that didn’t return were hit in areas where survivors weren’t.

Here’s a Claude artifact for you to simulate.

Dig deeper: Why SEO experts rarely share actual success stories

As I pointed out in “SEO grew up, a lot of SEOs didn’t,” indexing is becoming harder:

Google doesn’t have unlimited capacity. According to what we know, the Google index is static (about 400 billion documents).

Gone are the days when Google seemed like a ravenous beast, eager to consume everything.

With AI, we produce more stuff, so the quality bar for content to be indexed has to increase.

Technical SEO doesn’t directly grow your site, but for larger sites, mistakes can have serious consequences.

In the last two years, Google has become increasingly finicky.

Here’s an anonymized example of a large client domain with multiple indexing issues related to unintended parameter pages:

We don’t want to put Google into anaphylactic shock.

Pushing out more (bad or mediocre) content quickly might sound good, but it’s likely not what Google wants. It’s as if Google has new nutritional standards that we’re reluctant to follow.

Takeaways:

• Focus on building your brand. Start by reading “How Brands Grow” by Byron Sharp.
• Only serve Google your best dishes (content) and regularly check your menu (URL portfolio).
• Ensure a solid technical foundation – it’s like the basics in cooking. For large sites, monitoring is a must-have, not a nice-to-have.

Avoiding imbalance caused by overfocusing on SEO gains

Another thing we forget: SEO isn’t everything. One-sided nutrition neither works in real life nor in SEO.

We used to commission our content just like competitors, but with a 10% premium (known as the skyscraper technique). 

Some still think holistic pages are the way to go. Let’s assume this was still the case and ask some questions:

• What are the side effects of adding more content to an article?
• Do these articles convert better than they did before?
• How much better do you actually rank?

Common side effects include:

• A decrease in conversion rate as the article is meandering off path.
• Diluted brand perception due to a focus on “more” rather than “better.”
• Users are getting lost and not finding what they came for.

Searchers don’t want to devour “Infinite Jest” (which is 1,088 pages long) like articles all the time.

I like to follow the formula of Stephen King, which is that the second draft is the first draft, but you get rid of 10%.

Google’s Mueller recently had some nice words on this topic as well:

• “If you count the words in bestseller books, average the count, and then write the same number of words in your own book, will it become a bestseller? If you make a phone that has the same dimensions as a popular smartphone, will you sell as many as they do? I love spreadsheets, but numbers aren’t everything.”

A guide on “how to create X” doesn’t need an explanation of what X is, because people looking for this already know X. If they didn’t, they wouldn’t actively be searching for it.

This is a classic SEO and content marketing armchair problem I’ve faced many times. I’m not exempt, either. 

Early in my career, I told a B2B client to write all these articles. Looking back, such articles are often alienated from the actual target audience.

If you rank lower, you must compensate for the loss of traffic in conversions to maintain the same financial outcome.

When you sort a list in decreasing order (like CTR of organic traffic by position), “the value of the n-th entry is often approximately inversely proportional to n,” according to Zipf’s Law.

Basically, it means Position 2 will have half the value of Position 1.

Assuming we have a CTR of 25% for Position 1, Position 2 should have around 12.5%.

If you drop from Position 1 to 2, you have to make up for it by doubling conversions. This sounds like a lot. There’s a caveat, though.

We shouldn’t just look at organic traffic.

An increase in conversion rate means every traffic source should convert better, while more organic traffic is… well, more organic traffic.

Here are a few assumptions and an example:

• Organic search gets 1,000 page views (Position 1), the other channels 2,000 (= 3,000 in total).
• Conversion rate is low at 0.5%, so 5 from organic and 10 from other channels (= 15 in total).
• Making the article better (in spite of SEO), you drop to Position 2; however, your conversion rate increased by 50%.
• Now you get 500 page views from organic search (Position 2) and the conversion rate sits at 0.75%.
• You have less traffic and fewer conversions from organic search now (3-4 instead of 5).
• But, in total, you get 18-19 conversions instead of 15.

The example is simplified. But even if you drop organic conversions, you can still make more money if you consider the bigger picture.

Instead of eating each ingredient one by one (all channels as silos), you combine all flavors into one perfect bite (all channels working as unified clockwork) – the way Japanese cuisine intends.

Takeaways:

• SEO isn’t everything – you want to win, with or despite SEO.
• Better rankings = more traffic ≠ more conversions.
• You can offset worse rankings with improved conversions.

---

What to do when things get out of control

If things get out of control on a domain level, we face two situations:

• Having excess weight.
• Having shed that weight, preventing the relapse.

I suggest the following solutions to become and stay lean long-term.

Getting rid of the fat cells (= content pruning)

To determine if your domain is overweight, I’d check a couple of metrics:

• The % of indexed pages that should get organic traffic but haven’t for 3 months: A value of > 10% is something you should investigate.
• The % of pages with no organic traffic: Anything > 70% would pique my interest.
• The % of pages Google doesn’t want to index: If your sitemaps only contain pages to be indexed, more than 30% in “Crawled – currently not indexed” seems too high.

If you see those red flags, consider a content pruning project. 

This doesn’t mean you should just go out there and delete old stuff or URLs that “aren’t good for SEO.” 

If you just delete things, you might hurt your domain more than you would like:

In this case, content was simply deleted. 

Given that this is a publisher domain reliant on ad revenue, it likely had a significant impact.

Content pruning doesn’t just mean deleting stuff. It can also mean:

• Updating.
• Reworking.
• Combining.

Dig deeper: Improving or removing content for SEO: How to do it the right way

There are many popular examples, like IBM:

Or Progressive (moving pages from a subdirectory to a different domain):

As mentioned earlier, SEO isn’t everything. Here are some example metrics I’d consider.

For SEO:

• Traffic + Impressions.
• Number rankings (in top 3).
• Backlinks.

Contribution to the business model:

• Conversions (micro and macro).
• Revenue.
• Is it an important part of the customer journey?

User behavior (in relation to comparable page types – don’t compare apples to oranges):

• Time on page (relative to estimated reading time)
• Bounce rate
• Indicators of “a good session” (= this will most likely be custom events like sessions with 3 articles read for a publisher)

Editorial quality:

• Has an author?
• Article/H1/title length.
• Number of (original) images/videos.
• Last modified.
• Readability.
• Links to internal and external sources (e.g., to back claims up).

Tip: Create your own compound metrics. A high CTR is good, but doesn’t matter if post-click outcomes are poor. Prefer a combination of CTR and “happy users.”

I define thresholds for all metrics. If met, the URL goes into a whitelist (= “don’t touch yet”). I can’t tell you what good enough is; it depends on your domain. There isn’t a cookie-cutter template.

URLs not on the whitelist get points based on the outlined metrics. 

Then, you look at the total average, and everything below average continues to the next workflow step.

Then, you assign different “deletion criteria.” Here are three examples:

• Less than 250 words + title/H1 with ≤3 words.
• No author + updated over four years ago.
• Both.

This way, you can identify articles most likely to be empty calories, aka terrible content, if they meet all your criteria.

After that, I wouldn’t delete anything yet but look for opportunities to compare it to similar content to find potential redirect targets or ways to improve multiple articles by combining them.

Once your domain is fit again, the biggest mistake is becoming a victim of the jojo effect.

Keeping the weight away means changing the environment in a big way

Bad habits, like creating subpar content, often stem from environmental problems, such as the wrong incentives (e.g., a boss wanting higher traffic numbers) or mindset (e.g,. “SEO is everything”).

Ozempic can help you lose weight, but the weight was a symptom of underlying habits (like overeating and lack of exercise). 

We can’t just treat the symptoms; we must attack the root cause.

Here are examples of issues that lead to content pruning and how to prevent them:

• Editorial departments working in disconnected teams: Often due to organizational structure. While you can’t change this overnight, you can advocate for more collaboration rather than conflict.
• SEO not engaging with other departments: It’s partly your responsibility to initiate conversations. Listen and understand others first, before seeking to be understood or asking for favors. Your goal is to show how you can help them.
• Lack of content management guidelines: Creating content is one thing, but managing it is another. Content management needs clear guidelines, such as how to handle different lifecycle stages. If these don’t exist, explain their value and benefits to others.

Unfortunately, I can’t give you the exact blueprint for every problem. 

In any case, you can be an advocate. 

Avoid “telling others what to do,” which will likely lead to resentment. 

Instead, try to see things from the viewpoint of the people around you.

Content pruning, like any SEO initiative, runs on an empty stomach if you don’t work on the underlying systems. 

The best SEO initiative is the one we don’t need.

There’s a fair argument for content pruning (or Ozempic, if necessary, for faster weight loss). 

However, you must work on the underlying systems, or you’ll repeat this over and over again:

Making the cut: Doing less to achieve more can be a viable strategy

More can be better, but it doesn’t have to be. There’s great power in the art of subtraction. 

Anyone can create a lot of mediocre content and do more at the click of a button. Only a few can reduce and drive more results at the same time.

As Michael Porter puts it, “the essence of strategy is choosing what not to do.” 

You have to make decisions.

What I didn’t know for a long time is that “decide” comes from “decidere” in Latin, meaning “to cut sth. off.” 

Via negativa is a strategic mental model and symbolizes cutting off what you don’t need. So the obsession with quantity as our default option must die.

Understand: More isn’t always better. Less isn’t always worse. Do less but better for your best work.

Read more at Read More