Normally I write these types of posts at each anniversary, but so much has happened that I decided to do this now, order my thoughts, accept what is and move along.
I ended my last news update, the 3-year anniversary, with: “For the long-term I want to get rid of my dependance on Google. It has proven too big a risk for my income, health and everything that comes after those two, like being there for your family and partner.”
What I didn’t know then is that after having been de-indexed 3 times in a single year, a fourth de-indexation, courtesy of the September Helpful Content Update, was just around the corner.
That update took away about 50% of my traffic, and was quickly followed by the October spam update, which resulted in another 85% drop. A slight but temporary uptick happened with the November reviews update, but all in all my site is down to about 5% of its original traffic and hasn’t recovered like it did the previous 3 times.
As I mentioned in my last anniversary update there’s a one-on-one connection between traffic and affiliate sales/income, which means the site is doing around 5% of its financial peak. If I look at the sales numbers in Amazon, there are clicks and sales, but they’re not related to my content most of the time. Meaning I get commissions for non-bicycle related products.
How I got here
It’s September 13th, a day before the update rolls out, everything’s fine and I’m doing my research and blogging the way I’ve done before. How did I get to the point where Google decided it was a good thing to write several algorithms that would ultimately take away my source of income?
This is not my first business initiative that would generate arguably a considerable amount of money. I still have another business, which is a webshop selling WordPress themes and website templates called nexusthemes.com.
I decided to try and do my own thing in 2011. I went from a fulltime job as a software tester to doing that 2 days a week and trying to get something else off the ground for the rest of the week. One of the things I did, and the thing which finally started to make money to the point where I could quit my other job was nexusthemes.com.
At one point I had 10 people working for this company in a freelance capacity. Its success revolved around making a WordPress theme for every single business type that had search volume associated with it. Search volume means people are looking for my products by typing it, or a variation of it, into the Google search engine.
The essence of SEO
With that company I learned a certain recipe for success with respect to SEO or Search Engine Optimization:
- Find key words/phrases with search volume
- Ensure there is no or irrelevant competition in the SERP (Search Engine Results Page)
- Make a relevant webpage for the key phrase
- Have something to sell on the web page so you make money
- Rinse and repeat
“No competition” in step 2 in practice meant you’d need long-tail key phrases. For Nexus Themes it meant we would create 375 WordPress themes.
I stumbled upon this strategy by accident. If you look at the publication date of the first landing pages for WordPress themes we created, you can see that at the end of 2013 I still hadn’t figured out the SEO game.
I basically didn’t know what I was doing marketing-wise but was able to reverse engineer accidental financial success in early 2014, because I had made landing pages for products people were looking for in Google and had no competition.
One of the biggest initial successes were our yoga WordPress themes. And the only reason I made the first one was because a friend from my girlfriend was a yoga teacher and said she needed a website. So I made the theme, but she never used it, but it ended up being one of the key ingredients that showed me the outline of SEO success in 2014, alongside the actor theme, plumber theme, painter, and beautician WordPress.
I also made the nature photographer, personal trainer, dentist, and law firm themes in the same period. Themes which would sell extremely bad, either because no one was looking for them (no search volume for the nature photographer) or the competition was way too fierce (law firm and dentist).
From individual products to product lists
The nature of our business was that we were awarded traffic for landing pages, which held individual products. And to be able to create such a page, we needed to actually invest considerable resources. I think I created the first 100 themes before we decided to hire people to work alongside me. I would supervise the product production workflow, including the business analysis, theme design, and landing page copywriting process, besides doing it myself as well.
I think our peak in terms of traffic and sales was 2 years in, around the end of 2015. We’ve created roughly 250 themes by then, had maybe 5 designers, 2 support people, an accountant, business analyst and me and my partner taking care of everything else.
With a traffic peak in early 2016 we went downhill from there. The core reason was increased competition. And with increased competition, I mean out-of-this world competition. Competition which would push our landing pages out of the search results.
The characteristics of that competition were two-fold: marketplaces with lots of user generated content, and blog sites with list overviews of themes. The latter were pages with a number and the word “best” in them. So where we would have a page called “yoga wordpress theme”, our competition would have a page called “the best 24 yoga wordpress themes for yoga teachers and yoga studios for 2016”.
I researched the competition by looking at every single piece of content they would create in time, and by looking at the Internet Archive, to get a clear picture of the emergence of a certain SEO strategy of these competitors. A lot of these sites started around 2014, but figured out what type of page the Google algorithm rewarded with traffic and went all in, just as we had done on our way in 2014.
And because it’s much less work to create a blog post of a list of themes than it is to create an actual theme and then a landing page for that theme, we would be overtaken left and right with dozens of sites making thousands of blog posts that would rank higher.
So the landing page for the yoga wordpress theme that once was one of the seeds of success, would lose almost all of its traffic in a single year, because there were 25 lists of yoga wordpress themes that Google would rank higher, because they were lists.
We tried to emulate this list-strategy, which only hurt our existing sales. We were adding lists to our site that introduced keyword cannibalization, where lists were outperforming our landing pages with individual themes, but weren’t converting to sales at the same time.
The loss in conversion immediately resulted in 25% loss in revenue, and we were hurting our rankings for our existing products with this half-hearted lists initiative.
Although the list strategy still holds up today, the way we did it was only to game Google, get traffic, do sales, and we did a low-quality bad job, hoping it was good enough, which it wasn’t.
There are a number of other things we did that weren’t working for our company at the time, but the main takeaway for us was that posts with lists of products trump posts with an individual product in terms of getting traffic. And I didn’t see any exception to this rule in the SERPs. The only themes we still sold were themes with very low to no competition. If there was a website with a list that would be a multiple of our individual keyword, it would rank higher, no matter how crappy the content of that list page.
The start of my own SEO list strategy
In June of 2020, 6 years after I first got myself acquainted with the SEO list strategy of content creation I’d start both a YouTube channel and a website; this website.
My initial focus lay on the YouTube channel. For each YouTube channel I needed voiceover content. So here’s what I would do
- shoot video content of some bike repair or restoration project
- create a blog with screenshots of the video
- optimize the blog using a relevant key phrase I picked on the fly. The only requirement was that the key phrase was autocompleted by Google
In December 2021 I will turn on Adsense for restoration.bike and YouTube. At that time I had created more than 400 pieces of content, videos and blogs in less than 2 years in my spare time.
On February 22, this resulted in a payout of 171 euros. That’s great, but the ROI of this method of content marketing is super low and does not scale well. So I have answered your question “how are you going to make money with that?”, but not in a way that will really make a difference.
When you look at the analytics data of February 2022, the month I would start with affiliate marketing, it’s very clear that the SEO list strategy is just as relevant as it was for the WordPress theme business.
- 5 out 10 articles are lists of products, notably tires, which would serve as the template for my affiliate marketing in the months afterwards
- There’s a single product review, a single repair tutorial
- A case can be made that the other 3 articles (catalogs of bicycles) are also listed. All three of them offer lists of bicycles (GT, Trek, and Cannondale). The final conclusion would then be that 8 out of 10 of the top-performing blogs from a traffic standpoint are lists.
I didn’t need proof to know what was extremely obvious, and what would work marketing-wise from there on out. I would start testing the combination of product list articles and affiliate marketing of those products.
On February 13 2022, the first day of sales will be received on Amazon via affiliate marketing buttons that are placed by hand in existing articles (think: slick mountain bike tires).
On February 20, I see that I am putting a new blog live for the first time on restoration.bike, purely aimed at further testing affiliate marketing. That blog is a set of a total of 10 handmade articles about bicycle tires to test affiliate marketing on Amazon and Ebay.
On March 5, I will put the first tire blog live with the first version of a custom made injector plugin, which is a piece of software that creates a WordPress post based on content located in a Google spreadsheet. The Google spreadsheet is where I do all of my research about products.
In the following table you can see the date in relation to the number of unique pageviews and how many blogs I had on my site. So I started around 160 and created around 350 affiliate marketing posts in less than 2 years. This resulted in the traffic doubling from 30K to around 80K of unique pageviews in August of 2023, the month before the helpful content update.
blogs on site
All of those 350 affiliate marketing pages were lists of products. I didn’t just write them by hand. Across 23 different categories I listed around 3700 products and their specifications, and wrote hundreds of individual product reviews. I combined these product reviews into lists, added the relevant product images to them, wrote an intro text, and added a table product specifications.
I integrated my blog content into my site using custom fields and a template I built, which knew which product content should go where. In other words, I wasn’t creating every single WordPress post by hand using the block editor in the WordPress backend; it was a highly streamlined content creation process.
Getting Killed by Google
2023 was the big year of AI, with ChatGPT launching in March of that year. It was also the year my site got shut down 3 times by Google, which would roll out a record number of updates to their search algorithm. The third shut down has up to this point relegated my content creation efforts to the sidelines.
What could have been something that would see me making more money than I’ve ever earned in my life on a monthly basis, was effectively neutered as soon as I reached that goal I’d set for myself. I can’t know for sure, but missing out on the far majority of affiliate sales and its accompanying flat fee has set me back between 4000 to 5000 dollars per month. A number I reached after just over 3 years, which is kind of standard for blog income.
If a loss in traffic happened to you in this time period, as it did to me, Google has the following to say you:
Google Search’s helpful content system generates a signal used by our automated ranking systems to better ensure people see original, helpful content created for people in search results. […] The helpful content system aims to better reward content where visitors feel they’ve had a satisfying experience, while content that doesn’t meet a visitor’s expectations won’t perform as well. […] Any content—not just unhelpful content—on sites determined to have relatively high amounts of unhelpful content overall is less likely to perform well in SearchGoogle Search’s helpful content system and your website
- Content needs to be original
- Content needs to be helpful
- Content needs to meet visitor’s expectations
- All of your content (helpful and unhelpful) will receive less traffic, relative to the amount of unhelpful content your website displays.
That’s not saying much, especially not considering I based my content strategy on what people indicated they wanted to see from me (products lists), and I firmly believe I met visitor’s expectations based on click-through and conversion ratios, which were 30% and 9% respectively. Basically an absurd percentage was convinced of what I had to say and a huge number decided to buy something on Amazon because of it.
The decision to flag my site as being unhelpful and spammy wasn’t based on a solid review of my content but based on the fact I did an affiliate marketing strategy at scale.
An affiliate marketing strategy which helped the competition of my other company push me out of the market years before. A content marketing strategy that up to that point was the best strategy to attract traffic.
Besides the succinct helpful content update explanation, there’s also the Google E-E-A-T guidelines of creating content. If you read those guidelines they make sense for anyone not making a living or trying to make a living as a blogger. That includes SEO people. But simply looking at the Analytics data of blogs being successful in attracting traffic, paints an entirely different picture.
The reason is the current E-E-A-T guidelines were never the best practices of getting traffic. At least not up to the September update. The traffic numbers I’ve shown you and the articles which created those numbers are not in line with those guidelines.
I did create those E-E-A-T articles; they were my repair, maintenance and restoration tutorials. Those articles showed videos of my YouTube channel, photos of me doing my thing with bicycles, and content written entirely by hand. By using Google Ads I was able to make less than €200 euros per month with over 30.000 visitors. I made 20 times as much doing affiliate marketing the way I did, doing almost a million in sales within a year. And it all relied on the fact I received more traffic with those blog posts.
Google did a 180 and completely changed the rules of the game overnight, calling it E-E-A-T and helpful content. The company stating they want content written by people for people, used AI and machine learning to flag my site as being spammy, take away a huge part of my income, which took me over 3 years to build up. Yes, I’m pissed off and think Google and everybody else saying it’s my fault are nothing less than hypocrites.
Is the internet a better place without my content?
So Google set out to achieve a certain thing, which is arguably to make the internet a better place where people using their search engine are better served without my content taking up space in the Google SERP.
So what have I been replaced with? And is that content more helpful, whatever that might mean?
The first example is one I’ve already showed. The best puncture resistant road bike tires.
According to serp.api there isn’t an exact match for this key phrase.
The number one spot talks in general about puncture proof tires and also lists non-road bike tires. ChatGPT at least provides a majority of road bike tires. All of the listed tires would serve you. It’s a good list as far as I can tell.
Continental Grand Prix 4-Season
Continental Gatorskin BE
Pirelli Cinturato Velo
Pirelli Cinturato Road
Schwalbe Durano Plus
Schwalbe Marathon Plus road tyre
Specialized Nimbus Armadillo Reflect Puncture-Proof Tyre
Continental CONTACT Speed City puncture-proof tyre
Michelin Protek Max City Puncture-proof Tyre
Bontrager Hard Case Ultimate Reflective puncture-proof tyres
Schwalbe Durano S
Continental Gatorskin tyres
Vittoria Zaffiro Pro road tyre
Schwalbe Marathon Original
Schwalbe Marathon Plus
Michelin Power Endurance
Vittoria Rubino Pro Control G2.0
Pirelli P Zero Velo 4S: Pirelli P Zero Velo 4S
Next one is the best cycling gloves for hand numbness. Number one is a reddit post. It doesn’t list gloves, but talks about ulnar-nerve syndrome (hand numbness).
Number two is a non-exact match key phrase, talking about cycling gloves in general (best cycling gloves). The entire article doesn’t talk about hand numbness and or its related ulnar nerve issues. It’s no wonder its list is totally different from mine.
ChatGPT again seems to do a pretty good job, with multiple options similar to mine. Bontrager and Specialized are brands I would never consider, because they are not sold on Amazon. So I could never make money with them. So they could be perfect options to prevent hand numbness as far as I’m concerned.
Castelli Roubaix Gel 2 Gloves
Castelli Rosso Corsa Espresso Gloves
Castelli Rosso Corsa 2 W Gloves
Castelli Endurance Gloves
Castelli Arenberg Gel 2 Gloves
Pearl Izumi Select Gloves
Pearl Izumi Women’s Select Gloves
Giro Supernatural Gloves
Giro Women’s Supernatural Gloves
Giro LX Gloves
Giro Monaco II Gel Gloves
Giro Women’s Monica II Gel Gloves
Giro Women’s Jag’Ette Gloves
no gloves listed
Rapha Core Mitts
MAAP Pro Race Mitts
Endura Xtract Mitts
Galibier Specialissima Mitts
Le Col Cycling Mitts
Nalini Summer Mitts
Pearl Izumi Pro Air Mitts
Assos RS LF Targa gloves
Velocio Alpha winter cycling gloves
Giro Monaco II Gel Cycling Gloves
Pearl Izumi Elite Gel Cycling Gloves
Specialized Body Geometry Gel Long Finger Gloves
Bontrager Solstice Gel Cycling Gloves
Castelli Arenberg Gel 2 Cycling Gloves
Last example: lightest road bike helmets.
Again we can see the same pattern of a non-exact match general key phrase taking the number one spot (and two and three). Number four is an exact match key phrase (Best lightweight road cycling helmets for 2023). But maybe the year (2023) sets it back.
If you actually read the article, it becomes clear that its focus isn’t on displaying the lightest helmets per se, but rather on creating enough content to have an article that can rank, which it does. Although I have to say that based on my own criteria of 280 grams being the upper limit of lightweight, the majority of mentioned helmets fall within the lightweight category.
Again ChatGPT does a great job of listing a number of high-quality lightweight options, although the SPIN technology of the POC Ventral Air SPIN helmet has since been replaced by the MIPS version. So this is technically correct, were it not for the fact there’s an improved version on the market.
road.cc (number 4)
Giro Aether Spherical
Giro Aries Spherical
Lazer G1 MIPS
Lazer Sphere MIPS
Kask Protone Icon
POC Octal MIPS
Specialized Prevail II ANGI MIPS
MET Trenta 3K Carbon
Scott Centric Plus
Giro Aether Spherical MIPS
Bell Stratus MIPS
Lazer Z1 MIPS
MET Estro MIPS
MET Manta MIPS
Giro Helios Spherical
Giro Syntax MIPS
Giro Aether MIPS
Specialized S-Works Prevail II
Lazer Z1 MIPS
POC Ventral Air SPIN
So is the
Conclusion: AI will rule the information game
Based on this little experiment it’s clear you’re probably better of trusting AI in the form of ChatGPT than the big brand names taking the top rankings.
Google now serves up web pages which might have a certain amount of authority based on outdated SEO principles, but you’re not going to tell me that a non-exact match page is better than my highly researched exact match page. And even if it was, it’s definitely not better than ChatGPT, which is objective, doesn’t have to pay copywriters, SEO people, and other humans, is impartial, doesn’t get paid by big brand names, and can draw its conclusion based on far more content than any human ever could.
Even if it made a mistake in the third example, it’s only because it did so based on probably an outdated snapshot of relevant data. This is a small kink in the cable, which will get smoothed out as time progresses and AI will really start becoming our go-to source of information.
Obviously, I’m not the first one to say that AI will wipe out every straightforward content creator out there. I have already been wiped out courtesy of Google who is trying to play catch-up with ChatGPT and other tech companies in this AI arms race.
SEO as I knew it is very dead, and after all those years the Google god is under siege. If you think about it, the Google SERP is archaic. An AI conversational partner is the way forward in terms of providing us with our informational needs. How supply and demand and websites fit into that equation I have no idea. So until a company actually makes a version of the AI partner, which can live on our phone, I don’t know what it’s going to look like, but it might not be Google.
As for me in the short term; if I want to keep doing this online marketing thing, I have to turn myself from a human content creator into someone who specializes in AI to create content. I’m already convinced that AI in the form of ChatGPT can create a certain type of informational content better than any human on this planet.
In the end I’m just a small fish online, but even the biggest brands relying on human-made informational content will suffer. For me and my family, I just have to figure out my next step. And maybe the best thing would be to get as far from online content and AI’s disruptive nature as possible.
Johan van Seijen is the founder of restoration.bike. His cycling career has seen him at the starting line of classics such as the Amstel Gold Race and Liege Bastogne Liege. Realizing his racing capacity would fall short of what was needed he obtained a MS from the University of Amsterdam in engineering. His love for cycling changed into riding in an amateur capacity with his local cycling club TFC Weesp as a roadie and supporting MTB Noordwest as a mountain biker. He repairs, restores, and builds bicycles and shares his knowledge on YouTube, Facebook and this website.