Retro Dodo is a website dedicated to retro games and hardware. I recently added it to my feed collection and I sometimes take the time to read through its new articles. On April 3, 2024, the founder of Retro Dodo, Brandon Saltalamacchia, published an article titled Google Is Killing Retro Dodo & Other Independent Sites. I had my own struggle with big tech search last year when Bing arbitrarily blacklisted The New Leaf Journal from appearing in search results and then de-indexed the entire site, causing an unfortunate and still-unexplained situation that persisted for the better part of eight months. As the admin of a small website, albeit not one with Retro Dodo’s commercial ambitions, I was interested to read about their experience. While I do not agree with all of Retro Dodo’s take-aways about the current search landscape, the article does raise some interesting points that webmasters of small websites trying to produce humane (as in by humans, for humans) writing and media should consider before depending on Google, Bing, or social media algorithms to deliver traffic.

Retro Dodo is an independent blog and website that generates revenue through on-site ads, its YouTube channel, and book sales. It has on-site ads and also hosts a YouTube channel and several books under its banner. The on-site ads distinguishes it from The New Leaf Journal. While we have used The New Leaf Journal to promote our commercial endeavors (see some of Victor V. Gurbo’s articles about his music), the site proper avoids third-party scripts, much less ads or affiliate links. The point here is not about morality – but instead that having ads, especially Google ads, brings additional potential search engine issues that may not apply to slimmer projects such as mine. (Of course as I learned last year – Bing may ban you even when you have no ads or third-party scripts.)

Retro Dodo states that it saw steady growth from its beginnings several years ago until September 2023. Then the problems started:

Well, that all came to an abrupt end in September 2023 when Google decided to release an algorithm update that completely obliterated thousands of independent content businesses overnight, and we are one of them.

While Retro Dodo does not provide impression and visitor numbers, it posts clips from its Google Search Console account showing an 85% reduction in traffic from September 5, 2023 to March 31, 2024 (note: I displayed the results of our Bing de-indexing the same way).

From the outset – I will note that I am not convinced that whatever change Google made in September 2023 was specifically designed to nuke small websites from orbit – in part because I doubt that Google is particularly concerned with small websites one way or another. I only have access to my own Search Console statistics and that makes for an admittedly small sample size, but see my graph for the same 9/5/23 to 3/31/24 period depicted by Retro Dodo:

A Google Search Console clip showing The New Leaf Journal's impressions and clicks from September 2023 through March 2024. There is a general uptick from November through February.
The purple line shows impressions and the blue line shows clicks.

We had a relatively poor September and October followed by a sharp jump in November that has mostly held up – peaking in February and going down a bit in March. January and February 2024 were our two best Google months on record, but they were not unprecedented. For example, we were close to those numbers in March 2023 and also had some strong months in the spring and fall of 2022. Our Google numbers bounce around a bit and we have good and bad stretches within relatively stable upper and lower bounds, but given our steady, relatively low impression-conversion rate, much can probably be explained by day-to-day randomness. Our spikes sometimes correspond to new articles – see our excellent February where we were bolstered by the strong Google performance of my articles on the hair color of Yuki Itose and Itsuomi Nagi in the then-ongoing A Sign of Affection anime.

Taken together, the new algorithm changes over at Google seem to have had little-to-no effect on my site. One could speculate that the effect has been slightly positive based on our strong 4-month Google run from November 2023 through February 2024, but I do not think that the results are so out of line with past numbers to confidently draw that conclusion (especially in light of our weak October 2023 and slight decline in March 2024). For example, we had stretches in 2023 where we had similar numbers of clicks from a smaller number of impressions:

A Google Search Console clip showing The New Leaf Journal's impressions and clicks from December 2022 through March 2024. There is a general uptick from November through February.
The purple line shows impressions and the blue line shows clicks. This chart begins with December 3, 2022 and runs through the end of March, 2024.

However, while I have not had a site-wide negative experience with post-September 2023 Google, and while I think there are some material differences between my site and Retro Dodo’s that made the latter more susceptible to Google changes, Retro Dodo does raise several points that I am also concerned about – not only with respect to Google but also with respect to other search engines and algorithmic discovery tools.

I quote Retro Dodo’s first new Google concern:

Firstly, Google wants to completely eradicate users leaving their search results and will now show you their own ‘From Sources’ answers to search terms. These answers are taken without consent from the publisher’s content and are stolen from creators’ work so that Google can give you “their” answer instead, lowering the chances of readers exploring websites…

I agree entirely with this concern. I discussed it last year with respect to Google’s and Bing’s use of language models (or “AI”) to provide answers to questions instead of providing responsive websites to a search query (Retro Dodo discusses that specific issue later in its post). Even if these sorts of search engine answers cite to their sources, many visitors will not be inclined to visit the cited websites. In addition to the self-interested webmaster concern (less traffic), neglecting to check sources places too much faith in big tech language models. There are issues beyond the new effort to shove AI into every search. Search engines may select snippets for results that lower the likelihood that a searcher will click. Moreover, pre-existing instant answer tools, which have a tendency to appear above search results, may also reduce traffic. I see all of these concerns as symptoms of the same issue: Google and Bing are looking for ways to keep searchers from leaving Google and Bing.

Before continuing, I will add that there is a certain perversity to these different methods of “search engines” trying to answer questions instead of directing searchers to the answers. Many search queries return spammy results – and many supposedly authoritative websites produce globs of SEO spam content. Google promising to use AI to answer a question raises the nagging suspicion that Google Search will not return a website with a good answer at the top of traditional search results. One could say Google created a problem and then created a “solution” to their problem. Do note however that I remain skeptical that the general public has much interest in LLMs and AI chatbots, and Google has led the way in making them even less credible in recent weeks. The more immediate concern is Google pushing this stuff to the forefront of traditional search results.

Let us example Retro Dodo’s second Google concern:

[I]f you manage to scroll past Google’s stolen answer, you will undoubtedly be bombarded with sponsored ads, and lots of them. These results are purposely designed to look like normal results and can bend Google’s guidelines when it comes to content quality.

I will confess I do not know how bad ads are in Google because I do not actually use Google Search directly. If I need results drawn from Google’s index, I use Startpage. However, I agree that loading up searches with ads has the natural tendency to bury search results – especially on small screens. Retro Dodo keys in on “sponsored results” – which look more like traditional results than do most ads (see my discussion of some issues with sponsored results in my look at the Google index-dependent Norton Safe Search). I agree here that Google’s being the internet’s biggest advertiser creates some interesting incentives in its search engine much like how Facebook’s ad and data-collection interests ushered its transmogrification from social networking into social media.

Retro Dodo then describes what it sees as the most pressing problem of post-Setember 2023 Google:

Thirdly, and this is the biggest change I have ever seen in the search results over the last 5 years, is the fact that you will likely now see Reddit results upon searching for something.

I use a Firefox extension called uBlacklist to remove Reddit from most search results in my primary browser. I personally do not use Reddit, and I agree with Mr. Saltalamacchia’s position that Reddit’s place is not competing with well-thought out articles in general search queries (if one wants to search Reddit, he or she can easily employ domain-specific search queries). From my perspective, I doubt that an increase in Reddit appearances in results has a dramatic effect on The New Leaf Journal, but I can see why it would be a significant problem for a site that depends on tech reviews to generate traffic. The Retro Dodo article goes into the reasons it questions the timing and ethics of Reddit’s increased Google presence – but I will set those aside as beyond our scope.

Mr. Saltlamacchia described his frustration with the opacity of Google’s algorithmic changes:

It’s likely that our website structure isn’t perfect or that our content may likely be “over-optimised” for Google, but at least tell me. I am no SEO expert; I have a million plates to spin as a content website operator, and I cannot get everything right. If Google were to inform me what I was doing incorrectly then I could change that, but they don’t. They just keep hiding our work, month after month after month.

I had a similar complaint with respect to being blacklisted by Bing and not being able to get through to a human being for more than five months (although I am confident after having thoroughly investigated the issues that the problem was entirely on Bing’s end). Google Search Console is fairly robust (it is certainly better than Bing’s Webmaster Portal) and has alerted me to minor errors or potential issues. But when it does not provide actionable information to a site that loses 85% of its impressions and conversions over the course of six months, one could hardly say it is much more useful than Bing Webmaster when it is most needed.

I am not an SEO expert and other than performing basic SEO clerical tasks such as applying meta titles and descriptions to my posts, I do not worry about it much. I try to write interesting articles about what I want to write about and let Google, Bing, Yandex, and the rest rank my articles as they will. I sometimes have ideas for articles that may have search potential and take cues from what readers seem to be interested in – but in the end I cover what I want to cover. With that in mind, you can take my Google analysis of Retro Dodo with a grain of salt. It seems likely to me that Google took some kind of manual action against Retro Dodo after the algorithm change since the massive collapse in impressions and conversions seems extreme for mere ranking tweaks. What was the cause? The first two culprits I would suspect are how Retro Dodo uses ads or its schema markup. I am not an expert on staying out of trouble with ads since I never ran any of them. With respect to schema markup, however, I looked into the subject to see if I could make better use of JSON schema markup to improve the appearance of our articles in search. One area I looked at was reviews. My reading of Google’s review schema instructions left me with the impression that the review schema markup is more for aggregate user reviews than reviewers and that the product markup is more for things being sold. I could be off base – but I decided that the limited potential upside of maybe improving how some of my article previews appear in search was not worth the risk of potentially angering the Google and Bing algorithms (there is zero chance that someone will click on an article that is excluded from search). I would be curious whether it is possible that some of Retro Dodo’s schema markup on “products” and “reviews” triggered something in Google’s recent algorithm changes.

Manual actions aside – reviews are an oft-SEO-spammed subject, and I would be inclined to agree that since Reddit users love writing reviews, Google’s giving more weight to Reddit results would be ominous for review sites. I will add that another common Retro Dodo topic – “best of” lists – are also an SEO-spam favorite – which means there is significant low quality competition. Google favors certain domains more than it considers the quality of review results – with quality including whether the reviewer actually owns and used the thing being reviewed instead of scraping Amazon reviews or other sources for “review” information. So, I would proceed with writing reviews or best of lists under the assumption that quality is probably not the main factor for ranking on top of the results. Similarly, whileI saw some Hacker News commenters opine that Retro Dodo reviews sometimes contain uncorrected errors, I would be inclined to doubt that the sorts of errors that may appear in its reviews or even a dip in quality would lead to an 85% decrease in impressions and clicks.

The Retro Dodo experience brings us to one of my pet issues. Relying on big tech ad companies to sustain a website will always be a dangerous game – especially if you are counting on ads to sustain a business. This applies just as much to Bing (and derivatives thereof) and Yandex as it does to Google. I have issued similar warnings about big tech social media such as Facebook, X/Twitter, and the like. Complaining about the unfairness of it all accomplishes nothing. These companies do not care about small websites or their webmasters. They care about their own bottom lines and providing a baseline experience to the ordinary user (or product, it is a matter of perspective). What one algorithm update gives another may take. I personally see the point of running a small website (small website(s) in my case) as carving out my own place on the internet where people can find me and my writing. One reason I promote my RSS and ATOM feeds and also have a weekly newsletter is to make it possible for people to keep track of my writing at their leisure without sitting back and hoping that they will remember my name or the name of my sites and get it close enough that a search engine returns what they are looking for. To the extent I rely on big tech search, it is only in recognition of the fact that most internet users rely on Google (and to a much lesser extent, Bing and derivatives thereof) to find websites and articles.

Projects such as Retro Dodo should try to cultivate dedicated communities around their reviews – fora, mailing lists, or similar. Retro Dodo is powered by WordPress which is highly extensible in community-building directions for those who are so inclined. Going forward – I would like to see humane review sites (reviews by humans written for human beings) work together. I am sure there are similar game projects in Retro Dodo’s space that have slightly different points of emphasis. Cross-linking between projects serves as another way to help readers looking for genuine reviews and content find what they are looking for without relying on Google. See, for example, the recent RSS-driven site-to-site conversation between The New Leaf Journal and Yukinu Blog.

Reviews are in particularly bad shape on the internet ,and before reading about Retro Dodo’s unfortunate situation I had already been thinking about an interesting angle to approach review site challenges . This post reminded me to keep developing those ideas for future essay(s).

I conclude by wishing the team at Retro Dodo luck in figuring out a way to reverse the unfortunate Google nose-dive they have been seeing. Retro consoles are a strange, niche space and people who take the time to acquire and review them provide much more value than random Amazon reviewers or idle Reddit commenters.