Upon the recently announced Google update I've seen some people Tweet things like

• if you are afraid of algorithm updates, you must be a crappy SEO
• if you are technically perfect in your SEO, updates will only help you

I read those sorts of lines and cringe.

Here's why...

Fragility

Different businesses, business models, and business structures have varying degrees of fragility.

If your business is almost entirely based on serving clients then no matter what you do there is going to be a diverse range of outcomes for clients on any major update.

Let's say 40% of your clients are utterly unaffected by an update & of those who saw any noticeable impact there was a 2:1 ratio in your favor, with twice as many clients improving as falling.

Is that a good update? Does that work well for you?

If you do nothing other than client services as your entire business model, then that update will likely suck for you even though the net client impact was positive.

Why?

Many businesses are hurting after the Covid-19 crisis. Entire categories have been gutted & many people are looking for any reason possible to pull back on budget. Some of the clients who won big on the update might end up cutting their SEO budget figuring they had already won big and that problem was already sorted.

Some of the clients that fell hard are also likely to either cut their budget or call endlessly asking for updates and stressing the hell out of your team.

Capacity Utilization Impacts Profit Margins

Your capacity utilization depends on how high you can keep your steady state load relative to what your load looks like at peaks. When there are big updates management or founders can decide to work double shifts and do other things to temporarily deal with increased loads at the peak, but that can still be stressful as hell & eat away at your mental and physical health as sleep and exercise are curtailed while diet gets worse. The stress can be immense if clients want results almost immediately & the next big algorithm update which reflects your current work may not happen for another quarter year.

How many clients want to be told that their investments went sour but the problem was they needed to double their investment while cashflow is tight and wait a season or two while holding on to hope?

Category-based Fragility

Businesses which appear to be diversified often are not.

• Everything in hospitality was clipped by Covid-19.
• 40% of small businesses across the United States have stopped making rent payments.
• When restaurants massively close that's going to hit Yelp's business hard.
• Auto sales are off sharply.

Likewise there can be other commonalities in sites which get hit during an update. Not only could it include business category, but it could also be business size, promotional strategies, etc.

Sustained profits either come from brand strength, creative differentiation, or systemization. Many prospective clients do not have the budget to build a strong brand nor the willingness to create something that is truly differentiated. That leaves systemization. Systemization can leave footprints which act as statistical outliers that can be easily neutralized.

Sharp changes can happen at any point in time.

For years Google was funding absolute garbage like Mahalo autogenerated spam and eHow with each month being a new record. It is very hard to say "we are doing it wrong" or "we need to change everything" when it works month after month after month.

Then an update happens and poof.

• Was eHow decent back in the first Internet bubble? Sure. But it lost money.
• Was it decent after it got bought out for a song and had the paywall dropped in favor of using the new Google AdSense program? Sure.
• Was it decent the day Demand Media acquired it? Sure.
• Was it decent on the day of the Demand Media IPO? Almost certainly not. But there was a lag between that day and getting penalized.

Panda Trivia

The first Panda update missed eHow because journalists were so outraged by the narrative associated with the pump-n-dump IPO. They feared their jobs going away and being displaced by that low level garbage, particularly as the market cap of Demand Media eclipsed the New York Times.

Journalist coverage of the pump-n-dump IPO added credence to it from an algorithmic perspective. By constantly writing hate about eHow they made eHow look like a popular brand, generating algorithmic signals that carried the site until Google created an extension which allowed journalists and other webmasters to vote against the site they had been voting for through all their outrage coverage.

Algorithms & the Very Visible Hand

And all algorithmic channels like organic search, the Facebook news feed, or Amazon's product pages go through large shifts across time. If they don't, they get gamed, repetitive, and lose relevance as consumer tastes change and upstarts like Tiktok emerge.

Consolidation by the Attention Merchants

Frequent product updates, cloning of upstarts, or outright acquisitions are required to maintain control of distribution:

"The startups of the Rebellion benefited tremendously from 2009 to 2012. But from 2013 on, the spoils of smartphone growth went to an entirely different group: the Empire. ... A network effect to engage your users, AND preferred distribution channels to grow, AND the best resources to build products? Oh my! It’s no wonder why the Empire has captured so much smartphone value and created a dark time for the Rebellion. ... Now startups are fighting for only 5% of the top spots as the Top Free Apps list is dominated by incumbents. Facebook (4 apps), Google (6 apps), and Amazon (4 apps) EACH have as many apps in the Top 100 list as all the new startups combined."

Apple & Amazon

Emojis are popular, so those features got copied, those apps got blocked & then apps using the official emojis also got blocked from distribution. The same thing happens with products on Amazon.com in terms of getting undercut by a house brand which was funded by using the vendor's sales data. Re-buy your brand or else.

Facebook

Before the Facebook IPO some thought buying Zynga shares was a backdoor way to invest into Facebook because gaming was such a large part of the ecosystem. That turned out to be a dumb thesis and horrible trade. At times other things trended including quizzes, videos, live videos, news, self hosted Instant Articles, etc.

Over time the general trend was edge rank of professional publishers fell as a greater share of inventory went to content from friends & advertisers. The metrics associated with the ads often overstated their contribution to sales due to bogus math and selection bias.

Internet-first publishers like CollegeHumor struggled to keep up with the changes & influencers waiting for a Facebook deal had to monetize using third parties:

“I did 1.8 billion views last year,” [Ryan Hamilton] said. “I made no money from Facebook. Not even a dollar.” ... "While waiting for Facebook to invite them into a revenue-sharing program, some influencers struck deals with viral publishers such as Diply and LittleThings, which paid the creators to share links on their pages. Those publishers paid top influencers around $500 per link, often with multiple links being posted per day, according to a person who reached such deals."

YouTube

YouTube had a Panda-like update back in 2012 to favor watch time over raw view counts. They also adjust the ranking algorithms on breaking news topics to favor large & trusted channels over conspiracy theorist content, alternative health advice, hate speech & ridiculous memes like the Tide pod challenge.

All unproven channels need to start somewhat open to gain usage, feedback & marketshare. Once they become real businesses they clamp down. Some of the clamp down can be editorial, forced by regulators, or simply anticompetitive monpolistic abuse.

Kid videos were a huge area on YouTube (perhaps still are) but that area got cleaned up after autogenerated junk videos were covered & the FTC clipped YouTube for delivering targeted ads on channels which primarily catered to children.

Dominant channels can enforce tying & bundling to wipe out competitors:

"Google’s response to the threat from AppNexus was that of a classic monopolist. They announced that YouTube would no longer allow third-party advertising technology. This was a devastating move for AppNexus and other independent ad technology companies. YouTube was (and is) the largest ad-supported video publisher, with more than 50% market share in most major markets. ... Over the next few months, Google’s ad technology team went to each of our clients and told them that, regardless of how much they liked working with AppNexus, they would have to also use Google’s ad technology products to continue buying YouTube. This is the definition of bundling, and we had no recourse. Even WPP, our largest customer and largest investors, had no choice but to start using Google’s technology. AppNexus growth slowed, and we were forced to lay off 100 employees in 2016."

Everyone Else

Every moderately large platform like eBay, Etsy, Zillow, TripAdvisor or the above sorts of companies runs into these sorts of issues with changing distribution & how they charge for distribution.

Building Anti-fragility Into Your Business Model

Growing as fast as you can until the economy craters or an algorithm clips you almost guarantees a hard fall along with an inability to deal with it.

Markets ebb and flow. And that would be true even if the above algorithmic platforms did not make large, sudden shifts.

Build Optionality Into Your Business Model

If your business primarily relies on publishing your own websites or you have a mix of a few clients and your own sites then you have a bit more optionality to your approach in dealing with updates.

Even if you only have one site and your business goes to crap maybe you at least temporarily take on a few more consulting clients or do other gig work to make ends meet.

Focus on What is Working

If you have a number of websites you can pour more resources into whatever sites reacted positively to the update while (at least temporarily) ignoring any site that was burned to a crisp.

Ignore the Dead Projects

The holding cost of many websites is close to zero unless they use proprietary and complex content management systems. Waiting out a penalty until you run out of obvious improvements on your winning sites is not a bad strategy. Plus, if you think the burned site is going to be perpetually burned to a crisp (alternative health anyone?) then you could sell links off it or generate other alternative revenue streams not directly reliant on search rankings.

Build a Cushion

If you have cash savings maybe you guy out and buy some websites or domain names from other people who are scared of the volatility or got clipped for issues you think you could easily fix.

When the tide goes out debt leverage limits your optionality. Savings gives you optionality. Having slack in your schedule also gives you optionality.

The person with a lot of experience & savings would love to see highly volatile search markets because those will wash out some of the competition, curtail investments from existing players, and make other potential competitors more hesitant to enter the market.

Historically cloaking was considered bad because a consumer would click expecting a particular piece of content or user experience while being delivered an experience which differed dramatically.

As publishers have become more aggressive with paywalls they've put their brands & user trust in the back seat in an attempt to increase revenue per visit.

Below are 2 screenshots from one of the more extreme versions I have seen recently.

The first is a subscribe-now modal which shows by default when you visit the newspaper website.

The second is the page as it appears after you close the modal.

Basically all page content is cloaked other than ads and navigation.

The content is hidden - cloaked.

That sort of behavior would not only have a horrible impact on time on site metrics, but it would teach users not to click on their sites in the future, if users even have any recall of the publisher brand.

The sort of disdain that user experience earns will cause the publishers to lose relevancy even faster.

On the above screenshot I blurred out the logo of the brand on the initial popover, but when you look at the end article after that modal pop over you get a cloaked article with all the ads showing and the brand of the site is utterly invisible. A site which hides its brand except for when it is asking for money is unlikely to get many conversions.

Many news sites now look as awful as the ugly user created MySpace pages did back in the day. And outside of the MySpace pages that delivered malware the user experience is arguably worse.

Each news site which adopts this approach effectively increases user hate toward all websites adopting the approach.

It builds up. Then users eventually say screw this. And they are gone - forever.

Audiences will thus continue to migrate across from news sites to anywhere else that hosts their content like Google AMP, Facebook Instant Articles, Apple News, Twitter, Opera or Edge or Chrome mobile browser new article recommendations, MSN News, Yahoo News, etc.

Any lifetime customer value models built on assumptions around any early success with the above approach should consider churn as well as the brand impact the following experience will have on most users before going that aggressive.

One small positive note for news publishers is more countries are looking to have attention merchants pay for their content, though I suspect as the above sort of double modal paywall stuff gets normalized other revenue streams won't make the practice go away, particularly as many local papers have been acquired by PE chop shops extracting all blood out of the operations through interest payments to themselves.

Google's mission statement is "organize the world's information and make it universally accessible and useful."

That mission is so profound & so important the associated court documents in their antitrust cases must be withheld from public consumption.

Hey. The full exhibit list just posted in DC federal court for USA vs Google. J/k, they literally posted the numbers of all of the admitted exhibits which would be unsealed in a sane world where public interest is respected even more so because the defendant is insanely powerful. pic.twitter.com/FViD40xVmf— Jason Kint (@jason_kint) September 23, 2023

Before document sharing was disallowed, some were shared publicly.

Internal emails stated:

• Hal Varian was off in his public interviews where he suggested it was the algorithms rather than the amount of data which is prime driver of relevancy.
• Apple would not get any revshare if there was a user choice screen & must set Google as the default search engine to qualify for any revshare.
• Google has a policy of being vague about using clickstream data to influence ranking, though they have heavily relied upon clickstream data to influence ranking. Advances in machine learning have made it easier to score content to where the clickstream data had become less important.
• When Apple Maps launched & Google Maps lost the default position on iOS Google Maps lost 60% of their iOS distribution, and that was with how poorly the Apple Maps roll out went.
• Google sometimes subverted their typical auction dynamics and would flip the order of the top 2 ads to boost ad revenues.
• Google had a policy of "shaking the cushions" to hit the quarterly numbers by changing advertiser ad prices without informing advertisers that they'd be competing in a rigged auction with artificially manipulated shill bids from the auctioneer competing against them.

When Google talked about hitting the quarterly numbers with shaking the cusions the 5% number which was shared skewed a bit low:

For a brand campaign focused on a niche product, she said the average CPC at $11.74 surged to $25.85 over the last six months, amounting to a 108% increase. However, there wasn’t an incremental return on sales.

“The level to which [price manipulations] happens is what we don’t know,” said Yang. “It’s shady business practices because there’s no regulation. They regulate themselves.”

Early in the history of search ads Google blocked trademark keyword bidding. They later allowed it. When keyword bidding on trademarks was allowed it led to a conundrum for some advertisers. If you do not defend your trademark you could lose it, but if you agree with competitors not to bid on each other's trademarks the FTC could come after you - like they did with 1-800 Contacts. This set up forces many brands to participate in auctions where they are arbitraging their own pre-existing brand equity. The ad auctioneer runs shady auctions where it looks across at your account behavior and bids then adjusts bid floors to suck more money out of you. This amounts to something akin to the bid jamming that was done in early Overture, except it is the house itself doing it to you! The last auction I remembered like that was SnapNames, where a criminal named Nelson Brady on the executive team used the handle halverez to leverage participant max bids and put in bids just under their bids. The goal of his fraud? To hit the numbers & get an earn out bonus - similar to how Google insiders were discussing "shaking the cushions" to hit the number.

Halverez created a program which looked across aggregate bid data, join auctions which only had 1 other participant, and then use the one-way view of competing bids to put in a shill bid to drive up costs - which sure sounds conceptually similar to Google's "shaking the cushions."

"Just looking at this very tactically, and sorry to go into this level of detail, but based on where we are I'm afraid it's warranted. We are short __% queries and are ahead on ads launches so are short __% revenue vs. plan. If we don't hit plan, our sales team doesn't get its quota for the second quarter in a row and we miss the street's expectations again, which is not what Ruth signaled to the street so we get punished pretty badly in the market. We are shaking the cushions on launches and have some candidates in May that will help, but if these break in mid-late May we only get half a quarter of impact or less, which means we need __% excess to where we are today and can't do it alone. The Search team is working together with us to accelerate a launch out of a new mobile layout by the end of May that will be very revenue positive (exact numbers still moving), but that still won't be enough. Our best shot at making the quarter is if we get an injection of at least __%, ideally __%, queries ASAP from Chrome. Some folks on our side are running a more detailed, Finance-based, what-if analysis on this and should be done with that in a couple of days, but I expect that these will be the rough numbers.

The question we are all faced with is how badly do we want to hit our numbers this quarter? We need to make this choice ASAP. I care more about revenue than the average person but think we can all agree that for all of our teams trying to live in high cost areas another $___,___ in stock price loss will not be great for morale, not to mention the huge impact on our sales team." - Google VP Jerry Dischler

Google is also pushing advertisers away from keyword-based bidding and toward a portfolio approach of automated bidding called Performance Max, where you give Google your credit card and budget then they bid as they wish. By blending everything into a single soup you may not know where the waste is & it may not be particularly easy to opt out of poorly performing areas. Remember enhanced AdWords campaigns?

Google continues to blur dataflow outside of their ad auctions to try to bring more of the ad spend into their auctions.

Wow. Google. Years behind other browsers (aka monopoly power), Google is attempting to deprecate tracking system A (aka third party cookies) and replace it with another tracking system B (aka Topics) that treats sites as G data mules.

This is deceptive as hell comparing B to A. pic.twitter.com/hCBJgYr7qn— Jason Kint (@jason_kint) September 22, 2023

The amount Google is paying Apple to be the default search provider is staggering.

What is $18 billion / year buying ? The DoJ has narrowed in an agreement not to compete between Apple and Google: "Sanford Bernstein estimates Google will pay Apple between $18 billion and $19 billion this year for default search status" https://t.co/HmoZxCZkqm— Tim Wu (@superwuster) September 22, 2023

Tens of billions of dollars is a huge payday. No way Google would hyper-optimize other aspects of their business (locating data centers near dams, prohibiting use of credit card payments for large advertisers, cutting away ad agency management fees, buying Android, launching Chrome, using broken HTML on YouTube to make it render slowly on Firefox & Microsoft Edge to push Chrome distribution, all the dirty stuff Google did to violate user privacy with overriding Safari cookies, buying DoubleClick, stealing the ad spend from banned publishers rather than rebating it to advertisers, creating a proprietary version of HTML & force ranking it above other results to stop header bidding, & then routing around their internal firewall on display ads to give their house ads the advantage in their ad auctions, etc etc etc) and then just throw over a billion dollars a month needlessly at a syndication partner.

This is right -- Google was once an extraordinary product, but over time became stagnant & too grabby of random revenue as it ate its ecosystem. Makes it the right time to force Google to try and compete without reaching for its bribery checkbook
https://t.co/gDhtDMjfo0— Tim Wu (@superwuster) September 22, 2023

For perspective on the scale of those payments consider that it wasn't that long ago Yahoo! was considered a big player in search and Apollo bought Yahoo! plus AOL from Verizon for about $5 billion & then was quickly able to sell branding & technology rights in Japan to Softbank for $1.6 billion & other miscellaneous assets for nearly a half-billion, reducing the net cost to only $3 billion.

If Google loses this lawsuit and the payments to Apple are declared illegal, that would be a huge revenue (and profit) hit for Apple. Apple would be forced to roll out their own search engine. This would cut away at least 30% of the search market from Google & it would give publishers another distribution channel. Most likely Apple Search would launch with a lower ad density than Google has for short term PR purposes & publishers would have a year or two of enhanced distribution before Apple's ad load matched Google's ad load.

It is hard to overstate how strong Apple's brand is. For many people the cell phone is like a family member. I recently went to upgrade my phone and Apple's local store closed early in the evening at 8pm. The next day when they opened at 10 there was a line to wait in to enter the store, like someone was trying to get concert tickets. Each privacy snafu from Google helps strengthen Apple's relative brand position.

Google has also diluted the quality of their own brand by rewriting search queries excessively to redirect traffic flows toward more commercial interests. Wired covered how Project Mercury works:

This onscreen Google slide had to do with a “semantic matching” overhaul to its SERP algorithm. When you enter a query, you might expect a search engine to incorporate synonyms into the algorithm as well as text phrase pairings in natural language processing. But this overhaul went further, actually altering queries to generate more commercial results. ... Most scams follow an elementary bait-and-switch technique, where the scoundrel lures you in with attractive bait and then, at the right time, switches to a different option. But Google “innovated” by reversing the scam, first switching your query, then letting you believe you were getting the best search engine results. This is a magic trick that Google could only pull off after monopolizing the search engine market, giving consumers the false impression that it is incomparably great, only because you’ve grown so accustomed to it.

The mobile search results on Google require at least a screen or two of scrolls to get to the organic results if there is a hint of commercial intent behind the search query. Once they have monetized the real estate they are reliant on broader economic growth & using ad buy bundling to drive cross-subsidies of other non-search ad inventory, which may contain more than a bit of fraud. Performance Max may max out your spend without actually performing for anybody other than Google.

Google not only shill bid on lower competition terms to squeeze defensive brand bids and boost auction floor pricing, but they also implemented shill bids in competitive ad auctions:

Michael Whinston, a professor of economics at the Massachusetts Institute of Technology, said Friday that Google modified the way it sold text ads via “Project Momiji” – named for the wooden Japanese dolls that have a hidden space for friends to exchange secret messages. The shift sought “to raise the prices against the highest bidder,” Whinston told Judge Amit Mehta in federal court in Washington.

While Google's search marketshare is rock solid, the number of search engines available has increased significantly over the past few years. Not only is there Bing and DuckDuckGo but the tail is longer than it was a few years back. In addition to regional players like Baidu and Yandex there's now Brave Search, Mojeek, Qwant, Yep, and You. GigaBlast and Neeva went away, but anything that prohibits selling defaults to a company with over 90% marketshare will likely lead to dozens more players joining the search game. Search traffic will remain lucrative for whoever can capture it, as no matter how much Google tries to obfuscate marketing data the search query reflects the intent of the end user.

“Search advertising is one of the world’s greatest business models ever created…there are certainly illicit businesses (cigarettes or drugs) that could rival these economics, but we are fortunate to have an amazing business.” - Google VP of Finance Mike Roszak