August 9th, 2024. You wake up. You check your phone. Your monitoring dashboard - the one that sends you an alert when any of your clients' sites experiences a traffic anomaly beyond two standard deviations from the rolling average - has sent you fourteen alerts. Fourteen. In the time it took you to sleep (poorly, because you're the kind of person who checks dashboards before bed and after waking and sometimes in between, which is a behavioral pattern that a therapist might describe as "not great"). You open Analytics. The graphs look like the market on Black Monday. Every client. Every site. Every metric. Zero. Not declining. Zero. As if the internet had simply decided to stop sending people to websites, which is, as it turns out, more or less what happened.
You sit up in bed. The sheets are tangled around your legs because you are a restless sleeper, because you are an SEO professional, and SEO professionals sleep the way cats nap - with one eye open, ready to be startled by something you can't control and don't fully understand. Your phone is now buzzing continuously. The alerts are stacking. Client after client after client. The graphs all look the same. Flat lines. Not the gentle decline of a ranking drop or the seasonal dip of a slow Tuesday. Flat lines. The kind of flat line that, in a medical context, means someone should be calling a code.
You open Twitter. (You still call it Twitter. You will always call it Twitter. Elon can rename it whatever he wants. It's Twitter.) And Twitter is on fire. Or rather, SEO Twitter is on fire, which is a small corner of Twitter that is always at least mildly inflamed but is currently experiencing something closer to a five-alarm conflagration. Everyone is posting the same thing. "Is Google down?" "Google is down." "Anyone else seeing Google down?" "My traffic just went to zero." "My traffic went to zero too." "Is this a core update?" "This isn't a core update, Google is literally down." "What do you mean Google is down?" "I mean Google is down." "Google can't go down." "Google is down."
Google was down.
Forty-Seven Minutes
The outage lasted forty-seven minutes. That's the official number. Forty-seven minutes during which Google Search - the product, the infrastructure, the invisible utility that the entire internet economy treats the way a city treats its water supply, which is to say with the assumption that it will always be there and the utter inability to function when it isn't - was either unavailable or severely degraded. Not the ads. Not YouTube. Not Gmail. Google Search. The thing we built our careers on. The thing we've spent decades learning to optimize for. The thing that sends traffic to websites, which is the thing that makes websites worth having, which is the thing that makes the internet a commercial enterprise rather than an academic curiosity.
For forty-seven minutes, that thing stopped working.
I should be more precise. Google Search didn't disappear entirely. It wasn't as if google.com returned a blank page or a 503 error or a note saying "Gone fishing, be back later." What happened was more nuanced and, in some ways, more alarming. Google Search was intermittent. It would work for some queries and not others. It would return results for some users and errors for others. It would load the search page but not populate the results. It would serve cached results that were hours old. It was, in the technical sense, degraded - functioning at some reduced capacity, like a car engine that starts but only fires on three of four cylinders, enough to make you think it's working, not enough to actually get you anywhere.
The SEO industry spent those forty-seven minutes in a state I can only describe as collective psychological arrest. Not panic, exactly. Panic implies movement, adrenaline, the fight-or-flight response that propels you toward action. This was something else. This was the moment before panic, when the brain is still processing the input, still trying to reconcile what it's seeing with what it believes to be possible, still running the calculation that goes: this can't be happening, therefore it isn't happening, therefore what I'm seeing must mean something else, except it doesn't mean something else, it means exactly what it appears to mean, which is that the foundation of everything I've built my career on has cracked.
I know this because I was there. I was sitting in my home office, in a chair I've sat in for thousands of mornings, staring at dashboards I've stared at for thousands of mornings, and for the first time in my professional life, the dashboards showed nothing. Not bad data. Not confusing data. No data. The absence of data. The graphs had flat-lined and the real-time counters showed zero active users on every site and the Search Console performance report would later show a gap - a literal gap in the graph, a missing section, like someone had cut a piece out of a timeline with scissors - during the window of the outage.
My phone rang. It was a client. I won't tell you which one. It doesn't matter which one because they all called and they all asked the same question.
"What's happening?"
"Google is down," I said.
"What do you mean Google is down?"
And this question - this specific question, asked with this specific tone of bewildered disbelief, as if I had just told them the sun had failed to rise or gravity had taken the morning off - this question is the entire point of this article. Because the answer I wanted to give was: I mean Google is down, the same way a power grid goes down or a bridge collapses or a supply chain breaks, because Google is infrastructure, because it's a system built by humans, and systems built by humans fail, and the fact that this particular system has failed so rarely that its failure seems impossible is not evidence that it can't fail but evidence of how deeply we've allowed ourselves to depend on something we don't control.
What I actually said was: "It should come back soon. There's nothing we can do."
That second sentence. There's nothing we can do. I've been turning it over in my mind for months now. Because it's true. It was true in the moment and it's true in the broader sense and it describes, with uncomfortable accuracy, the fundamental position of every person and every business that has built its growth strategy on organic search. There is nothing we can do. Google works or it doesn't. Google sends traffic or it doesn't. Google ranks our pages or it doesn't. And when it doesn't - for forty-seven minutes or forty-seven days or permanently, because algorithm changes are just slow-motion outages that affect some sites and not others - there is nothing we can do. We are tenants. Google is the landlord. The entire SEO industry is a collection of people who have built elaborate, beautiful, carefully optimized structures on land they don't own.
The Math
Let me give you the numbers, because numbers make things real in a way that analogies don't, and the numbers here are sobering enough to justify a moment of genuine alarm.
Google processes approximately 8.5 billion searches per day. That's the commonly cited estimate. It's probably higher now. Let's use it anyway. 8.5 billion searches per day works out to roughly 5.9 million searches per minute. During a forty-seven-minute outage - or near-outage, or degraded service, whatever you want to call the thing that happened - approximately 277 million searches were affected. Not all of them failed completely. Some returned results. Some returned partial results. Some returned stale results. But 277 million search sessions experienced something other than the expected behavior of typing a query and receiving relevant results in under a second.
Now let's talk about money. Google's advertising revenue - the money that businesses pay to appear at the top of search results - was approximately $237 billion in 2023. That works out to roughly $650 million per day. Per minute, that's about $451,000. During a forty-seven-minute outage, approximately $21 million in advertising revenue was at risk. Not all of it was lost - some ads continued to serve, some searches still worked - but the exposure was $21 million. For forty-seven minutes. For less than one hour of one day.
That's Google's money. Now let's talk about everyone else's.
Organic search drives, depending on whose estimate you trust, between 50% and 70% of all website traffic across the internet. Let's use 53%, which is a conservative estimate from BrightEdge that I've seen cited so many times in so many conference presentations that I've developed a Pavlovian response to the number - I see "53%" and I automatically think "and the other 47% wishes it were organic too." If Google Search is responsible for 53% of all website traffic, and website traffic is responsible for (let's say conservatively) some percentage of all online revenue, then even a brief interruption in Google Search has a measurable economic impact that extends far beyond Google's own advertising losses.
E-commerce alone generates roughly $1.1 trillion per year in the US. If 53% of that traffic comes from organic search, and organic search goes down for forty-seven minutes, the theoretical revenue at risk is - I'll spare you the long division - approximately $51 million. For forty-seven minutes. From one country. In one channel. And this doesn't account for lead generation, for SaaS signups, for content publishers whose ad revenue depends on pageviews, for local businesses whose phone rings because someone Googled "plumber near me" and now the phone isn't ringing because Google isn't serving results for "plumber near me" and the pipe is still leaking and the homeowner is getting increasingly agitated and is probably going to call the first plumber they can find in their phone's contact list or, God help us, look for one on Bing.
The forty-seven-minute cost, across the entire internet economy, was almost certainly in the hundreds of millions of dollars. For forty-seven minutes. Of one product. Owned by one company. That we all decided, collectively, without signing anything or negotiating terms or even really thinking about it, to build our businesses on.
That's the math. The math is insane.
The Phone Calls
I took eleven phone calls on August 9th. Not during the outage itself - the outage was early enough that most of my clients' decision-makers hadn't started their workday yet, a small mercy for which I'm grateful, because if the outage had happened at 2 PM Eastern on a Tuesday I would have taken forty phone calls and my blood pressure would have required medical intervention.
The calls came in throughout the morning. Each one followed roughly the same pattern. The client had seen the traffic dip - either through their own analytics, which most clients check obsessively, or through an alert system, which some clients have set up because they've read an article about monitoring website performance and have implemented the monitoring without implementing the understanding of what the monitoring means. They had seen the dip. They had panicked. They were calling me to find out what had happened and, more importantly, whether it was their fault.
That question - "is it our fault?" - is the question that underlies every client interaction in SEO. Every phone call. Every email. Every Slack message that starts with "hey, quick question" and is never quick and is rarely just a question. The client's fundamental anxiety is not "is our traffic down" but "did we do something wrong." Because if they did something wrong, there's a fix. If the problem is internal, it's solvable. If the problem is that Google, the entire apparatus of Google, simply stopped working for forty-seven minutes and there's nothing anyone could have done and nothing anyone can do to prevent it from happening again, that's not a solvable problem. That's an existential condition. And people don't like existential conditions. People like solutions.
So I explained, eleven times, with varying degrees of patience and repetition, that Google had experienced an outage. That it was not their fault. That it was not my fault. That it was not anyone's fault, in the sense that outages are a feature of complex systems and Google is the most complex search system ever built and the remarkable thing is not that it went down but that it stays up. That the traffic would recover (it did, within hours). That the rankings would be unaffected (they were, as far as anyone could tell). That this was, in the grand scheme of things, a non-event - a blip, a hiccup, a brief interruption in a service that has been so relentlessly available for so long that its unavailability feels like a violation of natural law.
Most of them were reassured. Some of them weren't. One client - the CEO of an e-commerce company, a man who built his business on Google Shopping and organic search and who spends more on SEO per month than some people make in a year - asked me a question that I've been thinking about ever since.
"What if it happens again?"
"It probably will," I said. "Outages happen."
"No," he said. "What if it happens again and it doesn't come back?"
I didn't have an answer. I still don't.
The Landlord
Let me tell you what it's like to build your house on someone else's land, because that's what we've all done, every one of us, everyone who has ever invested time or money or attention in organic search as a growth channel. We've built on rented ground. And the rent, so far, has been free, which is the most dangerous kind of rent there is, because free rent is the kind that makes you forget you're renting.
I've been in SEO for over twenty years. In that time, I've watched the relationship between websites and Google evolve from something resembling a partnership - we make content, you organize it, everyone benefits - to something that more closely resembles feudalism. Google owns the land. We work the land. Google takes a cut (in the form of ads that push organic results further down the page, in the form of featured snippets that answer queries without sending a click, in the form of AI overviews that synthesize our content and serve it directly, saving the user the trouble of visiting our sites). We accept this arrangement because the alternative - not being on Google - is not really an alternative. It's an absence. It's a void. A business that doesn't appear in Google Search is, for most practical purposes, a business that doesn't exist.
This is the dependency. Not the technical kind - not the "we need Google's servers to stay up" kind, though that's part of it. The existential kind. The kind where your business model is predicated on the continued existence, continued availability, continued fairness, and continued goodwill of a single company's product. A product you don't pay for. A product you have no contractual relationship with. A product whose terms of service you've never read and whose algorithm you don't understand and whose future plans you have no insight into and whose decisions about your visibility are final, unappealable, and explained in blog posts written in a dialect of English that uses words like "helpful" and "quality" without ever defining what those words mean in any specific or actionable way.
We are dependent on Google the way a farmer is dependent on rain. We can do everything right - plant at the right time, use the right seeds, tend the soil with diligence and expertise - and if the rain doesn't come, the crop dies. We can build the best website, create the best content, implement the best technical SEO, acquire the best backlinks, and if Google decides to change how it ranks pages, or how it displays results, or whether it sends clicks to websites at all, the traffic dies. And we have no recourse. No contract to enforce. No SLA to invoke. No phone number to call. (Try calling Google. Really try. I'll wait. You'll be waiting longer.)
The outage was a forty-seven-minute reminder of this dependency. A reminder that most people processed, acknowledged, and immediately forgot, because the human brain is excellent at processing threats and even better at filing them under "probably won't happen again" and moving on. But it should not be forgotten. It should be framed and hung on the wall of every marketing department and every SEO agency and every conference stage where someone is about to give a presentation about maximizing organic traffic without mentioning, even once, the possibility that the entire channel could disappear.
Not because it's likely. I don't think Google is going to disappear. I don't even think Google is going to experience frequent outages. Google's infrastructure is, by any reasonable standard, the most robust computing system ever built by humans, and the fact that it went down for forty-seven minutes in twenty-plus years of continuous operation is actually a staggering achievement of engineering that we should probably be celebrating rather than hand-wringing about.
But the dependency is not just about outages. The dependency is about everything. It's about algorithm updates that can cut your traffic by 40% overnight, with no warning and no explanation and no appeal process. It's about feature changes - the introduction of featured snippets, the expansion of People Also Ask boxes, the addition of AI overviews - that reduce click-through rates on organic results by giving users what they need without requiring them to visit your site. It's about policy changes - the increasingly stringent requirements for E-E-A-T, the evolving definitions of "helpful content," the moving goalposts that mean what worked last year doesn't work this year and what works this year might not work next year. Each of these is, in its own way, a localized outage. A partial interruption in the service that you've built your business on. The difference is that an outage lasts forty-seven minutes and a core algorithm update lasts forever.
The Addiction
Here's the part that makes this genuinely tragic rather than merely cautionary. Everyone in the SEO industry knows about platform dependency. Everyone has read the articles. Everyone has heard the conference talks. Everyone has nodded along when someone says "diversify your traffic sources" and "don't put all your eggs in one basket" and whatever other folksy wisdom we deploy when we want to sound prudent without actually doing anything differently. Everyone knows.
Nobody diversifies.
Or almost nobody. I'm exaggerating for rhetorical effect, which I do often and which my editor (I don't have an editor, I am my own editor, which is its own kind of problem) would tell me to stop doing. Some businesses have genuinely diversified traffic sources. Some businesses get meaningful traffic from email, from social media, from referral links, from direct visits, from the whole rainbow of acquisition channels that exist beyond the monochrome of organic search. But most businesses - particularly the businesses that have invested heavily in SEO, the businesses that have SEO agencies and SEO consultants and SEO teams and SEO budgets - are overwhelmingly dependent on Google for their traffic.
I've seen the analytics. Across my client portfolio, the average percentage of total traffic from organic search is 67%. Two-thirds. For some clients, it's higher - I have an e-commerce client where organic search drives 82% of all sessions, which means that if Google sneezes this company catches pneumonia. These are not small businesses. These are not unsophisticated operations. These are companies with marketing departments and acquisition strategies and diversification plans that exist in PowerPoint decks and quarterly objectives and are undermined, daily, by the simple reality that Google traffic is too good to resist.
Because Google traffic is extraordinary. I need to say this, even though I've spent the last several thousand words explaining why depending on it is dangerous, because the reason we depend on it is not stupidity. The reason we depend on it is that organic search traffic has characteristics that no other acquisition channel can match.
It's free. Not zero-cost - SEO costs money, content costs money, technical optimization costs money - but free in the marginal sense. Once you rank, each additional visit costs you nothing. No cost-per-click. No cost-per-impression. No media spend. The traffic just arrives, day after day, like a river flowing past your property. You didn't build the river. You didn't pay for the water. You just positioned yourself correctly and the water comes.
It's intent-driven. People who find your website through Google Search are looking for something. Not browsing. Not scrolling. Looking. They have typed a query - a question, a need, a desire expressed in words - and your website has been presented as an answer. The conversion rates on organic search traffic are, for most businesses, higher than any other channel except direct traffic. Because the user arrived with intent. They wanted something. You appeared to have it. The transaction is half complete before they've loaded the page.
It's scalable. Social media traffic is proportional to your effort - you post, you get traffic, you stop posting, the traffic stops. Email traffic is proportional to your list - you can only send emails to people who've signed up, and the list grows slowly. Paid traffic is proportional to your budget - you can buy as much as you want, but every additional dollar of spend buys a little less because you're expanding into less interested audiences. Organic search traffic is proportional to your quality (for a given definition of "quality" that Google determines and periodically changes). Create a page that Google considers the best answer to a query that thousands of people search for every day, and you will receive thousands of visits every day, indefinitely, without additional effort. That's not marketing. That's a printing press. That's a money machine. That's the closest thing to passive income that digital marketing has ever produced.
It's measurable. You can track it. You can attribute it. You can see which queries brought which users to which pages and what those users did after they arrived. Not perfectly - Google has obscured much of the keyword data behind "not provided," which is a whole other article and a whole other grievance - but well enough to make decisions, to calculate ROI, to justify budgets, to prove that the investment is working.
Free. Intent-driven. Scalable. Measurable. These are the four properties of organic search traffic, and collectively they form one of the most compelling value propositions in the history of marketing. They are also, collectively, the reason that nobody diversifies. Because why would you? Why would you invest in a channel that costs money per click when you have a channel that's free? Why would you invest in a channel that reaches uninterested browsers when you have a channel that reaches active searchers? Why would you invest in a channel that requires constant feeding when you have a channel that generates returns for months or years from a single investment?
The answer, of course, is that you would invest in those other channels because the free channel is controlled by someone else and can be taken away at any time for any reason or no reason at all. But that answer requires you to weigh a hypothetical future risk against a concrete present benefit, and humans are terrible at that calculation. We always choose the present benefit. We always take the free traffic. We always lean into the dependency because the dependency feels so good.
It's an addiction. I don't use that word casually. I've written about SEO and anxiety before, about the compulsive checking and the cortisol spikes and the behavioral patterns that a clinical psychologist would find clinically interesting. But the addiction here is not to checking rankings. The addiction is to the traffic itself. To the growth curve. To the dashboard that shows organic sessions climbing up and to the right, that magical up-and-to-the-right line that is the universal symbol for "things are going well" and that is, in the context of organic search, a line drawn by a system you don't control and that can redraw it at any time.
When you're addicted to something, the prospect of not having it is more frightening than the knowledge that it's harming you. Smokers know cigarettes cause cancer. Alcoholics know alcohol destroys relationships. And SEO professionals know that organic search dependency is a strategic vulnerability. But the knowing doesn't change the behavior because the alternative - less traffic, slower growth, harder work, higher costs, the long and unglamorous grind of building direct relationships with customers instead of borrowing them from Google - is too painful to contemplate when the easy option is right there, available, free, one keyword research session away.
So we keep optimizing. We keep building on rented land. We keep pretending that the forty-seven-minute outage was an anomaly rather than a preview. And when the next outage comes, or the next algorithm update, or the next feature change that reduces our clicks by 20%, we'll panic for a few hours and write some tweets and then go right back to doing what we were doing, because the alternative is too hard and the present is too easy and the dependency is too deep to break with willpower alone.
What Actually Happened
I should tell you what actually happened during those forty-seven minutes, beyond the numbers and the phone calls and the existential dread. Because the experience of an outage is different from the analysis of an outage, and the experience is what I think about when I can't sleep, which is more often than I'd like.
For the first five minutes, nobody knew what was happening. The dashboards were showing anomalies but not zeros - the degraded service was sending some traffic, intermittently, in amounts that were low enough to trigger alerts but not low enough to suggest a complete failure. It looked, to the untrained eye, like a particularly aggressive algorithm update. Like the kind of traffic drop that happens when Google decides to re-evaluate a site's quality and temporarily suppresses its rankings while the re-evaluation is in progress. I've seen that before. I've lived through that before. The stomach-drop moment when the graphs go south and you don't know why and you won't know why for days or weeks because Google doesn't tell you why, they just do it, and you figure it out later, or you don't, and either way the traffic does what the traffic does and you are a passenger, not a driver.
So for the first five minutes, I thought it was an update. I thought it was targeted. I thought it was my clients. I thought it was me. This is the SEO professional's default assumption when something goes wrong: it's personal. It's about my sites. It's about my work. Google has looked at what I've done and found it wanting and is now punishing me, specifically, for some transgression I committed or failed to commit, some rule I broke or failed to follow, some piece of content that was too thin or too long or too optimized or not optimized enough. The narcissism of the SEO professional in crisis is breathtaking. We think we're important enough for Google to punish individually. We are not. Google processes 8.5 billion searches a day. It does not know who we are.
By minute seven, I checked Twitter. And by minute seven, Twitter had already diagnosed the problem, because SEO Twitter is nothing if not fast, and what I saw there changed the emotional register from personal panic to something larger and stranger. It wasn't me. It wasn't my clients. It was everyone. Every site. Every industry. Every geography. The entire organic search ecosystem, worldwide, was experiencing the same thing simultaneously, which is a sentence I had never expected to write and which still feels surreal months later, like describing a day when the wind stopped blowing everywhere at once.
By minute fifteen, the helplessness had fully set in. Not the productive helplessness that leads to problem-solving - the kind where you see a problem, acknowledge that it's beyond your control, and focus on what you can control instead. The unproductive kind. The kind where you sit in your chair and refresh a dashboard that shows zero and refresh it again and it still shows zero and you refresh it again because the act of refreshing is the only act available to you and doing something, even something futile, is more bearable than doing nothing. I refreshed my dashboards more times in those forty-seven minutes than I normally refresh them in a week. I was doing the very thing I've written about and warned against and lectured clients about, the compulsive checking, the anxiety-driven refreshing, and I was doing it with full awareness that it was pointless and irrational and that I would write about it later as evidence of my own hypocrisy, and I did it anyway, because self-awareness is not the same as self-control and never has been.
By minute thirty, something interesting happened. I stopped checking. Not because the anxiety had passed but because the anxiety had peaked and crested and begun to transform into something else. Acceptance, maybe. Or resignation. Or the kind of exhausted calm that descends when you've been scared for long enough that your adrenal glands simply run out of material. I sat in my chair and I looked out the window and I thought about what my business would look like if Google Search went away permanently. Not as an exercise. Not as a strategic planning session. As a genuine, full-bodied confrontation with the possibility that the thing I'd spent my career building expertise in could simply cease to exist.
The answer was not comforting. Without Google Search, I have no clients. Without clients, I have no revenue. Without revenue, I have - what? Twenty years of knowledge about a system that no longer exists? A set of skills - crawl analysis, content optimization, link equity evaluation, technical auditing - that are specific to a platform that isn't there? I'm a farrier in a world of automobiles. I'm a typewriter repairman. I'm an expert in a thing that doesn't exist anymore.
Obviously this was catastrophizing. Google didn't go away. Google came back in forty-seven minutes. My clients are fine. My career is fine. Everything is fine. But the forty-seven minutes gave me a window - a small, uncomfortable, unwelcome window - into a future that is not impossible, just improbable, and the difference between impossible and improbable is the difference between "don't worry about it" and "maybe you should have a plan."
I don't have a plan. I'm being honest about that.
The Diversification Lie
After the outage, the think-pieces arrived. They always do. The outage happened on a Friday, and by Monday the SEO blogs were full of articles with titles like "The Google Outage: Why You Need to Diversify Your Traffic Sources" and "Don't Put All Your Eggs in One Basket: Lessons from the Google Downtime" and "5 Traffic Sources Beyond Google That Your Business Should Be Investing In." (The "5" in that last title is optimistic. Most businesses struggle to maintain one traffic source, let alone five.)
I read them all because reading SEO blogs is part of my job and because I am a glutton for punishment. And they were all right. Every one of them was factually, strategically, objectively right. You should diversify your traffic sources. You should invest in email. You should invest in social. You should invest in direct brand building. You should invest in referral partnerships. You should not build your entire business on a single platform that you don't control.
They were right and they were useless. Because telling someone to diversify their traffic sources is like telling someone to exercise more and eat less. It's correct advice. It's obvious advice. It's advice that everybody already knows and almost nobody follows because following it requires sustained effort, delayed gratification, and the willingness to invest in channels that will return less, cost more, and scale slower than the one they're already dependent on.
Let me show you why the diversification argument fails in practice. Say you're a mid-market B2B SaaS company. You're getting 10,000 organic visits a month. Your conversion rate on organic traffic is 2%. That's 200 conversions a month. Your average deal value is $5,000. Organic search is generating $1 million a month in pipeline. Not revenue - pipeline. But pipeline is what sales teams live on, and $1 million a month is enough pipeline to keep the lights on and the investors happy.
Now say you want to diversify. You want to reduce your dependency on organic from 70% to 50%. You need to replace 3,000 monthly visits - the visits that would take you from 70% organic dependency to 50% - with traffic from other sources. Where does that traffic come from?
Email: You'd need a list of approximately 60,000 subscribers (assuming a 5% open rate and a click rate that gets you to the site). Building that list takes years. Maintaining it takes staff. Running campaigns takes time and tools and copywriting and segmentation and all the other things that email marketing requires, which are considerable, because email marketing is a discipline, not a side project, and treating it like a side project is how you end up with a list full of unengaged contacts and a sender reputation that Gmail has quietly flagged as "meh."
Social: LinkedIn might get you there, if you're B2B, if you post consistently, if your content resonates, if the algorithm doesn't decide to show your posts to twelve people and your CEO's golfing photos to twelve thousand. Social media traffic is volatile. It spikes when a post goes viral and collapses when it doesn't. It is, in its own way, as platform-dependent as organic search - you're just trading one landlord for another, and the new landlord (LinkedIn, Facebook, Twitter, whatever platform is currently in favor) is just as capricious and just as opaque about its algorithmic decisions.
Paid: You could buy 3,000 visits. At, let's say, $5 per click for a mid-market B2B keyword (which is optimistic, for some industries the CPC is $50 or more), that's $15,000 a month. $180,000 a year. To replace traffic you were getting for free. The CFO will love that conversation. "Hey, you know that traffic we were getting at no marginal cost? I'd like to start paying $180,000 a year for an equivalent amount, as a hedge against a forty-seven-minute outage that has happened once in twenty years." Good luck.
Direct/brand: This is the real answer. Build a brand so strong that people type your URL directly. Build a reputation so solid that people recommend you by name. Build a product so good that people come back without being told to. This is the right strategy. It's also a five-to-ten-year strategy, and it requires a level of investment and patience and brand-building sophistication that most companies don't have and most marketing departments aren't structured to deliver.
Every path to diversification is harder, slower, and more expensive than the thing it's supposed to replace. That's why nobody does it. That's why the think-pieces get written and get shared and get nodded at and get filed away and nothing changes. Because changing is expensive and the status quo is free and the outage only lasted forty-seven minutes and it probably won't happen again and even if it does we'll deal with it then and besides, have you seen our organic traffic numbers this quarter? They're up 23%. Up and to the right. The line goes up and to the right. Everything is fine.
The Algorithm Update Is a Slow Outage
I want to make a connection that I think most people miss, because they think about outages and algorithm updates as different categories of events. They're not. They're the same event at different speeds.
An outage is sudden. Total. Visible. Your traffic goes to zero and you notice immediately and you call your SEO consultant and you ask "what's happening" and you get an answer. An algorithm update is gradual. Partial. Often invisible until you go looking for it. Your traffic drops 30% over two weeks and you might not notice for another two weeks after that, depending on how closely you monitor (and I've already written about the pathology of monitoring too closely, so let's split the difference and say you notice at a reasonable interval).
But the effect is the same. In both cases, Google made a decision - to go down, or to change how it ranks pages - and in both cases, you had no input, no warning, no recourse, and no choice but to accept the outcome. In both cases, the traffic disappeared. In both cases, the revenue was affected. In both cases, you were reminded, with varying degrees of gentleness, that you don't own your traffic. You borrow it.
The September 2023 Helpful Content Update killed sites. I don't mean it hurt them. I mean it killed them. Websites that had been receiving hundreds of thousands of organic visits per month - websites with real audiences and real revenue and real employees - saw their traffic drop by 80%, 90%, sometimes more. Overnight. Without warning. Without explanation. Without recourse. Google published a blog post explaining what the update was "designed to do" using language so general that it could apply to literally anything, and then stopped talking about it, and the affected site owners were left to figure out what had happened and whether they could fix it and in many cases the answer was: you can't fix it, because there's nothing to fix, because you didn't break anything, Google just changed its mind about whether your content is "helpful" and there's no appeal and there's no explanation and there's no person you can talk to and the traffic isn't coming back.
That's not a forty-seven-minute outage. That's a permanent outage for specific sites. A selective, ongoing, indefinite disruption of service that looks different from a technical outage but feels the same to the people affected. The graphs go flat. The revenue disappears. The phone stops ringing. And the cause is the same: dependency on a system you don't control.
I've worked with sites hit by algorithm updates. The conversations are identical to the conversations I had during the outage. "What happened?" "Is it our fault?" "What can we do?" And the answers, depressingly, are often the same. Google changed something. It might be your fault or it might not. We can try things and hope they work. There's nothing we can do that's guaranteed to fix it. There's nothing we can do.
That phrase again. There's nothing we can do. The most honest sentence in SEO. The sentence we never put in our proposals or our case studies or our conference presentations. The sentence that would make every client in every industry reconsider their organic search investment if they heard it often enough. We can optimize. We can improve. We can build the best possible case for Google to rank us. But we can't make Google do anything. We can't compel Google to send us traffic. We can't enforce a right to rank. We can ask. We can hope. We can do our best.
And then Google decides.
The Honest Conclusion
I want to end this with something useful. Something actionable. Something that justifies the time you've spent reading these several thousand words about a forty-seven-minute event that happened months ago and that most people have already forgotten. I want to give you a framework, a strategy, a set of concrete steps that you can take to protect your business from the dependency that I've just spent twenty pages describing.
But I don't have one. Not really. Not honestly.
I have the standard advice: diversify your traffic sources. Build your email list. Invest in brand. Create direct relationships with your customers. Don't depend on any single platform. These are all correct. They are all things you should do. They are all things that I, an SEO consultant whose entire livelihood depends on the continued existence and relevance of organic search, tell my clients to do, because I care more about their long-term success than I do about my short-term relevance, or at least I like to think I do, on my better days.
But I also know that most of them won't do it. And I know why. Because organic search traffic is too good. Too free. Too easy. Too effective. The drug is too pure. The dependency is too deep. The forty-seven minutes weren't long enough to change behavior. An outage changes behavior when the pain is severe enough and sustained enough to overcome the comfort of the status quo, and forty-seven minutes of zero traffic is painful but not sustained, scary but not scarring, a warning but not a wound.
What would change behavior? I don't know. Maybe an outage that lasts a day. Maybe an algorithm update that wipes out an entire industry vertical. Maybe AI search actually replacing traditional search in a way that makes organic rankings irrelevant. Maybe something I can't predict, because the nature of dependency risk is that the disruption comes from a direction you weren't looking, at a time you weren't expecting, in a form you hadn't imagined.
What I know is this: on August 9th, 2024, for forty-seven minutes, every SEO professional in the world was reminded that they are building on rented ground. That the foundation of their career, their clients' businesses, their industry's existence is a product built by someone else, maintained by someone else, and controlled by someone else. That the traffic they've spent years optimizing for can disappear in the time it takes to make a cup of coffee, and that when it disappears there is nothing - absolutely nothing - they can do about it except wait and hope that it comes back.
It came back. In forty-seven minutes, it came back. The graphs recovered. The traffic returned. The clients calmed down. The think-pieces were written and the hot takes were tweeted and then everyone went back to work, back to optimizing, back to building on the same rented ground as before, because what else are you going to do? Where else are you going to build? The ground is free and the building is familiar and the landlord has been reliable for twenty-five years and forty-seven minutes of unreliability isn't enough to make you move.
I went back to work too. I'm being honest. I checked my clients' rankings the next morning. I looked at the traffic dashboards. I confirmed everything was back to normal. I breathed a sigh of relief. I moved on. I went back to doing exactly what I'd been doing before the outage, which is optimizing for Google, depending on Google, building on Google's land, with full knowledge that the land isn't mine and never was.
I know it's rented. I know the lease has no term. I know the landlord can change the terms whenever they want. I know all of this. And I keep building anyway, because the building is what I know, and the ground is where the traffic is, and the traffic is what the clients need, and the clients are what pay the mortgage, and the mortgage is what keeps the roof over my children's heads, and if that's dependency then fine, it's dependency, and I'll manage it the way you manage any dependency, which is one day at a time, with your eyes open, hoping the thing you depend on keeps working.
Forty-seven minutes. That's how long it took for the whole thing to crack. And that's how long it took for us all to plaster over the crack and pretend it wasn't there.
The crack is still there. It's always been there. We just choose not to look at it, because looking at it means acknowledging something that no amount of optimization or diversification or strategic planning can fix, which is that we built an industry on a product we don't own, and the product can be turned off, and we can't stop it.
The outage lasted forty-seven minutes. The lesson should last forever.
I give it a week.