Around the fifth and sixth of August the ranking volatility trackers went red, and the sequence that followed was so precisely choreographed that you could have set a watch by it.
First the trackers spike. Then the forums fill with people posting screenshots of their own graphs, which are the same graph. Then somebody with a large enough following gives it a name, and the name is always either a month or an animal. Then, by the afternoon, agencies are sending clients an email about the August update, and the email contains the phrase "we are monitoring the situation closely," which is a sentence that has never once been false and has never once been informative.
I have been through perhaps a hundred of these. I have sent that email. I am not above any of this.
What a Tracker Actually Is
Let us be unglamorous about the instrument for a moment, because almost nobody who cites it has thought about what it physically is.
A volatility tracker watches a fixed sample of keywords across a fixed sample of domains and reports how much the positions moved in aggregate compared to a baseline. That is it. That is the entire mechanism. It is a motion detector.
Consider what that instrument cannot tell you, structurally, no matter how good the vendor is. It cannot tell you why anything moved. It cannot tell you whether Google shipped one change or forty in the same week, which is the normal state of affairs. It cannot distinguish a ranking system change from a spam enforcement action from an index refresh from a serving infrastructure hiccup from a seasonal shift in what people are searching for. It cannot tell you whether the movement touched your vertical, because your vertical is probably not in the sample in any meaningful proportion. And it cannot tell you whether the thing that moved is a thing you would want.
What it can tell you is that the needle moved. That is a real signal and it is worth having. It is a weather station. It is not a diagnosis, and the entire industry reads it as a diagnosis.
Why We Keep Naming Things
The naming habit is a fossil. There was a genuine period, roughly 2011 through 2016, when Google announced discrete, named, individually meaningful updates. Panda was a thing. Penguin was a thing. They had launch dates, they had targets, and you could sensibly ask whether you had been hit by one. Naming them was correct.
That world is gone. Google now ships continuously, most changes are never announced, the announced core updates roll out over weeks and overlap each other, and the systems interact in ways that make a single causal story almost always wrong. Asking "what was the August update" is like asking what the traffic on the motorway was about. There was no about. There was just traffic, produced by ten thousand independent decisions, and the only reason it has a name is that naming it makes it discussable, and discussable things can be sold.
Because that is the actual function. Volatility is an excellent product. It sells tool subscriptions, because you need the tracker to see the thing the tracker told you to worry about. It sells emergency audits. It justifies a retainer in a quarter when the retainer was looking hard to justify. It gives an account manager something urgent to say on a call. I am not accusing anyone of fraud. I am pointing out that an entire economy has grown up around an indicator, and economies that grow around an indicator tend to keep that indicator loud.
What I Do When a Client Panics
The call comes in and someone says they have heard there was an update and are we affected. Here is the sequence, and it is deliberately boring.
First, I ignore the trackers entirely and open their data. Not rank tracking. Search Console, impressions and clicks, segmented. The question is not "did the internet move," it is "did we move, where, and by how much relative to how much we normally move."
Second, I check normal variance. Most sites have a weekly rhythm and a seasonal rhythm, and a startling number of update panics are a Tuesday being compared to a Sunday, or August being compared to June in a B2B vertical where the entire buying population is on holiday. If the movement sits inside the band the site normally wanders around in, there is no event to explain.
Third, and this is the one people hate, I check whether we did it to ourselves. A deploy went out. Someone changed a robots directive. A CDN rule started serving something strange to a crawler. A template lost its canonical tags. A migration that happened six weeks ago finally finished being recrawled. In my experience the single most common cause of a client-visible traffic drop that coincides with an industry-wide volatility event is an unrelated self-inflicted wound that nobody would have noticed if the forums had been quiet. Every company has a graveyard of old decisions and they do not stay buried on a schedule.
Fourth, only if all of that is clean, I segment by query type, page type and intent, and look for a pattern with a shape. Real ranking system changes usually have a shape. They hit a category of page, or a category of query, or a class of site. Undifferentiated movement across everything is usually noise or is usually infrastructure.
Fifth, and most importantly, I do nothing for a while. This is the hardest advice to sell and the most consistently correct. Core updates take weeks to settle, positions oscillate during rollout, and the number of times I have watched a company make a panicked structural change in week one and then be unable to tell whether the recovery in week four was the fix or the settle is genuinely too high to count. If you change five things during a rollout you have not fixed your site, you have destroyed your ability to learn anything from the event.
The Boring Conclusion
None of this means updates do not matter. They do, some of them are brutal, and sites die. It means the tracker going red is not information about you, the name somebody gave it is not information about anything, and the gap between those two facts is where a lot of money gets spent badly.
I have written before about what compulsive rank checking does to a person, and the industry version of that compulsion is the volatility tracker, refreshed hourly, by people who cannot act on what it says.
In a volatility week, the calmest person in the room is usually the one who checked their own data first.