Possibly Approaching Super-Exponential

Two weeks ago I wrote that this site had reached roughly ten thousand daily organic impressions in ninety days. That number is now out of date, and the way it went out of date is the reason for this piece.

The last seven complete days averaged 14,590 impressions a day. The single best day, 15 September, was 18,736. The figure in the earlier article was 4,804 a day. So the traffic is three times what it was six weeks ago.

That is not the interesting part. Tripling is just a big number. The interesting part is the shape.

What the weekly totals actually say

Week beginning	Impressions	Ratio to prior week
22 August	42,527	—
29 August	58,210	1.37
5 September	86,894	1.49

A curve growing by a constant multiple every week is exponential. That alone is what most people mean when they say hockey stick. It is a strong result and it is also an ordinary one.

What we appear to have is not that. The multiple itself went up, from 1.37 to 1.49.

Why that distinction matters

Take the natural log of each week's ratio and you get the growth rate itself: +0.314, then +0.401.

In exponential growth that number holds steady. It is the constant in the exponent. When it rises, the exponent is not a constant any more, and the curve is steepening faster than exponential. That is what the phrase super-exponential describes.

Two increments are not a proof of anything. I want to be plain about that before anyone quotes this. But the direction is the whole question, and the direction is up.

The climb, in plain dates

The first day this site recorded more than ten thousand organic impressions was 28 August: 10,292. Eighteen days later, on 15 September, it recorded 18,736. That is an increase of 82 percent, and the article reporting the ten thousand figure was barely a week old when it was overtaken.

The run into that peak was not smooth, and nobody should pretend otherwise. There were down days inside it. But the direction across those eighteen days is not in dispute.

Why this stops at 15 September

Every figure here ends on 15 September. What Search Console reports after that date is, I believe, wrong, and the useful part is that it can be shown to be wrong using Google's own columns.

Search Console reports four things each day: clicks, impressions, click-through rate and average position. Three of them are measured. The fourth, click-through rate, is just clicks divided by impressions. So the four have to agree with each other, and when they do not, one of them is broken.

The three numbers that have to hold together

Average position did not move. It was 6.8 across 8 to 15 September and 6.6 across the days that followed. If anything it improved slightly. No ranking event happened.

Clicks did not move either. They stayed in the same band they had held all month, within about fifteen percent of the prior week's daily average. Real people kept arriving on the site at close to the same rate.

Impressions, alone, fell away sharply.

Now put those together. Click-through rate is a function of where you rank. Rank in the same place, get the same share of clicks. But with clicks roughly flat and impressions falling, the reported click-through rate had to climb, and it did, by around half again, ending at more than double the rate recorded on 15 September.

That is the contradiction. A site does not double its click-through rate in three days while its average position sits still. There is no mechanism for it. The only way the arithmetic works is if the impression count is too low, because impressions are the denominator and the denominator is the only term that can be wrong without anything real changing.

A second engine says the same thing

Bing was not reporting a decline in that window at all. Across 11 to 17 September it recorded an average of 377 impressions a day for this site against a hundred-day average of 105, running at more than three times its own baseline while Google's figures fell away.

Two independent engines cannot both be right about the same days. And Bing agreeing with Google's own clicks, against Google's own impressions, settles which column to doubt.

What I think is actually happening

Search Console runs two to three days behind by design. That is documented and it is not a complaint. But anyone who has watched this tool across a few sites for long enough has seen it do something else: get stuck, under-report impressions for a stretch, and then catch up days or sometimes a week later. I have seen it on ExecSearches.com more than once over twenty seven years.

I think that is where this is. The traffic did not go anywhere. The counting did.

So the last number I will stand behind is 18,736 on 15 September, and everything above stops there. When Search Console catches up I will publish whatever it says, including if it proves me wrong.

What I am not claiming

I am saying possible, and the word is doing real work.

Three weekly totals give two growth increments. Two increments establish a direction, not a law. The category itself is new, which means some of this is a market being measured for the first time rather than a market growing. Search volume is seasonal and September is a hiring month. And a curve of this shape cannot continue, because nothing does. Something will bend it.

What I will defend is narrower and I think more useful: over the period measured, the growth rate was itself increasing. That is a statement about three numbers, and the three numbers are above.

Where it is coming from

Not from a single page. The board now carries 844 live positions in governance, risk, compliance, privacy, audit and AI governance, each one read by a person before it published. Around them sit role guides, job description templates, certification material and practice questions, all cross-linked.

A structure like that compounds in a way a flat list does not. Every new page gives the existing pages somewhere to point, and every page that earns a position lends a little of it onward. Add a page to a list and you have a longer list. Add a page to a web and you change the web.

That is the honest mechanism behind a rising growth rate, and it is also why the rate cannot rise forever. The web is not infinite, and eventually a category runs out of questions.

An invitation, and I mean it

I studied mathematics at Purdue and I did proofs there. I am aware that a claim like this is exactly the kind that people make loudly and support thinly, which is why every number in this piece is one you can put in front of somebody.

If you think the reading is wrong, I would genuinely like to hear it. Bring your professor. Bring the book. And if you have numbers I do not have, I want to see them. Wanting the numbers is the entire point.