Taking the direction out of a flow figure leaves a measure of pure activity. It answers how much money is moving through these funds while taking no view at all on which way any of it went.

That makes it a reading on engagement rather than on sentiment, and the two disagree more often than not. Some of the quietest periods for net direction have been among the busiest for sheer activity, with money arriving and leaving in similar quantities.

What it actually measures

High activity means the funds are being actively used rather than merely held. Positions are being taken and unwound continuously, which makes the complex easier to move in and out of at size.

Low activity describes neglect rather than contentment. Money that is neither arriving nor leaving is money that nobody is currently thinking about, whatever price happens to be doing.

It rises around events regardless of their character. Good news and bad news both produce activity, which is precisely what this measure is built to capture and precisely why it says nothing of direction.

Sustained high activity is different from a spike. One is a change in how the complex is being used and the other is a single event working through it.

Engagement and conviction are different conditions

Heavy activity gets read as strong interest, which is fair, and then as bullish interest, which is not. The measure has had the direction removed on purpose.

What it is genuinely good for is spotting when the complex has gone quiet. Extended low activity has preceded both continued neglect and abrupt re-engagement, and knowing which state you are in is worth something even when the reading says nothing about direction.

Record Days: a ranked list of extremes is a chart of events, not of behaviour

Extremes cluster rather than scattering evenly through the record. Sessions of this size tend to arrive in groups around a single event rather than turning up one at a time.

Each bar corresponds to something that actually happened on the day: an announcement, a shock, a rebalancing. The list is therefore a record of external events at least as much as it is a record of investor conduct.

That makes it useful for calibration and poor for inference. Knowing how large a single session can get is worth having in mind; treating the ranking itself as a pattern is reading a list of accidents as though it were a trend.

What it does not tell you

It scales with the size of the complex, so a reading from the early months and one from today are not directly comparable however similar the figures look.

Direction is removed by construction, so this can never say anything about appetite.

A single large institutional order produces a high reading from exactly one decision, which is not what the word activity suggests.

It nets nothing at all, so a day of heavy buying and a day of heavy selling produce an identical reading.

It cannot separate activity from turnover. Money arriving and leaving the same week counts twice and describes no change in position.

How to read it

High. A large amount of money is moving through the funds.

Moderate. Activity in line with what this complex ordinarily sees.

Low. Little money is moving in either direction.

On the ETF dashboard, Flow Size runs next to Record Days, Traded Volume and Weekly Flows.

Common questions

Why strip out the direction?

Because the question here is the quantity in motion, not which way it went. Busy and keen are separate conditions, and only the first of them is being counted.

What does low activity mean?

Neglect rather than contentment. Money that is neither arriving nor leaving is money nobody is currently thinking about, whatever price is doing.

Does heavy activity mean buying?

No, and taking it that way is the usual error. Direction was stripped out on purpose, which leaves the measure unable to say anything about appetite at all.

Are old and new readings comparable?

Not directly. The figure scales with the size of the complex, so a level that was extraordinary at launch may be ordinary now.

Can one order produce a high reading?

It can, easily. A single large institutional order moves a large amount of money on one decision, and nothing in this measure distinguishes that from a hundred smaller ones.

Why put both directions on one scale?

Because the two sides are rarely symmetric on this complex, and how far each of them reaches is the comparison worth making. Separate scales would conceal precisely that.

What does a ranked list not show?

Sequence. There is no time axis on the bars, so nothing about when these sessions happened or how close together is visible from the ranking itself.