You open the breakdown expecting a distribution and you get a single line. One asset class holding effectively all of the book, the rest at zero or close to it. On a strategy that was described to its investors as multi-asset, that single line is the most important thing on the screen, and it is worth more attention than any return figure sitting above it.
The finding is not that concentration is bad. Concentrated books are often the good ones. The finding is that the book being run and the book being described have separated, and the gap between them is the thing that gets litigated later, in a review, under pressure, when a position has already gone wrong.
Rule out the view before you call it a finding
Everything on the performance tab is conditioned on two controls, and a single populated row is a conclusion only after both have been checked.
The view selector runs ALL ACCOUNTS (REAL), BLOCKCIRCLE ONLY, MTE, PAPER, MANUAL and WALLET, with a cumulative all accounts row above the individual accounts. Those are different populations. A breakdown drawn on BLOCKCIRCLE ONLY says nothing about assets held elsewhere. A breakdown that includes PAPER is mixing a simulated account into a picture of real exposure. If the equity book lives at one broker and the crypto lives in wallets, selecting the wrong view produces a single class result that is an artefact of the selection rather than a fact about the mandate.
The period controls run 1D, 7D, 14D, 1M, 3M and 1Y, and they matter more than they look. A breakdown drawn over a short window on a book that trades will reflect what was held recently rather than what the strategy holds structurally. If the answer changes materially between 1M and 1Y, the concentration is a current state rather than a standing characteristic, and those two require different conversations.
Cross-check against the holdings side
The performance view is a derived picture. The holdings side is closer to the raw position set, and it is the faster corroboration. On the book in this account the tiles read 4 unique positions against 8 connected sources, and the 24 hour mover tile shows the same stablecoin, USDC, occupying both the best and the worst slot at 0.00 percent.
That combination is diagnostic. When one stablecoin fills both ends of the mover tile, there are not enough instruments with a moving price in the book to populate both. Four unique positions, one of which is behaving as cash, is a book with two or three real bets in it. If a breakdown shows a single populated class, and the holdings tiles show four unique positions across eight sources, the two agree and the finding stands.
If they disagree, the disagreement is the more valuable output. A holdings table showing instruments across several markets underneath a breakdown showing one class means either the classification layer is mapping instruments into the wrong buckets, or the breakdown is drawn on a narrower population than the holdings view. Both are worth resolving before any number from either screen goes into a document.
Deliberate concentration versus unfilled intent
Once the finding survives those checks, it splits into two cases that look identical on screen and are completely different in substance.
Deliberate concentration is a manager who has decided, with reasons, that the opportunity set in the other classes does not justify capital right now. The evidence for that reading is a documented process that covers the other classes, a pipeline of ideas that were examined and rejected, and stated conditions under which capital would move. A concentrated book with that behind it is a manager doing their job.
Unfilled intent is a book that is concentrated because nothing else was ever built. No coverage, no pipeline, no rejected candidates, no stated conditions. The mandate promised breadth, the operation delivered one bet, and the diversification the risk framework assumes simply is not there. This is the case that damages a firm, because the framework will keep producing risk numbers that depend on correlations between exposures that do not exist.
The question that separates them is short and you should be able to answer it about your own book in one sentence per class: what did we look at in this class this quarter, and why did none of it clear. A manager who can answer that has a concentrated portfolio. A manager who cannot has a single strategy with a multi-asset description on the front.
What the allocator will actually ask
Assume the finding reaches an allocator, because eventually it does. Four questions arrive, in roughly this order, and each has a right and a wrong way to be met.
- Is this what we agreed. The answer refers to the mandate document, not to performance. If the mandate permits the concentration, say so and cite the clause. If it does not, saying so first is enormously better than being told.
- Is it deliberate or is it drift. This is where the pipeline evidence goes. Volume of ideas reviewed per class, and the conditions under which capital moves.
- What happens to the risk numbers. A book with one populated class has no meaningful diversification benefit in it, so any risk figure built on cross-class correlation is describing a portfolio you do not have. Recompute the standard measures on the actual position set and present those instead of the framework defaults.
- What is the path back. Either a deployment timeline with dates and target weights, or an explicit request to amend the mandate to match how the book is genuinely run. Both are respectable. Continuing to report against a mandate you are not following is not.
The failure mode in all four is treating the question as an attack. It is a governance question with a documentary answer, and the manager who has the documents ready converts a difficult meeting into a routine one.
Documenting the gap so the next review has a baseline
Whatever the resolution, the finding needs to leave a record, because the value of catching it is mostly in the comparison to next quarter.
Record the breakdown as it stood, with the view and the period that produced it stated alongside, so a future reader can reproduce exactly the same picture rather than a similar one. Record the corroborating position count and source count from the holdings side on the same date. Record which of the two cases you concluded, and the evidence you relied on.
Then set the review cadence to something that will actually catch the drift. Monthly on the composition, quarterly on the mandate comparison, and an immediate review whenever the number of populated classes changes in either direction. A book going from one populated class to three is as much a governance event as the reverse, because it usually means capital moved into markets where the process has not been exercised in a while.
The discipline worth keeping is that the breakdown is read as a statement about the mandate rather than as a chart. Every quarter, one sentence: this book is currently a single class book, deliberately, for these reasons, with these conditions for changing. When that sentence is written down before anyone asks for it, the finding stops being a problem and becomes part of the record.