Your portfolio already has a benchmark. You just haven't seen it.

No — under one Technology label, the seven have seven distinct exposure profiles. Idiosyncratic share spans 29% to 63% across the cohort, sorted below by the stock-specific component (the layer a stock-picking mandate is paid for):

The dominant subsector differs for each. A PM targeting semiconductor exposure has a direct instrument pairing (NVDA or SMH). A PM targeting systems-software exposure does not get it from AAPL, AMZN, or NVDA — only from MSFT or IGV. The cascade makes the mapping from thesis to instrument explicit for each position.

Counterintuitive result. The idiosyncratic layer collapses from 46% to 7% as cross-stock correlation diversifies stock-specific risk away. The market layer stays at 41% — systematic risk is not diversifiable within a long-only sector-concentrated book; SPY overlay remains the only instrument that moves it.

The reframe: an equal-weight Mag7 book diversifies away the layer the mandate is paid for. For a PM evaluated on residual return, equal-weighting across the seven hands back the 39pp of idiosyncratic budget that made the individual positions interesting in the first place.

This is the difference between the seven names and the seven names as a book. Holding them individually preserves the stock-picking surface (29–63% of each name's variance). Equal-weighting blends those independent residuals against each other; the cross-correlations cancel out the part you wanted to harvest.

A concentrated allocation to one or two of the names (where you have a specific thesis) keeps the idio exposed. A diversified allocation across all seven turns the cohort into something closer to a sector ETF — which is fine if that's the thesis, but it's not what most PMs think they're doing when they buy "Mag7."

Two distinct read-outs, not one.

1. Per-stock Risk DNA. For each Mag7 position, the cascade reads market → sector → subsector → residual. The subsector is the named exposure (FDN, IGV, SMH, XRT, CARZ, RSPT) — the layer where the thesis lives. The residual is what you're being compensated for if the thesis works.

2. Cohort-level diversification math. When the cohort is held as a basket, the residual-layer collapse from 46% to 7% should be visible — that's the warning that "Mag7 ETF" and "concentrated Mag7 positions" are different exposures even though they hold the same seven tickers. The reporting needs to surface that they're different, not paper over it.

Practical implication for a long-only book: if the residual portion of Mag7 is what your thesis is about, don't equal-weight. Concentrate where the thesis is sharpest, let the others be smaller positions or pure-systematic vehicles (SMH, IGV, FDN). If the thesis is just "tech beta," the cohort is the wrong tool — XLK or QQQ is honest about what's being bought; the cohort framing has 8% lower expected residual rent for the same realized factor exposure.

The two-step framing — per-stock DNA + cohort math — is what separates a stock-picking program from a sector-overweight tilt.