Multifamily diligence has a gravitational center, and it is the T12: twelve months of actuals, the document every buyer normalizes, every lender reads, every broker leads with. The T12 deserves its position. It is also, by construction, a rearview mirror, a record of what the previous owner's decisions earned, and the underwriting question is never what the property earned. It is what it can earn under your decisions, and the answer to that question lives in the unit mix.
Mix is strategy, not inventory. Studios, one-bedrooms, and two-bedrooms are different products with different renter pools, different demand depth, different turnover behavior, and different reposition economics, and a property's mix is a fixed bet on a specific slice of the rental market. Underwriting it means asking whether the bet fits the submarket: a two-bedroom-heavy property in a young single-renter market and a studio-heavy building near family demand are both mismatches wearing full occupancy, because occupancy hides mismatch until the market softens and the mismatched units go soft first.
Comps inherit the same logic. Market rent is not one number for the property; it is a number per unit type, sometimes per floor plan, and blending them produces the same distortion blanket assumptions produce everywhere else. The practical failure looks like this: the pro forma carries a healthy average market rent, the one-bedrooms genuinely support it, the two-bedrooms do not, and the reposition budget gets spent renovating units into a rent the comp set never actually promised for that type. Per-type comps are more work. They are also the difference between a rent thesis and a rent average.
Sequencing is where mix meets the calendar. A reposition plan does not renovate the property, it renovates units, in some order, at some pace, and the order matters: leading with the unit type carrying the widest spread to market front-loads the return on capital, while renovation pace has to respect what the team can execute and what the market can absorb without the property competing against its own vacant renovated inventory. The plan's economics are the sum of per-unit decisions arranged in time, which is exactly the structure a blended model cannot hold.
The T12 still opens the analysis, normalized honestly: strip the previous owner's anomalies, find the real expense run rate, reconcile the rent roll against the collections. Then the work turns forward, and forward is unit-level. PropCalc™ models multifamily with the mix, the per-type rents, and the reposition plan carried through a full capital stack and waterfall, and there is a fictional multifamily deal on this site, linked below, built to realistic norms. Open it and change the reposition pace. What that single input does to the returns is the entire argument of this page, in one slider.