Maybe you could enlighten the OP, then. How would you perform a H&BU analysis on one, unsubdivisable site with two separate zonings - not overlays. It seems you guys are WAY out of my league.
Two separate zonings doesn't mean 2 separate properties.
Analyzing for the contributory value of each of the components can be pretty straightforward if you have data for each.
So let's say there are 2 residential units + 2 storefronts equaling 3000sf between them. The typical buyer would be buying the property for the income and the financing would (probably) be under non-residential terms. On an income basis an appraiser could develop a price/rm or price/unit value for the residential component by analyzing 5+ units (due to their non-res classification) in a comparable size range, and analyzed for the multi-unit residential components the same way - by looking for multi-tenant residential in that overall size range.
You could even blend the cap rate you're using based on the contributory of each component to the net income. In the example above, if the if the residential use contributed 65% of the net and the commercial only added 35% of the net and these components had different cap rates you could built the blended net that way.
.65 x .0500 = .0325. (for the residential component)
.35 x .0750 = .0263 (for the multi-retail component)
Overall Rate .0588 / NOI = Value by Income
That's not the only way to do it nor is it necessarily even the way investors in that market are operating - it always depends and we always have to look. But it's one way. You can do similar with the value indicators in a Sales Comparison
$30,000/rm x 6 rentable rooms = $180,000 (for the residential component)
$200/sf x 1500sf = $300,000 (for the commerical component
Value by Sales = $480,000
Again, it would be better if you had a good selection of 2res+2ret sales comps to use as direct comparables and just do a price/sf, but IRL these are non-res propertes; not SFRs. So we never get datasets like that. Hence the utility of sometimes switching over to an alternate methodology.
We always have to work with what we have, as opposed to what we wish we had.