I've been pondering the scalar possibilities of selling Magic cards as a business model.  

Scalar businesses are models in which return on investment increases non-linearly in respect to cost.  Cost can be time or effort, or anything else that's a finite resource.  Consider a typical retail model in which a middle man purchases large lots of [product] from a wholesale outlet and sells them individually, at a markup, to retail customers.  

Simply, by subtracting the cost of goods sold and their operating expenses (staff, utilities, physical plant, etc.) from their revenue, you see the business's profit.  If they buy [product] at $x.00 and sell [product] at $(x*2).00, then their margin per [product] is $x.00.  If their overhead is $(100 * x).00, they will need to sell 100 [product]s to break even:

for x = $10.00,

Cost of Goods Sold:  $1000.

Operating Expenses: $1000.

Net Sales (100 units): $2000.

Profit: $0

At 100 products, the theoretical business breaks even.  Let's say they sell 200 units:

Cost of Goods Sold:  $2000.

Operating Expenses: $1000.

Net Sales (100 units): $4000.

Profit: $1000

Now we're making money!  The problem is that our profit increased $1000, but so did our expenses.   What if we extrapolate to 1000 units?

Cost of Goods Sold: $10,000.

Operating Expenses: $1000.

Net Sales: $20,000.

Profit: $9000.

After we deduct operating expenses, which we're assuming to be a fixed cost (it rarely is), we see that our costs grow linearly with our revenue.  This is a Bad Thing.  Your overhead will also inevitably shoot up.  Look at a store like Best Buy, or any other retail operation.  More sales imply more customers in the store, which require more staff, more floor space, a larger physical plant, more electricity, more HVAC, and so on.  The model scales poorly, but they make up for it with volume.

~

Selling cards has a similar issue of scale.  Take a retailer like Star City Games, which is probably the closest analogue to Best Buy in our industry.  Their value propositions are in-stock product, a trusted brand name and excellent customer service.  Their prices are rarely the lowest, but they are a safe, known quantity.  As they sell more cards, their inventory grows and so do their operating costs and cost of goods sold.  They are, presumably, a very profitable enterprise, so this is not an issue for them.  Their executive staff know how to run a business of scale, but this model just doesn't translate to the individual trader.

How does the math change when a bigger warehouse isn't an option?  When hired help isn't an option?  How do you scale a model once a single employee is a non-negotiable factor?  Most traders do this for fun, or as ancillary income, and they do not want to deal with expansion, hiring, or anything of the sort. We have to assume that, other than a friendly helping hand, the trader is doing all the work by hand.  To figure out where to get the most efficient use of time and capital, we need to analyze our expenses.

Shipping scales on a transaction basis;  units per transaction & number of transactions.  It costs more money to ship more cards, and the more times you ship cards, the more shipping fees you pay.

Transaction Fees scale on a dollar amount basis.  CardShark, eBay, and other services all take a cut of your revenue with no regard to your other costs.  This is the convenience fee for access to an efficient marketplace.   Selling Standard singles requires a marketplace, since there are so many people selling these cards.  You need to put your listing in front of the most eyes at the lowest prices.  I call this "grinding retail".  The more cards being sold, the greater need there is for a market infrastructure.

Storage and Supplies scale on a inventory quantity basis.  Binders, boxes and sleeves scale linearly with the quantity of cards you own, but there may be some marginal savings when buying these in bulk.  These savings probably don't amount to enough to factor into the model.

Time scales on a transaction basis and an inventory quantity basis.  This is the real killer; the reason I decided to "cut the fat" on my retail operation and move to a more agile model (backpack trading) had everything to do with my return on time invested.  Any time not spent selling or buying inventory was wasted time.  When you're plunging a clogged toilet at 10pm on a Friday, you're not making money.  Sure, there's the value of having a clean store so you can attract customers that will ostensibly sell or buy cards, but that time would be better spent just making sales. The only thing that doesn't cost more time is larger sales.  I once sold a Beta Black Lotus.  It took about 5 minutes to pack, ship and insure, and I closed a sale within 24 hours.  If anything, the high dollar items more more easily than Standard singles.

These are your main costs.  I'm certain there are others worth factoring, but this brings us to a reasonable conclusion.  Examining the points where the model does and does not scale, we see the following:

To reduce shipping costs, sell fewer cards less frequently.

To reduce transaction fees, sell cards that do not require a market infrastructure (like the Black Lotus example).

To reduce storage and supply costs, hold fewer cards in inventory.

To reduce time expenditure, sell fewer cards, less frequently, and sell cards with the greatest liquidity.

This round-about discussion on scalar business has led to a somewhat obvious conclusion: you should be selling fewer cards, less frequently, for more money.  The more liquid the card, the better.  

A Beta Black Lotus is a great example of this: it's dead simple to locate a buyer, the card is worth more than almost every "normal" card in existence, and you are only processing a single card in a transaction.  

There is another option: sell complete sets or playsets.  This does not dodge the shipping cost issue, but it can cut down on storage costs.  You don't need to binder your rares if you're just going to box them up and sell them as a set.  You also get the efficiency of a single transaction, which is much better than selling singles or playsets individually.

~

This is all well and good in theory, but it's not exactly easy to trade up to a Lotus.  Traders also need to maintain liquidity in their trade binder. If you operate on a cash basis, this is easier, but most traders don't have the scratch to plop down a few grand right out of the gates.  

The solution, it seems, is to slowly trade up.  

Turn your $5 Standard cards into $20 Standard Mythics.  Turn those into Legacy staples.  Turn your Legacy staples into Legacy Power (The $200+ stuff).  Turn that into Vintage Power, and then keep upgrading your Vintage Power.  Clearly, the difficulty grows along with the price of the cards, but this is the path to scalar Magic trading.

There is also much to be said for big trade-downs.  I had a few Dual Lands sitting around, and I wasn't particularly hard-up for cash.  I took a Tropical Island and acquired a stack of Standard and Modern staples at approximately dealer buy price.  You can get away with this because other players understand that they must pay a premium for trading "up".  This is a great way to "cash out" your higher-end cards and reinvest the value into more liquid assets.  Repeat the process until you want/need to cash out.

Ultimately, there is a sweet spot somewhere along this path.  It is not easy to trade up to a Black Lotus.  

The sweet spot seems to be the stable Legacy cards like Dual Lands.   They command enough of a price tag that each transaction is of a worthy dollar amount, and they sell easily enough so you can avoid eBay and CardShark's fees.  They are always in demand and always will be.  

Cards like Candelabra of Tawnos and The Tabernacle at Pendrell Vale are more expensive, but the market is nowhere near as liquid.  

Non-land cards always run the risk of being banned or effected by metagame shifts (like Mental Misstep), whereas lands are usually a safe bet.  

I would conclude that Dual Lands and similar assets are the best way to achieve a respectable volume of dollar sales without investing a disproportionate amount of time or effort.   If nothing else, a stockpile of Dual Lands can facilitate larger trade-ups.  

They are the gateway to bigger cards, always liquid, and seem to be a very stable investment.