I just typed this up last night for an interview. Because it's already typed up and sitting on deck, I see no reason not to share it with backers. So here's a bit of a detailed walkthrough of how I put together a $3 game ...
First and foremost is obviously the budget. With a $3 backer level, the Kickstarter fees typically run ~10%, so that's $0.30 right there. And the cheapest international postage I can get, but to the most expensive country, is $1.20. Most of Europe costs that much, as does Australia. So I added that to the Kickstarter fee ($0.30 + $1.20 = $1.50) and subtracted the sum from the backer level ($3.00 - $1.50 = $1.50), and that was my per-unit production and packaging budget.
The postage also gives me my size and weight limits. I looked up the size and weight limits on the USPS Shipping Calculator and verified them against the Royal Mail Shipping Calculator. Once I had the parameters for shipping as a letter, which included the length (11.5"), height (6.125"), thickness (0.25"), proportion (length/height ratio: 2.5 ≥ X ≥ 1.3), and weight limit (1.0 oz for $1.20), I had the limits for the game, including packaging.
Packaging has weight, and because I'm working with a very clear weight limit, I had to take it into account. The absolute lightest envelopes that I could find were 11# Tyvek business envelopes. I ordered a sample from envelopes.com for $1, and weighed it three times on my Weight Watchers kitchen scale (which is accurate to the nearest 20th of an ounce, whereas the Stamps.com scale is only accurate to the nearest 10th of an ounce) and got 0.11 oz. To save weight, I could have the company print both shipping and return addresses on the envelopes, which would save the weight of both a shipping and return label. The envelopes cost $48.00 for 150, or $0.32 each to purchase and ship. Printing them with a processed address list would cost $0.38 each, regardless of number. I marked that as a tentative expense that I could save by addressing all envelopes by hand, but calculated that into the cost just in case. Shipping the envelopes to me would cost $4.20, which when spread across my 150 backer goal comes to $0.028 each. So the total cost of packaging was $0.32 + $0.38 + $0.028 = $0.728. That's my per-unit packaging budget. Subtract that from what remained after the Kickstarter fees and postage and you get $0.772. That's my per-unit production budget.
To find the weight limit of whatever material I was going to use, I took the postal limit and subtracted the weight of the Tyvek envelope to get 1.0 oz - 0.11 oz = 0.89 oz. That's the absolute heaviest my game could get before I'd have to pay more postage. And I absolutely CANNOT afford to pay more postage, so that's a hard limit, there is no flex room whatsoever there.
The USPS size restrictions wouldn't affect me that much, because it would be hard to get an envelope big enough to be considered a package while still snaking in under that 1.0 oz weight limit. The 0.25 inch thickness limit isn't a big deal, except that a box is completely out of the question. The thickest cardstock commonly available from custom printers is 14 pt, which means 0.014 inches thick. 0.25 inches / .014 inches = 17 (inches cancel out), so at most I could have 17 cards layered on each other and still make it under the thickness limit. So that was my hard mailable limit for game area: 17 cards x 6 inches x 9 inches = 918 card sq inches. That's far higher than the affordable limit for area, so I never even got close to 918 sq inches. Note: the affordable limit for area is a soft limit, because the prices fluctuate from company to company and depending on the volume that you purchase, so it wouldn't do any good to say much about that.
14 pt card stock weighs 308.5 grams per sq meter, or 0.007 ounces per sq inch. Because I have 0.89 ounces to play with, I divided the weight per sq inch into the weight and got my size restriction by weight, 0.89 ounces / 0.007 ounces per sq inch = 127 sq inches (ounces cancel out). That's certainly less than 918 sq inches, so that was my hard limit for size.
So I sat down with $0.772 to spend on each game, and all game components combined would have to fit onto no more than 127 sq inches of printed material.
The secondary, downstream problem was that I knew from previous experience that gamers don't like their counters to be too thin. 14 pt is a nice cardstock, but players would be happier with thicker stock. So I contacted prints all over the country, some in Canada, and some overseas, to see what thickness of cardstock they have and what kind of pricing I'd be looking at. I collected 22 quotes from different printers, and put together a production plan that would begin with the absolute cheapest printer at 14 pt, then when a volume discount would make up for the difference in price I would go with the cheapest 16 pt printer, then 18 pt, and finally make the big granddaddy leap to the 24 pt. Now, 24 pt was the highest I could realistically go, because even though there were printers with 30 pt stock and thicker, the price just got too high. So I set the 24 pt as the highest possible stretch goal. Then I recalculated the printable area based on a 24 pt stock: 0.007 ounces per sq inch / 14 pts = 0.005 ounces per sq inch pt, then multiply that by 24 pt to get 0.005 ounces per sq inch pt * 24 pt = 0.012 ounces per sq inch (pt cancels out). Divide that into the 0.89 ounce weight limit, and you get 0.89 ounces / 0.012 ounces per sq inch = 74 sq inches (ounces cancel out).
Prior experience had also taught me that players don't care about the space limitations I work with, they just want a game that's easy to play. Having rules on the back of the board saves space and money, but makes the game difficult to learn without having a pdf of the rules handy. And not everyone has the capability to have the pdf handy, so I didn't want to rely on that. So the first component was going to be the board, and the second component would be the rules sheet. The rules sheet would be handled less than the board and counters, once the game was learned, so I wasn't as worried about making sure that it's nice a thick. Printing on whatever paper would fit the weight limits would be sufficient.
So to find the weight limit for the rules sheet, I had to calculate the weight of the board. A standard number 10 envelope, regardless of its material, is 4.125 in * 9.5 in. To give a bit of breathing room and make packaging easier, I decided that the board would either be 4 in x 9 in, or would fold to that size. A quick calculation informed me that the board would have to be a single piece: 4 in x 9 in = 36 sq in ... and if it was folded in half, it would be 72 sq in ... but I have a 74 sq in limit, and I needed some of that weight for a rules sheet, so 36 sq inches was the final decision. The I multiplied the board size by the weight per area to get the weight of the board: 36 sq inches * 0.012 ounces per sq inch = 0.432 ounces (sq inches cancel out). Subtract the weight of the board from the available weight that I had, and that was the remaining weight I could use for rules and anything else: 0.89 ounces - 0.432 ounces = 0.458 ounces.
One more thing I wanted to add was a zip lock bag. Once again falling back on prior experience, I knew that some kind of packaging would make the game more appealing and easier to store, and would also protect it during shipping. So I found a place that could sell me 2 mil 4" x 9" zip lock bags in bulk at $0.14 each. I weighed the 4 mil 6"x 9" zip locks that I already had and got 0.33 ounces. I multiplied that by half to get the approximate 2 mil 6" x 9" zip lock weight. Then I multiplied that by 2/3rds (because 4x9 is 2/3rds of 6x9) to get the approximate weight of a 2 mil 4" x 9" zip lock: 0.33 ounces * 0.5 = 0.165 ounces ... 0.165 ounces * 2 / 3 = 0.11 ounces. Subtract that from the remaining weight, and that gives me the weight available for the rules sheet: 0.458 ounces - 0.11 ounces = 0.348 ounces.
Because the board was going to be 4" x 9", it makes sense that the rules sheet should either be that size or fold to that size, just for aesthetics. So I perused the various paper weights available at the different printers, using 36 sq inches as my template, and settled on 100# gloss text, which weighs 0.0034 ounces per sq inch. Multiply the weight per size by the size to get the weight of the rules sheet printed on that paper: 36 inches per sheet * 0.0034 ounces per inch = 0.1224 ounces per sheet (inches cancel out). Remembering that I have 0.348 ounces to play with, and each 4" x 9" sheet would weigh 0.1224 ounces, that means that I can have a rules sheet that is either two 4" x 9" sheets or a sheet that folds in half to 4" x 9". I opted for the latter, because it would be cheaper to print one folded sheet than two separate sheets. So the weight of the rules sheet would be: 0.1224 ounces per sheet * 2 sheets = 0.245 ounces (sheets cancel out). Subtract that from the remaining weight to get the margin of error that I can be afforded in my calculations: 0.348 ounces - 0.245 ounces = 0.103 ounces ... which is more or less a 10% margin. Subtract that from the beginning weight to get the shipping weight of the game: 1.0 ounces - 0.103 ounces = 0.897 ounces.
The math I just did here for you here is one thousandth of an ounce off the total I have on the itemized cost and weight list on the Kickstarter project page, which I'll attribute to rounding errors.
In summary, I treated it like a math project, not an entertainment project. I measured absolutely everything, as accurately as possible, then I obtained physical samples from the printers and measured those, then I run a variance analysis and always used the worst possible scenario. I input the information in a giant spreadsheet to come up with the materials available to me at the best possible price. Then I created a game that fit within those logistical parameters. The theme and gameplay were the easy parts.
My mathematical model suggests we'll go into the final 48 hours with about $8300 in funding, so I'm not going to call it for the super deluxe Kickstarter Edition quite yet. But any experienced backer knows that the final 48 can be a wild ride, it's nearly impossible to predict what can happen. It could be flat, or it could bump us up another $3000. I would think it'd be closer to the former, because who'd be on the fence about a $3 pledge? But you never know.