Someone asked me how recipes and articles evolve, here at FeedMe headquarters. I thought I would show the process I used to come up with the caramel candies.
I usually start with a "recipe survey." I poke around on recipe sites and find the ingredients and quantities recommended there. I draw a table so I can see them all, side by side. Here's the one I did for the caramel article:
A couple of interesting things happen at this stage. First is that the range of measurements is usually very wide. It's amazing that such wildly varying recipes all work. (Sometimes, I later discover that they don't!) The second surprise is that there are always a few whacky ideas. In this case, it's Alton Brown (the AB column). He, alone, includes cream of tartar and soy sauce (!). He's also one of two to include salt. They all should, because candy makers know that a pinch of salt amplifies flavor in sweet products. But I decided to skip the soy sauce after reading the comments from readers. Some liked it, some thought it weird. I wasn't going for weird today.
Next, I note the major differences and guess at their effect. In this case, we have brown sugar. I like that idea. We have one major disagreement between the condensed milk group and the ones that don't use it. I think I'd like to avoid that but I also know condensed milk is very common in caramel and caramel-like recipes (notably, dulce de leche), so I decide to try that first.
I pick one key ingredient and scale the recipes so all use the same amount of that ingredient. Usually, it's easy but in this case, the key ingredient, sugar appears in different forms. Sugar and brown sugar have similar sweetening power and corn syrup is 75% as sweet so I tweaked the numbers so sugar plus brown sugar plus 75% corn syrup added up to the same amount. I counted the condensed milk's sweetness at about half that of sugar. It's all pretty approximate.
Next, I eyeball the recipes and come up with a single consensus recipe. I favor recipes that have good reviews or come from sources I trust (like Cook's Illustrated, most Food Network sources, and a few magazines).
All this sounds challenging but it's easier and faster than the explanation. Since I am not a baker, approximations are ok!
Next step is to make my consensus recipe. In this case, my reference recipe had a step that sounded questionable to me: It had me add the dairy at the start. That sounded like it would curdle the milk products and make a grainy caramel.
It did. It was obvious even before the boil:
but didn't hurt the finished product, which was delicious.
You might notice the foil-lined pan in the picture above. After ten minutes of careful work with a kitchen torch to peel away foil, I learned foil sticks. The final recipe calls for parchment.
I decided I didn't want the condensed milk and switched the recipe around, adding dairy after the sugar was caramelized. I made a second batch which was just about right but I let the temperature rise to 150 degrees. It was delicious but harder than caramels usually are. I had discovered how to make a Sugar Daddy! So I added that as an option for the third recipe, which is the one I published. I also experimented with the cream and noted in the final that you can exchange milk for some of the cream.
Also see: Cooking Without Recipes: Carrot Soup Case Study, which describes the recipe development for carrot soup.