Baking in Process Improvements

Today I want to talk about process improvements with an example.

Recently, a relative of mine was busy preparing for a garage sale. She had also wanted to make a double batch of chocolate chip cookies, but ran out of time. So I offered to make the cookies with her daughter, who I will call Spunky Shoes. She handed me a printout of the recipe, which I have listed below:

Soft and Chewy Chocolate Chip Cookies

(from Danielle Rye’s Live Well Bake Often)

Ingredients

• 2 and 3/4 cups (345 grams) all-purpose flour (spooned & leveled)
• 1 teaspoon baking soda
• 1 teaspoon salt
• 1 cup (230 grams) unsalted butter, softened to room temperature
• 1 cup (200 grams) brown sugar
• 1/2 cup (100 grams) granulated sugar
• 2 large eggs at room temperature
• 2 teaspoons pure vanilla extract
• 1 12-ounce package (2 cups) semi-sweet chocolate chips

Instructions

1. In a large mixing bowl, whisk together the flour, baking soda, and salt. Set aside.
2. In the bowl of a stand mixer fitted with the paddle attachment, or in a large mixing bowl using an electric mixer, cream together the butter, brown sugar, and granulated sugar for 1-2 minutes until well combined. Add in the eggs and vanilla extract one at a time and mix until fully combined.
3. Slowly mix in the flour mixture and continue mixing until just combined, then mix in the chocolate chips, making sure to scrape down the sides of the bowl as needed.
4. Cover the cookie dough tightly and transfer to the refrigerator to chill for at least 2 hours.
5. Once the dough is almost chilled, preheat the oven to 350°F. Line large baking sheets with parchment paper or silicone baking mats and set aside.
6. Remove the cookie dough from the refrigerator, using a 1.5-2 tablespoon cookie scoop, scoop the cookie dough and drop onto the prepared baking sheets. Make sure to leave a little room between each ball of cookie dough as they will spread a little while they bake.
7. Bake in separate batches at 350°F for 10-12 minutes or until the edges of the cookies are lightly golden brown and the top is set. Remove from the oven and cool on the baking sheet for 5 minutes, then transfer the cookies to a wire rack to finish cooling.

Notes

Cookies may be stored in an airtight container on the counter for up to 5 days.

That’s it! How hard could it be?

A few things to note:

• I have included this recipe content for illustrative purposes only. If you are going to try the recipe, please use the most up-to-date version of Ms. Rye’s recipe from Live Well Bake Often. The printout, and the content I listed above, are from 2019. There are some minor differences in the instructions, and I do not want to be held responsible for any typographical errors I may have made. Also, the website has additional guidance not included in the recipe itself.

• My cooking and baking skills are not particularly good. The last time I baked a batch of cookies was probably 15 years ago, and they were the molasses cookies from the Silver Palate Cookbook, which are very easy and which I had made at least five or ten times previously.

We proceeded to follow the directions to make the cookie dough, although near the end, Spunky Shoes began to lose interest, and went off to do something else, so I was on my own after we put in the chocolate chips.

At baking time, I found a small ice cream scoop, which looked about right, and ran a test batch of nine cookie scoops on a large baking sheet, to leave plenty of room for the dough to spread during baking. Here is what the resulting “Batch Zero” looked like:

Very little spreading, and the bottom of each cookie was a little burnt.

Okay. I figured the oven’s temperature control had some DC offset and ran a bit hotter than the setting, so I lowered the temperature to 340°F. This time I flattened the scoops of cookies prior to baking, and proceeded with Batch 1. And I set a timer for only ten minutes this time.

While the cookies were in the oven, I was wondering what adjustments I would have to make next — and it made me think of Terry Holdt’s notes on the NMOS silicon process at MOS Technology during the development of the 6502:

I wondered if early wafer fab process engineers did this kind of tweaking from one batch to the next: things like increasing or decreasing the flow rate, or the temperature, or the spin speed for applying the photoresist. Admittedly, they had reasonably good instrumentation to measure the results, and I did not. But then again, I was just making chocolate chip cookies.

I could at least make some notes, and start taking pictures. So I did:

Here is part of Batch 1, yielding 12 cookies total; eight on one baking mat and four on the other. These SILPAT silicone mats were really useful — no mess! And this time the cookies turned out much better. While they were baking, I kept putting scoops of dough into a bowl so I would have them all ready by the time the next batch started.

Great! On to Batch 2, where I felt confident enough to make 15 cookies on the large sheet, after which I ran out of pre-scooped dough bits, so I ran just one baking sheet. This time I also shaped the edges of the dough before baking, to try to make them more circular.

When I took them out of the oven, they were starting to touch at the edges, and they were a little underdone after only ten minutes, so I had to put them back into the oven for a few minutes. Maybe the increase in the number of cookies in close proximity moderated localized heating in the oven, and that meant I needed to cook them longer?

In Batch 3, I reduced the number of cookies on the large sheet to 13, and increased the time to 11.5 minutes. And I had caught up with enough pre-scooped cookies to run a full batch of two sheets. This turned out well:

I maintained the same routine for the Batch 4, my last batch, with a minor optimization of cookie placement on the small baking mat, for a total of 21 cookies:

At this point, it took me 20 minutes total for the batch of 21 cookies, including the time to put the dough on the mat, shape it, take a picture, put it in the oven, start a timer for 11 minutes 30 second cooking time, scoop enough cookie balls for the next batch, take the cookies out of the oven, take a picture, and put them on the cooling rack.

Compare with Batch Zero which probably took me 30 minutes total, and I only got 9 burnt cookies out of it.

With a couple of practice runs:

• I increased the throughput (9 cookies in 30 minutes → 21 cookies in 20 minutes with a pipelined dough scooping step)
• I improved the quality (burnt misshaped blobs → well-baked circular cookies)

What else can we take away from this exercise?

Process Improvements in Engineering

Continuous process improvement in engineering is extremely important. If you are repeating or you anticipate repeating any aspect of your work, learn from it!

Maybe you feel like you’re in a rush and don’t have the time. But it doesn’t take long to document at least some rudimentary notes on how you do things, and then reflect on how they could be improved. A few percent extra overhead in time for process improvements can improve quality, make efforts more systematic, and decrease both the average and worst-case times of various tasks. This last point is probably the most important, in my opinion. Task latency can really blow up when you get stuck on something unfamiliar, so the next time through, the “long tail” of the probability distribution of the time to complete a task becomes much less likely.

Even if you don’t learn much yourself, documenting your processes is important because it makes it easier for a more experienced person to review your work and point out areas of improvement. I’m sure my cookie-baking exercise will make experienced bakers roll their eyes and wonder why I didn’t do X, Y, or Z… but that’s what happens when you’re experienced and I’m not. I really wish I had known about the recipe website, because it includes all sorts of guidance, which would have helped, including some of the effects of the different ingredients. For example:

The brown sugar adds moisture and creates chewy chocolate chip cookies while the granulated sugar helps them spread.

A light went on in my head when I read this afterwards — the brown sugar I used was nearly as hard as a rock, and I had to crush it into smaller pieces, so I’m guessing it didn’t have much moisture. This probably affected the cookie texture and the spreading of the cookies during baking.

True mastery of a subject involves more than just following directions; it requires a deeper understanding of how things work, including the practical knowledge of real-world effects that a simple theory may not cover.

If you think you might want to try something different but don’t have the time — well, try a pilot project. For example, suppose you have some complex system with a microcontroller and a computer database, and you want to know if it scales well to higher bandwidth. Instead of trying to get your complex system to actually run at higher bandwidth, try an experiment with a microcontroller that sends simple mock data to the computer database at high speed, and then see where the bottlenecks are. If you find places that are potential speed bottlenecks, then you have a better idea of what impacts the overall system capacity and can plan to improve those pieces of the system as time allows.

You can automate parts of your work that were manual, and make them less prone to human error. One recent effort I made was to automate some analysis of our issue-tracking project in JIRA. There are a lot of reasons I dislike JIRA, but the basic issue database functionality works well, and there is a REST API, so it is relatively easy to write a Python script to query JIRA and obtain the results as a JSON document, which you can analyze in Python. In my case, I wanted to understand how many issues were open in the different parts of our project, so we could assess if there were any areas that were not receiving enough attention. Now that I’ve automated it, I can repeat the analysis in seconds, and I’ve added automated linking to the output report, so that I can look at specific issues in JIRA in more detail.

Finally, it is important to create a culture where process improvement is encouraged. If your engineering staff wants to improve productivity and has the capability to improve, but management discourages them from spending “extra” time on this, so that they can remain more “focused” on the work they are doing, then that interest and ability to become more productive is obstructed. You might be amazed at the effect of a few words from a project leader.

Wrapup

Today I started with an example of baking chocolate chip cookies, where I learned from the results of each batch to adjust temperature and cookie placement technique until I had improved the throughput and quality of my cookie-baking.

Continuous improvement in engineering is similar: we can improve productivity and quality when we learn from our efforts and understand how to do it better.

Keep improving!

© 2025 Jason M. Sachs, all rights reserved