A Year of Documenting Lunches
Sometime last year, I started writing down what I ate for lunch every day. There was no grand motivation behind this. I did not begin with a hypothesis to test or a health goal to pursue. I was reorganizing some of my personal tracking files, which I keep in a simple spreadsheet, and I noticed that I had columns for various daily metrics but nothing about food. It occurred to me that lunch, specifically, might be interesting to track because it is the meal where I have the most variability and the most conscious decision-making. Breakfast for me is almost always the same: oatmeal with a banana and black coffee. Dinner depends on what groceries I have and what I feel like cooking. But lunch sits in an odd middle ground where I am making an active choice in the middle of a workday, often while thinking about other things, and I was curious about what patterns might emerge if I actually paid attention.
The system I use is deliberately simple. I have a spreadsheet with columns for the date, a brief description of the meal, where I ate it, whether it was homemade or purchased, and an optional notes column for anything worth recording. I do not track calories or macronutrients. I am not interested in optimizing my diet. I just wanted a record, the way you might keep a logbook or a journal. The entry takes about thirty seconds to fill out, which is important because any tracking system that requires more effort than that tends to get abandoned within a few weeks.
After accumulating roughly a year of data, I can report the following observations, which I acknowledge are interesting primarily to me and possibly to no one else. The most common lunch I eat is a sandwich. This was not surprising, but the specifics were mildly illuminating. The default sandwich is turkey and Swiss on whole wheat with mustard and lettuce. I eat this, or a close variant of it, approximately twice a week. The second most common lunch is leftover dinner from the previous night, which accounts for maybe another two days per week. The remaining days are split between soup, salads, and the occasional meal eaten out at a nearby deli or cafe.
The seasonal variation was more pronounced than I would have predicted. During the colder months, soup became much more frequent, sometimes appearing four or five times in a single week. I make a large batch of soup on weekends, usually something simple like lentil, minestrone, or potato leek, and eat it throughout the week. In the warmer months, sandwiches and salads dominated, and I was more likely to eat something cold or room temperature. This makes intuitive sense, but seeing it laid out in the data was satisfying in a way I find difficult to explain. There is something gratifying about confirming, with evidence, a pattern that you vaguely suspected but never verified.
The homemade versus purchased ratio was heavily weighted toward homemade. Out of roughly three hundred and sixty-five entries, I estimate that fewer than forty were purchased from a restaurant or takeout establishment. Most of those purchased meals clustered around certain periods: a week when my kitchen was being repaired, a stretch where I was unusually busy with a project, and a few scattered occasions when I simply did not feel like preparing anything. The average purchased lunch cost between eight and thirteen dollars, which I noted mostly out of idle curiosity. Over the course of a year, the total spent on purchased lunches was somewhere in the neighborhood of four hundred dollars. This seems reasonable to me, though I have no particular benchmark for comparison.
One pattern that surprised me was the repetition within a given week. I tend to eat the same thing for lunch several days in a row, then switch to something different for the next several days. This is partly a function of how I shop and cook. If I make a pot of soup, I eat soup until it is gone. If I buy a loaf of bread and deli turkey, I eat sandwiches until the bread is stale or the turkey runs out. There is a natural cycle to it that I was not fully aware of until I saw it in the data. The average run of the same basic meal is about three consecutive days, after which I switch. I do not plan this consciously. It just happens as a consequence of how I provision my kitchen.
The notes column has turned out to be the most interesting part of the record, though it is also the most sparsely populated. I only write a note when something is worth remarking on, which is not often. Representative entries include: "Tried adding pickled onions to the turkey sandwich. Good." "Lentil soup came out too thick, added broth to thin it." "Ate at the deli on Main Street, got the grilled chicken wrap. Adequate." "Forgot to eat lunch until 3pm, had crackers and cheese at my desk." These are not revelatory insights. They are small observations that, accumulated over time, form a kind of texture. Reading back through a few months of entries is like leafing through a very mundane diary. It is boring in the way that daily life is boring, which is to say, not entirely.
I have occasionally been asked why I bother with this. The question is fair. The data does not serve an obvious purpose. I am not using it to lose weight or save money or optimize my nutrition. The honest answer is that I find the act of recording inherently satisfying. I have always been a person who keeps records. I track my reading, my household maintenance tasks, my walking routes. The lunch log is just another instance of this tendency. There is a quiet pleasure in the completeness of it, in knowing that if I wanted to, I could look up what I ate on any given Tuesday in the past year. The fact that this information is almost entirely useless does not diminish the pleasure. If anything, the uselessness is part of the appeal. It is documentation for its own sake, a practice of attention applied to the most ordinary of daily activities.
I intend to continue the project for at least another year. With two years of data, seasonal patterns will be clearer and I will be able to compare year over year. Do I eat more soup this winter than last? Has the sandwich ratio changed? These are small questions, but any long-term dataset becomes interesting once it has enough points to reveal a trend. The project requires almost no effort, generates no stress, and provides a quiet satisfaction each time I update the spreadsheet. As personal experiments go, that qualifies as a success.
If anyone is considering something similar, my only advice is to keep the system simple. Track one meal, not all of them. Use a format that takes seconds to update. Do not try to measure everything. The goal is not precision. The goal is simply a record of ordinary days.
— Kevin Vadala