The hours spent on a research project tend to fall into two categories. There are hours when you are actually thinking: forming judgments about what the evidence means, identifying the argument you want to make, working out how your sources speak to each other. And there are hours spent on everything else: collecting sources, reformatting files, searching for something you read three weeks ago, building citation lists, annotating for organizational reasons rather than analytical ones.
The second category, for most researchers on most projects, is the larger one. This is not a productivity failure or a discipline problem. It is a structural feature of how research workflows have evolved, and it has real costs for the quality of the work that gets done.
What Actually Counts as Busywork
The term "busywork" risks sounding dismissive, as if these tasks are optional or pointless. They are neither. Sources need to be collected. Files need to be accessible. Notes need to be findable. The question is not whether the logistics of research need to happen, but how much of a researcher's time and attention they should consume.
Typical research busywork falls into four overlapping categories.
Collecting and reformatting. Downloading PDFs, converting file formats, screenshotting content that sits behind paywalls, copying URLs into citation managers, tagging browser bookmarks, creating folders for a project that will change shape three times before it is finished. This is real work, but it produces no insight.
Re-finding. Returning to a source that was read earlier but not annotated in a way that makes it useful now. Spending thirty minutes locating something you are certain you saved but cannot retrieve. Running the same search across your notes and browser history multiple times over the course of a project.
Duplicate reading. Reading the same section of a paper or document a second time because the notes from the first reading did not capture what the current question requires. Not because the source was unimportant, but because the note-taking was done at the wrong level of specificity.
Administrative annotation. Writing summaries for organizational purposes, tagging sources for later sorting, maintaining a running list of what has been read and what has not. Useful as infrastructure, but not productive as analysis.
None of these activities is wasteful in isolation. Together, they add up to a proportion of project time that is difficult to justify once you actually measure it.
The Time Budget Problem
The specific numbers vary by project type and the researcher's habits. But across the people who have joined our early-access program, one pattern holds: the ratio of logistics time to thinking time is worse than most researchers had consciously acknowledged before they started paying attention to it.
Consider a policy analyst working on a 40-source brief for a mid-size think tank. Before writing a sentence of analysis, they might spend eight to ten hours on source collection, format conversion, and organizational annotation. That is not unusual. On a project with a 20-hour total time budget, it means less than half the available time reaches the level of genuine thinking. The analysis gets written in the hours that remain after the filing is done, which are rarely the researcher's best hours.
The concern is not that organization is unimportant. Well-organized source material is genuinely faster to work from. The concern is whether the current ratio is structurally necessary, or whether it reflects a deficit in the tools available rather than the irreducible minimum overhead of research.
The Cognitive Cost Compounds the Time Cost
Time is not the only cost. Every interruption from analytical work to perform a logistical task carries a cognitive penalty beyond the minutes the task takes.
When a researcher stops to locate a source they cannot find their notes on, or re-reads a section to reconstruct what they extracted from it, they lose the mental context they were holding. The partial argument being assembled in working memory has to be partially rebuilt when they return. The re-entry cost is real and it accumulates across a project.
This pattern becomes most visible on long projects with many sources. As the source count grows, so does the surface area for logistical interruptions. A project with 15 sources has 15 potential re-finding moments. A project with 60 sources has many more, and each one carries its exit and re-entry cost.
The researchers in our early-access group who reported the most frustration with their workflows were not the ones struggling with hard analytical problems. They were the ones spending their best hours on filing, and arriving at the actual analysis already fatigued.
What the Busywork Displaces
The insights that tend to get crowded out are the ones that require sustained, uninterrupted engagement with the material. The connection between paper A and finding B that neither paper makes explicitly. The methodological tension in a literature that only becomes visible when you have held all of it in mind at once. The implication that follows from two sources taken together, but not from either one alone.
These are not insights that can simply be recovered later by spending more time. They arise from a particular quality of attention. When that attention is fragmented by logistics, those insights become harder to reach, and more likely to be missed entirely. The finished work shows the absence in ways that are difficult to name precisely: arguments that are technically correct but not quite sharp, analyses that summarize rather than synthesize, conclusions that follow from the evidence but do not add to it.
What We Are Not Claiming
There is no version of research that has zero logistics overhead. Sources need to be collected and organized. The setup work matters, and doing it carelessly creates a different set of problems downstream.
We are not saying that every researcher should simply spend less time on logistics by willpower. The logistics tasks that consume the most time are largely mechanical: finding, collecting, reformatting, and annotating for organizational purposes. These are the tasks most amenable to structural handling, without requiring the researcher's analytical attention.
Pilgrim was built around this specific problem. When the collection and organization work happens in the background, the researcher's first interaction with a project is the material itself, already gathered, ready for the thinking. The hours recovered are not recovered by working faster. They are recovered by redirecting time from tasks that a tool can handle toward the analysis that only the researcher can do.
That said, we want to be clear about the boundary: Pilgrim does not do the analysis. It does not identify the tension between two papers. It does not write the argument. What it removes is the overhead that sits between the researcher and that work. Whether that overhead was consuming four hours or twelve on your last project, the time and attention it consumed are now available for the part that actually required you.