My work has been mostly digital for over twenty-five years. As a result, I have a lot of files lying around. Some are things I’m currently working on, while others are older and less relevant to my current needs. But all of it is important to me. And it’s not just files; I also keep archives of my digital communications. This includes email mailboxes going back almost twenty years, chat threads, and (more recently) Slack channels. I also manage several critical long-term databases, including a password manager.

Making this stuff findable is a challenge to those of us who work with computers. A new computer user will soon discover that saving all her files to the desktop doesn’t scale. And while search systems in modern operating systems (such as Spotlight on the Mac and Cortana in Windows) are pretty good, they’re not omniscient — and even if they were, you often wouldn’t know what to search for to find the information you need. As a result, we still create folders to store files for longer-term retrieval.

How do you decide how to group these files? There are many options open to you. For example, if you’re a freelance designer, you could create a folder for each project, a folder for each client, or a folder for each type of file. After a while of doing this, you’ll discover this, too, doesn’t scale. Then you’ll experiment with nesting these folder structures: a folder for each client which contains folders for that client’s projects.

This folder/file structure is an information architecture designed for an audience of one: your future self. You need to be able to retrieve the things you need when you need them, which will probably be many months (or even years) in the future. This requires that you make predictions about how you will expect to look for the things you need at a different time.

This entails articulating your mental model of your digital stuff. For example, if you’re a freelance designer, client projects are most likely central to your work. As a result, projects will be an important factor in how you organize your personal information environment. (I have such a structure on​ my computer: I have a folder called “Projects” where I keep all my project files together.) If you follow this route, you’ll want to establish a consistent naming scheme for these folders, ideally featuring the project name as the label.

One of the advantages of this method is that it allows you to replicate this structure in systems other than your computer’s file/folder structure. For example, you could create a series of folders in your email application where you keep conversation threads related to each project. For sanity’s sake, you should make the naming structure consistent across systems; the project’s folder in the Finder should have the same name as the project’s folder in the mail application.

But of course, not everything you deal with is a project. So you also need to understand what the difference is between a project and not-a-project. You’ll need to store and find many things that are not-a-project. How do you group them? In my case, I have a folder in my computer called “Reference” where I keep the not-a-project stuff that I may need to refer to in the future. Under this folder, I have several sub-folders grouped by subject matter. For example, I one of these folders contains documents for the house where my family and I live.

How granular do you go with the grouping? That depends on several factors, including how many documents you plan to keep in a folder and how frequently you expect to need them. A folder that contains several hundred files probably needs sub-folders. My “House” folder has less than a dozen files in it, so it doesn’t have any sub-folders in it.

Of course, that folder isn’t called “House.” I’ve lived in several houses in the past twenty-five years. When I look for these files, I need to know which house I’m looking at. The labeling of folders is important if you want to make groupings unambiguous. (In my case, the folder’s label is the house’s street address.)

This structure didn’t emerge spontaneously; it evolved over a long period of time. There was a time in my life when I hadn’t lived in more than one house, so I didn’t need a complicated file/folder structure to store my house-related information. Personal information architectures change over time as your information needs change.

As with all information architectures, creating and managing an effective personal IA requires discipline to group and label things consistently, and flexibility to recognize when grouping and labeling schemes require tweaking. Achieving a balance between consistency and flexibility can lead to a structure that feels natural; an extension of your mind.