Skip to content
Home » Blog » Summaries » Personal Knowledge Management » The Science of Managing Our Digital Stuff (Summary)

The Science of Managing Our Digital Stuff (Summary)

The science of managing our digital stuff - Summary

When I need to find a document on my computer, I navigate through folders rather than search for it. Even though search seems more efficient. For years, I assumed this was just habit or stubbornness. I believed newer technologies like desktop search and tagging systems would eventually replace my manual organization methods.

According to research by Bergman and Whittaker, this preference isn’t random or irrational. It’s rooted in how our brains work. When we navigate folders, we use ancient brain structures that have evolved for physical navigation. Search, by contrast, requires verbal processing resources that are far more limited.

What looks like technological backwardness turns out to be cognitive wisdom. Our brains have spent millions of years perfecting navigation and only decades with search.

Table of Content

About the Book

The Science of Managing Our Digital Stuff upends everything we thought we knew about organizing our digital lives. Through rigorous studies involving thousands of users, Bergman and Whittaker reveal why we stubbornly stick with folders despite supposed better alternatives like search and tagging. The book introduces “curation” as the missing framework for understanding personal information management, explaining what we do with our files, emails, and photos and why we do it. Most importantly, it translates these insights into practical design principles that work with our brains instead of against them. It shows how personal information tools should be built for real humans, not idealized users.

About the Authors

Ofer Bergman and Steve Whittaker have spent nearly two decades solving the mysteries of how people manage their digital stuff. Bergman brings sharp experimental precision to understanding user behaviour. At the same time, Whittaker, a Professor at UC Santa Cruz with previous stints at Bell Labs and IBM, blends academic rigour with real-world application. They’ve published over forty pioneering papers that shaped our understanding of personal information management. Their complementary perspectives, bridging psychology, information science, and computer interfaces, give this book its unique power to explain why we organize our digital worlds the way we do.

Practical Takeaways

Organizational Strategies That Work

  • Create shallow, broad folder hierarchies
    Research shows that folder structures with fewer levels but more folders at each level are optimal for retrieval. Aim for folder depth of about 3 levels with 10-20 items per folder.
  • Use location for important items
    Place critical files on your desktop or in easily accessible locations. This works with how our brains naturally process information.
  • Demote files rather than delete them
    For files of uncertain value, consider “demoting” them (making them less visually prominent) rather than deleting them outright. This preserves potentially useful information while reducing visual clutter.
  • Organize by project rather than file type
    Group related information by projects or topics regardless of format (documents, emails, bookmarks) to overcome the “project fragmentation problem.”
  • Trust your preference for navigation
    If you prefer navigating folders over searching, that’s not a bad habit – it’s cognitively efficient. Navigation uses less verbal working memory than search.

File Management Techniques

  • Name folders with consistent conventions
    Use clear, time-based labels for consistency, especially for frequently accessed collections like photos.
  • Clean up folder views for better retrieval
    Visually distinguish between versions of the same file (such as making older versions less prominent) to spot the latest version at a glance.
  • Use recency and frequency information
    Leverage the fact that recently and frequently used items are more likely to be needed again. Many systems already support this through “recent documents” features.
  • Preserve context with your content
    When possible, maintain connections between related items. For example, keep emails linked to the attachments you’ve saved from them.

Email Management

  1. Separate actionable from informative emails
    Treat emails requiring action differently from those that are just informative.
  2. Keep actionable items visible
    Use your inbox as a to-do list for items requiring action, or create a specific “action required” folder that you check regularly.
  3. Don’t create too many email folders
    Research shows that too many email folders (especially tiny ones with just 1-2 messages) don’t improve retrieval and can hinder it.

Collaborative Information Management

  • Be cautious with shared repositories
    When collaborating, be aware that retrieving files from folders created by others is five times more difficult than from self-created folders.
  • Consider email for critical shared files
    Despite cloud storage’s popularity, email attachments lead to better retrieval outcomes for important shared files.
  • Organize shared repositories carefully
    If using shared repositories, create clear organizational structures and guidelines with collaborators.

Personal Information Management Mindset

  • Think of organization as communication with your future self
    When organizing files, consider how your future self will think about this information when looking for it.
  • Active organization improves memory
    The act of organizing information helps you remember it, which assists later retrieval. Don’t rely solely on automatic organization.
  • Understand the curation lifecycle
    Recognize the three phases of personal information management: keeping (deciding what to save), management (organizing saved information), and exploitation (retrieving when needed).
  • Accept some overkeeping
    People naturally tend to keep more than they need. Rather than fighting this tendency, develop strategies to manage it effectively.
  • Trust subjective organization
    Your personal, subjective way of organizing information is valid and valuable precisely because you’re organizing for yourself.

These takeaways reflect both the empirical research findings presented in the book and the practical implications of the user-subjective approach to personal information management developed by the authors.

Core Thesis

Personal Information Management (PIM) is fundamentally different from other types of information management, requiring unique theories, research approaches, and design principles that recognize the subjective, self-directed nature of how individuals curate their personal digital information.

Key Points

  • PIM concerns how individuals store, organize, and retrieve their personal information for future use, which the authors define as “curation” rather than “consumption.”
  • Current information management theories focus primarily on public information and ignore the specific challenges of personal information management.
  • PIM involves three key processes: keeping (deciding what to save), management (organizing saved information), and exploitation (retrieving information when needed).
  • Despite technological advancements in these alternatives, people strongly prefer navigation through folder hierarchies over search, tagging, or group information management.
  • The preference for folders has cognitive and neurological bases: navigation uses brain structures evolved for physical navigation and requires less verbal working memory than search.
  • The authors propose the “user-subjective approach” to PIM design, with three principles focusing on subjective importance, project classification, and context.
  • The book presents empirical studies and practical implementations of these principles, demonstrating their effectiveness in real-world applications.

Contextual Framework

The authors position PIM as a distinct field requiring its own theoretical foundation. They draw upon cognitive psychology to explain user behaviors and preferences, and employ cognitive neuroscience to understand underlying mechanisms of information organization and retrieval.

The work emerges during a period of significant technological change (2010-2016), with rapidly expanding personal digital collections and evolving tools for organizing and accessing digital content. The authors frame PIM as fundamentally a self-directed communication process, where users organize information for their future selves, contrasting with traditional information management designed for others to access.

Part I: Personal Information Management: The Curation Perspective

This section establishes the theoretical foundation for understanding PIM as curation rather than consumption. The authors argue that current information theories focus almost exclusively on discovering and consuming new public information, overlooking how people manage their collections.

Chapter 1: Personal Archives and Curation Processes

The authors introduce the concept of information curation—how individuals select, organize, and access personal collections. They present a communication metaphor for PIM: unlike traditional information management, where an information professional organizes for different users, PIM is “a solipsistic interaction between a person and him- or herself at two different times.” This allows for subjective organization methods that might not make sense to others.

The chapter defines personal information as anything users preserve for future access and establishes the three-stage curation lifecycle: keeping, management, and exploitation. It identifies key properties of information items that affect curation:

  • Information vs. action orientation: Some items require action by specific deadlines, while others are purely informative
  • Uniqueness: How replaceable or recoverable information is
  • Accumulation status: Whether information automatically accumulates or must be actively preserved

Chapter 2: Keeping This chapter examines how people decide what information to keep.

The authors note that making decisions is difficult because it requires predicting future needs. Studies across different media (paper, email, contacts, photos, web) show people tend to “overkeep” due to:

  • Deferred evaluation: postponing decisions about information value
  • Information overload: inability to process all incoming information
  • Loss aversion: people prefer avoiding potential loss over gaining equivalent benefits

For actionable items (like emails requiring response), keeping patterns are different from informative items. In practice, the “one-touch” ideal of immediately processing and discarding items rarely happens. Studies show that people keep about 70% of emails, with actionable messages almost always retained until completion.

Chapter 3: Management

This chapter explores how people organize their information. Management requires predicting how one will think about information at retrieval time—a difficult cognitive task. The authors distinguish between:

  • Semantic cueing: organizing information by conceptual similarity
  • Temporal cueing: organizing by time or urgency (especially for actionable items)

They contrast filing (systematically categorizing into hierarchical folders) with piling (less structured organization, often chronological). Both approaches have advantages and disadvantages. Filing requires more upfront effort but potentially offers a better retrieval structure, while piling requires less effort but may reduce findability as collections grow.

Management strategies vary by information type. Email requires different approaches for actionable versus informative messages. Photos often receive minimal organization, which causes later retrieval problems. The authors also note the influence of personality traits on management strategies.

Chapter 4: Exploitation

This chapter discusses how people retrieve their personal information. Unlike searching for new information, exploitation involves familiar, personally organized information. The authors identify two main strategies:

  • Navigation: manually traversing hierarchies of folders and subfolders
  • Search: specifying properties of target items via queries

Their research shows that folder structure affects retrieval success and efficiency. Shallow hierarchies with moderately sized folders (around 10-20 items) optimize retrieval. For email, opportunistic strategies (scrolling, sorting, searching) are more prevalent than folder-based retrieval. Photo retrieval for older photos is often unsuccessful due to poor organization. Web page retrieval surprisingly favours navigation over search, with most web revisits occurring through links and the back button rather than search.

Part II: Hierarchical Folders and Their Alternatives

This section explores why folder hierarchies remain the preferred method for PIM despite proposed alternatives. The authors test different technologies against folders and explain why navigation is cognitively preferred.

Chapter 5: The Search Alternative

The authors examine the “search everything” approach, which proposes that advanced search engines should replace hierarchical folder navigation. Despite major improvements in search technology (real-time indexing, incremental search, cross-format search), their studies show users still strongly prefer navigation (56-68% of retrievals) over search (7-15%). The reasons include:

  • Search is used primarily as a “last resort” when location is forgotten (for only about 25% of files).
  • Improved search engines do not lead to reduced reliance on folder organization.
  • Navigation provides consistency and incremental feedback that search lacks.

Chapter 6: The Tagging Alternative

This chapter examines tags as an alternative to folders. Tags theoretically offer advantages: allowing multiple classification of items and eliminating hierarchies. While tags work well for social web content, the authors’ studies show that in personal information management:

  • Users strongly prefer folders over tags for both storage and retrieval.
  • People rarely apply multiple tags to items (92% of items received only one tag).
  • Users prefer hierarchical organization over flat storage.
  • Location-based retrieval is preferred over non-location-based methods.

The authors suggest this preference exists because single classification is cognitively simpler than multiple classification, and most users successfully remember where they stored their files.

Chapter 7: The Group Management Alternative

This chapter explores Group Information Management (GIM), where multiple users share and organize files in common repositories (like Dropbox or Google Drive). While GIM has theoretical advantages, the authors’ research shows:

  • Users strongly prefer email-based file sharing (86% of files) over shared repositories.
  • File retrieval failure rates are significantly higher for GIM (22%) than PIM (13%).
  • Files in folders created by others are 5 times more likely to fail retrieval than those in self-created folders.
  • GIM fails because organization is subjective; what makes sense to one person may not to another.

They identify four reasons PIM outperforms GIM for retrieval: the subjectivity of classification, the benefits of active organization for memory, the importance of episodic memory, and the reduced sense of control in GIM.

Chapter 8: Why Is Navigation the Preferred PIM Retrieval Method?

This chapter provides cognitive and neurological explanations for the preference for navigation over search. Two studies present convergent evidence:

  • A dual-task study shows navigation requires less verbal cognitive attention than search, leaving more mental resources available for primary tasks.
  • An fMRI study reveals navigation activates bilateral posterior brain regions used for physical spatial navigation, while search activates left-lateralized frontal regions associated with verbal processing and working memory.

These findings explain why navigation is preferred: it uses evolutionary older brain regions specialized for spatial navigation, requiring less of the limited verbal working memory resources often needed for other tasks.

Part III: The User-Subjective Approach to PIM Systems Design

This section presents the authors’ design framework for PIM systems based on their empirical findings. They propose three principles that exploit the self-directed nature of PIM.

Chapter 9: The User-Subjective Approach

The authors introduce their approach to PIM design, arguing that since the organizer and retriever are the same person, PIM systems should exploit subjective information attributes. They identify three subjective attributes:

  1. Importance: How valuable is the information to the user?
  2. Project: What activity or topic does the information relate to?
  3. Context: In what circumstances will the information be used?

These attributes inform three design principles for PIM systems.

Chapter 10: The Subjective Importance Principle

This principle states that visual salience and accessibility should correspond to an item’s subjective importance. It has two subprinciples:

  • Promotion: Important items should be highly visible and accessible.
  • Demotion: Less important items should be visually de-emphasized but kept in context.

The authors developed and evaluated three implementations:

  1. GreyArea: Allows users to demote files to a grey area at the bottom of folders.
  2. DMTR: Automatically demotes unused mobile phone contacts.
  3. Old’nGrey: Greys out older versions of files to highlight the most recent version.

All three implementations showed significant improvements in retrieval efficiency and user satisfaction.

Chapter 11: The Subjective Project Classification Principle

This principle addresses the “project fragmentation problem,” where information related to a single project is scattered across different applications and folders. It states that information items from the same project should be grouped together regardless of format or application. The authors propose a “single hierarchy solution” implemented as ProjectFolders, which would allow documents, emails, web favourites, and other items related to a project to be stored and accessed together.

Chapter 12: The Subjective Context Principle

This principle states that information should be retrieved in the same context in which it was previously used. The authors identify four types of context:

  1. Internal context: User’s thoughts while interacting with information.
  2. External context: Other items concurrently being used.
  3. Social context: Other people related to the information.
  4. Temporal context: The state of the item when last used.

They present implemented designs for each context type:

  • ChittyChatty: Links notes to recordings of meetings to recreate internal context.
  • ItemHistory: Shows documents, web pages, and emails that were open simultaneously.
  • ContactMap: Organizes information by social connections.
  • PiccyChatty: Uses social information to highlight important content.
  • Starlight: Manages temporal context for actionable emails.

Nuanced Perspectives

The authors acknowledge multiple perspectives throughout their work. They recognize arguments for alternative folder-based navigation and present them fairly before offering counterevidence. They note that while folders have limitations, proposed alternatives (search, tags, GIM) fail to address the fundamental cognitive processes involved in PIM. They also acknowledge that certain technologies may work well in one domain (e.g., tags for Web 2.0) but not in PIM.

The book recognizes individual differences in management strategies, noting that personality traits influence organizational preferences. The authors carefully present their user-subjective design principles abstractly, encouraging multiple possible implementations rather than prescribing specific designs.

In discussing future directions, they maintain openness to emerging technologies that might improve PIM while cautioning against assuming that technologies successful in other domains will transfer effectively to personal information management.

Underlying Assumptions

The book operates from several key assumptions:

  1. Information management is not one-size-fits-all; PIM has unique requirements because users organise for themselves.
  2. Understanding cognitive processes is essential for designing effective PIM systems.
  3. Empirical research, rather than intuition, should guide design decisions.
  4. Individual differences in organisational strategies are significant and should be accommodated.
  5. The fundamental curation processes (keeping, management, exploitation) persist despite technological changes.
  6. Active user involvement in an organization is valuable and should not be eliminated by automation.
  7. The familiarity users have with their own organization schemes is a strength to be leveraged, not a problem to be solved.

The authors’ methodology reflects a commitment to ecological validity, using studies of actual user behaviour with personal information rather than artificial lab tasks whenever possible.

Connections & Implications

The work connects PIM research to broader fields, including cognitive psychology, neuroscience, human-computer interaction, and information science. Emphasizing PIM’s subjective, personalized nature challenges dominant information management paradigms that focus on standardization and objective organization.

The practical implications are significant for designing personal information systems. Rather than attempting to replace folder hierarchies with search or tagging, the authors suggest enhancing folders with features that support subjective attributes. Their design principles could inform the development of future operating systems, email clients, and file management tools.

The book also has implications for teaching information management skills. The authors caution against blanket recommendations for “best practices,” noting that effective strategies depend on individual differences and specific contexts.

The authors acknowledge that emerging technologies like machine learning, cloud storage, and wearable devices will create new challenges and opportunities for PIM. However, they suggest that understanding the fundamental cognitive processes involved in curation will remain essential, regardless of technological advancement.