Backlog
A backlog is an accumulation of unfinished tasks, orders, or materials awaiting processing, commonly used to describe pending work in professional, administrative, or operational contexts. The term originates from the late 17th century, referring literally to a large log placed at the back of a hearth fire to sustain burning, with its figurative sense of a reserve or buildup emerging in the 1880s to denote stored-up obligations or inventory.[1][2] The concept appears across various fields. In business and manufacturing, it often refers to unfulfilled orders or production queues indicating demand and potential delays.[3] In project management, especially Agile and Scrum, it is a prioritized list of tasks or features guiding development.[4] In computing, backlogs describe queue accumulations in networks or systems that can cause performance issues.[5] Publishing uses the term for pending submissions, such as in academic journals facing review delays.[6] In arts and entertainment, it applies to unreleased music tracks or a gamer's list of unplayed video games.[7][8]Etymology and general definition
Origin of the term
The term "backlog" originated in the late 17th century in American English as a compound of "back" and "log," referring to a large log of wood placed at the rear of a hearth fire to provide sustained burning and heat. This literal usage symbolized a substantial, enduring source of fuel, with the earliest recorded instance appearing around 1680 in colonial contexts, including a 1684 reference by Puritan minister Increase Mather describing a fireplace log.[9][2] By the late 19th century, the word evolved into a metaphorical sense denoting an accumulation or reserve for later use, drawing from the back log's role as a slow-burning reserve. The figurative application first emerged in the 1880s, initially in contexts of stored resources such as timber; the sense of an "accumulation of unfulfilled orders" dates to around 1904, and soon extended to business jargon for unfulfilled orders or tasks.[2][10] Linguistically, "backlog" functions primarily as a noun, with no significant spelling variants, though the derived adjective "backlogged" describes situations delayed by such accumulations, as in overloaded systems or workflows.[1]Core concept
A backlog refers to an accumulation of uncompleted work, tasks, orders, or materials that arises when incoming demands exceed processing or production capacity, often due to delays, resource limitations, or inefficiencies in workflow.[3] This buildup creates a reservoir of pending items awaiting attention, distinct from routine queues by its tendency to grow over time if not addressed. The term's etymological roots trace to the late 17th century, originally denoting a large log placed at the back of a hearth fire to sustain burning, evolving by the early 20th century to signify reserves or arrears of work.[1][11] Key characteristics of a backlog include the frequent need for prioritization to allocate limited resources effectively, as unmanaged accumulation can create bottlenecks that hinder overall progress and amplify delays across systems.[3] Quantitatively, backlogs are typically measured by metrics such as the total number of pending items or the estimated time required to clear them, providing indicators of operational strain.[12] In historical contexts, early 20th-century industrial reports highlighted backlogs as markers of productivity challenges; for instance, during economic booms, manufacturing sectors experienced backlogs that prevented the full adoption of efficiency-enhancing technologies, signaling imbalances between demand and capacity.[13] Organizationally, backlogs often signal underlying inefficiencies, such as inadequate staffing or process bottlenecks, which can erode morale and increase costs through prolonged delays.[14] However, they can also represent opportunities, particularly when stemming from high demand, as in pending orders that forecast future revenue and growth potential.[3] Psychologically, persistent backlogs may foster stress among teams by creating a sense of overload, yet proactive management can transform them into strategic assets by revealing areas for capacity expansion.[15]Business and manufacturing
Order backlog
In business and manufacturing, an order backlog represents the accumulation of customer orders that have been received but remain unfulfilled, often quantified in monetary terms (e.g., $520 billion) or units (e.g., thousands of aircraft). This metric captures contractual commitments scheduled for future delivery, arising primarily when customer demand surpasses current production or supply capabilities.[16][17] A common metric for assessing order backlogs is the backlog ratio, defined as: \text{Backlog ratio} = \frac{\text{current backlog value}}{\text{annual revenue}} This ratio measures the duration of secured future revenue in years, providing insights into sales stability and operational visibility for forecasting purposes. For example, ratios exceeding 3x indicate robust demand pipelines that buffer against market fluctuations.[18][19] In the aerospace sector, Boeing's order backlog stood at $520 billion at the end of 2023, encompassing over 5,600 commercial airplanes with multi-year delivery schedules that span several years due to complex production cycles. Similarly, in construction, supply chain disruptions—such as material shortages and tariff-related delays—have prolonged backlogs; as of September 2025, the industry average was 8.5 months, reflecting challenges in fulfilling orders amid global logistics bottlenecks.[20][21][22] To manage order backlogs, firms often adopt just-in-time (JIT) production strategies, which synchronize material inflows with order fulfillment to minimize delays, reduce excess inventory, and accelerate throughput. High backlogs signal strong market demand, correlating with positive stock price movements as investors view them as reliable predictors of future earnings growth.[23][17] Historically, the post-World War II economic boom in U.S. manufacturing led to unprecedented order backlogs in the 1940s and 1950s, fueled by pent-up consumer demand and industrial reconversion from wartime production. Unfilled orders escalated from $4 billion in 1939 to $74 billion by 1942, with durable goods comprising the bulk; post-1945, backlogs persisted at elevated levels—though smaller than wartime peaks—due to labor shortages and supply constraints, sustaining high output through the decade.[24][24]Production backlog
In manufacturing, a production backlog refers to the accumulation of scheduled production runs, assembly tasks, or orders that remain uncompleted due to constraints exceeding the firm's capacity, such as machine downtime or labor shortages. This gap between demand and output disrupts operational flow and can signal inefficiencies in resource allocation. Machine downtime, often resulting from equipment failures or unscheduled maintenance, directly contributes to backlogs by halting production lines and extending fulfillment times. Similarly, labor shortages, exacerbated by factors like workforce retirements and skill gaps, reduce the workforce available for assembly and processing, leading to delayed tasks. Key metrics for analyzing production backlogs include lead time backlog, defined as the total hours of delayed production across scheduled runs, which quantifies the extent of postponements and their impact on delivery timelines. Throughput impacts are commonly measured using overall equipment effectiveness (OEE), a composite metric calculated as the product of availability (ratio of operating time to planned production time), performance (actual output rate versus ideal rate), and quality (good parts produced versus total parts), revealing losses from downtime, speed reductions, and defects that exacerbate backlogs. Low OEE values, typically below 60% in many facilities, correlate with increased backlog accumulation by highlighting unproductive time that prevents meeting demand. Illustrative examples include automotive assembly lines, where the 2021 global semiconductor shortage caused significant production backlogs for Tesla, resulting in order wait times exceeding four months for models like the Model Y due to halted vehicle assembly. In food processing, seasonal harvest fluctuations can create backlogs, as seen in California's tomato industry, where delays in harvesting and transportation bottlenecks led to excess crops overwhelming processing capacity at once, straining facilities during peak periods. A notable historical case is the 1970s oil crisis, triggered by the 1973 OPEC embargo, which imposed fuel shortages and quadrupled oil prices, leading to widespread production backlogs in the global energy sector and related manufacturing industries through supply chain disruptions and reduced operational capacity. This event forced refineries and petrochemical plants to curtail output, creating delays in downstream production of plastics and fuels that rippled across manufacturing. To mitigate production backlogs, lean manufacturing principles, including kanban systems, visualize workflow via cards or digital boards to limit work-in-progress, prevent overproduction, and signal replenishment needs, thereby reducing delays by up to 70-90% in lead times for adopting manufacturers. Enterprise resource planning (ERP) software further aids by providing real-time tracking of production schedules, inventory, and bottlenecks, enabling proactive adjustments to maintain throughput and minimize unfulfilled orders.Project management
Task backlog
In traditional project management methodologies such as Waterfall, tasks are typically planned upfront in a fixed scope, with any backlog referring to outstanding tasks or deliverables that require assignment, execution, or completion due to delays, to advance the project toward its objectives.[25] This list serves as a repository to track progress against the predefined project scope from the initial planning phase.[25] Prioritization of tasks often employs techniques like the MoSCoW method, which categorizes items as Must-have (essential for success), Should-have (important but not vital), Could-have (desirable if time allows), or Won't-have (excluded for the current scope).[26] Task estimation typically relies on time-based metrics, such as projected hours or man-days, to allocate resources and forecast completion timelines.[27] Common tools for managing and visualizing task lists include Gantt charts, which display task dependencies, durations, and milestones in a timeline format, often implemented via software like Microsoft Project.[28] For instance, in construction projects, delays such as those from permitting processes can stall subsequent activities like site preparation or material procurement, extending overall timelines. Similarly, in other fields like marketing, unexecuted tasks such as content creation can disrupt schedules and resource allocation. Key risks associated with unmanaged task lists include scope creep, where uncontrolled additions of new requirements strain budgets or deadlines without formal change approvals.[29] A historical example is the CONFIRM project, a late-1980s initiative by AMR Information Services to build an integrated travel reservation system, which ballooned from an estimated $50 million to over $600 million in costs due to poor task management, inadequate planning, and escalating unresolved development issues, ultimately leading to its cancellation in 1992.[30]Agile methodologies
In agile software development, the product backlog represents an emergent, ordered list of everything known to be needed in the product, including features, enhancements, bug fixes, and other requirements, serving as the single source of work for the Scrum Team. Owned by the product owner, who is accountable for its content, prioritization, and alignment with the product goal, the backlog evolves through ongoing refinement to reflect changing needs and new insights. Items are often articulated as user stories in the format "As a [type of user], I want [some goal] so that [some reason]," facilitating clear communication of value and requirements.[31][32] The sprint backlog, in contrast, consists of a selected subset of product backlog items chosen for a specific sprint—a time-boxed period typically lasting one to four weeks—along with the sprint goal and an actionable plan for delivering an increment of potentially releasable product. Committed to by the development team during sprint planning, it includes task breakdowns to guide daily work and is updated as progress unfolds, ensuring focus on achievable outcomes within the iteration. This structure promotes transparency and adaptability while limiting scope changes mid-sprint to maintain team velocity.[31] Key processes in managing agile backlogs include refinement, where the Scrum Team adds details, estimates effort (often using story points), and reorders items to ensure the top of the product backlog is "ready" for future sprints; this ongoing activity usually consumes no more than 10% of the development team's total capacity. Velocity, calculated as the average amount of product backlog converted to increment per sprint, helps teams forecast completion rates for upcoming sprints and enables the product owner to predict broader release timelines by analyzing historical trends. Popular tools like Jira support these processes by centralizing backlog prioritization, story point estimation, and sprint integration, while Trello offers a visual, card-based alternative for simpler backlog tracking and collaboration.[31][33][34][35] The backlog concept originated in the Scrum framework, developed independently in the early 1990s by Ken Schwaber and Jeff Sutherland to address complex product development challenges, and first co-presented by them at the 1995 OOPSLA conference. Over three decades, Scrum backlogs have emphasized iterative prioritization and team commitment, distinguishing them from traditional project task lists by integrating continuous feedback and value-driven ordering tailored to software delivery. In comparison to Kanban, another agile method, Scrum backlogs operate within fixed time-boxed sprints for disciplined execution, whereas Kanban uses a continuous-flow backlog that allows dynamic pulling of work without iterations, prioritizing workflow visualization over time-bound commitments.[31][36][37] A notable example is Spotify's squad model, where cross-functional squads function as autonomous mini-teams, each maintaining a dedicated product backlog tied to a long-term mission—such as improving the Android client—to guide feature delivery and ensure alignment with organizational goals through collaborative prioritization among product owners. In the 2020s, adaptations for remote work have amplified the use of digital backlog tools, with teams leveraging platforms like Jira for asynchronous refinement sessions and virtual estimation via tools such as Scrum Poker to overcome time zone barriers and sustain collaborative backlog management in distributed environments.[38][39]Computing
Queue backlog
In computing, a queue backlog occurs when the number of items—such as data packets, requests, or operations—accumulated in a queue data structure exceeds the system's processing capacity, leading to delays. Queues in this context are typically implemented as first-in-first-out (FIFO) structures or priority-based queues to manage orderly processing, and they are fundamental in operating systems for handling asynchronous events like input/output (I/O) and network traffic. This buildup arises when the arrival rate of items surpasses the service rate, as described in queueing theory applications to computer systems.[40] Common types of queue backlogs include I/O backlogs, where read/write requests to storage devices like disks form queues managed by the operating system's block layer. For instance, in the Linux kernel, pending I/O requests are held in per-device queues to optimize access patterns and prevent device overload.[41] Network packet backlogs, on the other hand, occur in routers and host kernels when incoming packets overwhelm processing, such as when the kernel's protocol stack cannot keep pace with high-speed interfaces; each CPU core maintains a separate backlog queue to buffer these packets before further handling.[42] Key metrics for analyzing queue backlogs include queue length, which measures the number of pending items, and increased latency, reflecting the time items spend waiting. These can be quantified using Little's Law from queueing theory, which states that the average queue length L equals the arrival rate \lambda multiplied by the average time W an item spends in the system: L = \lambda W This relationship, proven for stable systems under steady-state conditions, helps predict backlog growth and system performance without needing detailed internal models. Little's Law, a fundamental result in queueing theory applicable to computer systems.[40] Examples of queue backlogs include web server overloads during traffic spikes, such as in distributed denial-of-service (DDoS) attacks throughout the 2010s, where SYN flood techniques exhausted TCP connection queues by sending incomplete handshake requests, causing legitimate HTTP requests to backlog and leading to service denial.[43] In databases, transaction queues can backlog under high concurrency, as seen in systems like PostgreSQL where long-running transactions delay processing of queued operations, amplifying write latency.[44] Management strategies for queue backlogs involve buffering techniques to bridge speed mismatches between producers and consumers, such as double buffering where one buffer fills while another is processed, and load balancing to distribute requests across multiple queues or processors. These approaches trace back to the evolution of operating systems from batch processing in the 1950s–1960s, where jobs were queued offline for sequential execution on mainframes to maximize resource utilization, transitioning to real-time and multiprogrammed environments by the 1970s to handle interactive workloads.[45][42]System backlogs
In computing, a system backlog refers to the accumulation of pending system-level tasks, such as software updates, error logs, or maintenance operations, that await processing or review within hardware and software ecosystems. These backlogs arise when the rate of incoming tasks exceeds the system's processing capacity, leading to delays in execution across operating systems, servers, or distributed environments.[46] Common types of system backlogs include print job queues in spoolers, where multiple print requests accumulate if the printer or spooler service is overwhelmed or stalled.[47] Operating system update backlogs, such as those in Windows Update, occur when security patches and feature updates queue up due to deferred installations or network constraints.[48] Error log backlogs in monitoring tools build up from unprocessed diagnostic data generated by applications or infrastructure, potentially overwhelming storage and analysis capabilities.[49] Key metrics for managing system backlogs include backlog age, which measures the time elapsed since tasks entered the queue, and clearance rate, defined as the proportion of pending items resolved over a period, often improved through automation.[50] For instance, in asynchronous systems, metrics like event age help identify aging tasks that risk becoming obsolete.[50] In cloud computing, AWS Lambda invocation backlogs form when asynchronous events, such as API calls or stream records, exceed concurrency limits, causing invocations to queue and delay processing.[51] Similarly, mobile app crash report backlogs accumulate in tools like Firebase Crashlytics, where unreviewed stability issues from user sessions pile up, hindering rapid debugging.[52] Mitigation strategies emphasize asynchronous processing to decouple task submission from execution, allowing systems to handle bursts without immediate failure, alongside scheduled automation like cron jobs for periodic backlog clearance.[5] These approaches have evolved from batch processing systems of the 1950s–1960s to time-sharing and multiprogramming in the 1970s, and later to cloud-native architectures since the 2010s that leverage scalable queues and auto-scaling for resilient backlog management.[45][53] This progression builds on foundational queue concepts by integrating them into distributed, event-driven ecosystems.[54]Publishing
Academic journals
In academic journal publishing, a backlog refers to the accumulation of peer-reviewed and accepted manuscripts that await allocation to specific issues due to constraints such as limited page budgets or fixed publication schedules.[55] These delays occur primarily after acceptance, as journals prioritize bundling articles into periodic issues, often leading to extended waits beyond the peer review phase.[6] The primary causes of such backlogs include high volumes of submissions overwhelming editorial capacity, inefficiencies in peer review coordination, and traditional print-based limitations on space.[56] Overall submission-to-publication times average 9-18 months across disciplines, with science, technology, and medicine (STM) fields experiencing shorter lags compared to social sciences or humanities.[57] Journals with higher impact factors tend to face larger backlogs, as elevated prestige attracts more submissions, correlating moderately with lower acceptance rates and extended processing.[58] High-impact journals like Nature exemplify these challenges, with total handling times for accepted manuscripts averaging around 4 months in the 2020s, though post-acceptance production backlogs can extend this to 6 months or more amid surging demand.[59] The rise of open-access models has mitigated some delays through online-first publication strategies, allowing accepted papers to appear digitally before print compilation and reducing average review and publication times compared to subscription-based journals.[6] Management of backlogs has evolved significantly with the advent of electronic publishing in the 1990s, when early digital journals and internet distribution began decoupling content from physical constraints, enabling faster dissemination.[60] However, the proliferation of predatory journals—low-quality outlets promising rapid publication for fees—has exacerbated perceptions of delays in legitimate venues by offering illusory shortcuts that undermine trust in rigorous processes.[61] During the COVID-19 pandemic from 2020 to 2022, submissions to academic journals surged, particularly in STEM fields, with many outlets reporting increases of 10-87% in non-pandemic-related manuscripts alongside a wave of COVID-focused research, swelling backlogs by 20-50% in affected disciplines.[62][63] This influx intensified existing pressures, prompting some journals to accelerate online releases to manage the overload.[64]Patent applications
In the context of intellectual property, a patent backlog refers to the accumulation of patent applications that have been filed but await examination by patent office examiners. This pending inventory is tracked and managed by major offices such as the United States Patent and Trademark Office (USPTO) and the European Patent Office (EPO), where it represents applications not yet reviewed for novelty, inventive step, and other patentability criteria.[65][66] Key metrics highlight the scale of these backlogs, including average pendency time—the duration from filing to final disposition—and the total number of unexamined applications. At the USPTO, total pendency averaged approximately 26.1 months as of early 2025, with first office action pendency reaching 20.5 months in fiscal year 2023. The USPTO's backlog exceeded 826,000 unexamined applications in January 2025, having been reduced below 800,000 by September 2025 through targeted initiatives.[67][68][69] The primary causes of patent backlogs include surging application volumes driven by technological advancements, such as the post-2000 tech boom, alongside chronic examiner shortages and resource constraints at patent offices. For instance, global patent filings rose sharply in the late 1990s and early 2000s due to the internet growth spurt, particularly in software-related inventions, which overwhelmed examination capacities and contributed to a backlog surge at the USPTO.[70][71][72] Internationally, the USPTO's backlog is significantly larger—about three times that of the EPO—while the Japan Patent Office (JPO) maintains a comparatively smaller pending inventory, with around 290,000 applications filed in 2022 but lower unexamined stocks due to efficient staffing and processes.[73][74][75] Efforts to address these backlogs have included the Patent Prosecution Highway (PPH), a multilateral initiative launched in the mid-2000s to accelerate examinations by leveraging work products from participating offices. Originating as a pilot between the USPTO and JPO in 2006 and fully implemented by 2008, the PPH allows applicants to request expedited review in one office based on positive determinations from another, thereby reducing duplication and pendency times across jurisdictions like the EPO and JPO.[76][77] In the 2020s, patent offices have piloted AI-assisted tools to further streamline examinations; for example, the USPTO's Artificial Intelligence Search Automated Pilot Program, launched in October 2025, uses AI to generate pre-examination prior art search reports, aiming to enhance efficiency and reduce manual search burdens. Similarly, the EPO has tested AI-driven search technologies to improve prior art identification and throughput.[78][79][80]Arts and entertainment
Music
Backlog is a joint compilation album by the British electronic music duo Leftfield and the project Djum Djum, released in 1992 on the Outer Rhythm label.[81] It assembles remixes and versions from Leftfield's debut singles "Not Forgotten" (1990) and "More Than I Know" (1991), alongside contributions from Djum Djum, capturing the duo's nascent experimentation in electronic sounds.[81] Issued as a CD in the UK on November 30, 1992, the album reflects the burgeoning electronic music landscape of the early 1990s, blending progressive house and techno elements without achieving mainstream commercial success.[81] The tracklist comprises 10 pieces, emphasizing dub and house influences through layered remixes and additional production by Leftfield members Neil Barnes and Paul Daley on select cuts:-
- Not Forgotten (Original Mix) – 6:41
-
- Not Forgotten (Fateh's On The Case Mix) – 6:12
-
- Not Forgotten (Dub Mix) – 4:46
-
- More Than I Know (12" Mix) – 6:43
-
- Not Forgotten (Hard Hands Mix) – 7:37
-
- More Than I Know (10K Mix) – 8:36
-
- More Than I Know (More Mix) – 7:29
-
- More Than I Know (Even More Mix) – 4:22
-
- Difference (Steng Mix) – 7:06 (Djum Djum)
-
- Difference (Cake Mix) – 6:42 (Djum Djum)