Core Primitive
Produce multiple outputs in one focused session rather than one at a time.
Four videos in one day
A YouTuber with two million subscribers once described his production schedule in an interview that stunned his audience. He does not film one video per week. He films four videos in a single day, once a month. Four sets, four scripts, four performances — back to back, in a rented studio, over eight hours. Then he spends the rest of the month editing, promoting, and planning the next batch.
His viewers see a weekly upload cadence. What they do not see is that the weekly cadence is powered by a monthly production session. The creator is not more talented than the thousands of YouTubers who film one video at a time, scrambling each week to set up equipment, find the right lighting, get into performance mode, and produce a single output before tearing everything down. He is more structurally efficient. He pays the setup cost once and amortizes it across four outputs. He enters performance mode once and stays there. He rents the studio for one day instead of four.
This is batching. And it is not a hack for content creators. It is a production principle that applies to any output you produce regularly — and that becomes essential once you commit to the high-frequency shipping cadence from the previous lesson.
You learned to ship early and ship often. Now you need the production strategy that makes "often" sustainable without burning through your cognitive budget one output at a time.
The primitive: produce multiple outputs in one focused session
Batching means grouping similar outputs together and producing them in a single focused session rather than producing each one individually as it comes due.
The principle is simple. The reasoning behind it is grounded in decades of research on how human cognition handles — and fails to handle — task transitions.
When you switch from one type of work to another, your brain does not flip a clean toggle. It drags residual attention from the previous task into the new one. This residual attention — what researchers call "attention residue" — degrades your performance on the new task until it fully dissipates, which takes far longer than most people realize.
Sophie Leroy, in her 2009 study at the University of Minnesota, demonstrated that when people switch from Task A to Task B, their performance on Task B suffers measurably because part of their attention remains anchored to Task A. The effect persists even when Task A was completed before the switch, though it is worse when Task A was left unfinished. Leroy's finding: "People need to stop thinking about one task in order to fully transition their attention and perform well on another. Yet, results indicate it is difficult for people to transition their attention away from an unfinished task."
The implication for output production is direct. If you write a lesson, then switch to answering emails, then switch to editing a different lesson, then switch to a meeting, and then switch back to writing — you are paying the attention residue tax at every transition. Each switch costs you cognitive clarity. Each recovery period eats into your productive time.
Gloria Mark, Daniela Gudith, and Ulrich Klocke at the University of California, Irvine, found that it takes an average of twenty-three minutes and fifteen seconds to fully return to a task after an interruption. Even self-initiated interruptions — where you choose to switch tasks — carry significant recovery costs. The twenty-three-minute figure has become one of the most cited statistics in productivity research, and for good reason: it means that a knowledge worker who switches tasks ten times in a day loses nearly four hours to recovery alone.
Batching eliminates most of these switches. Instead of producing one output, switching to something else, and then producing the next output later (paying the full setup and recovery cost each time), you produce multiple outputs of the same type in sequence. You enter the cognitive mode once. You stay in it. You exit once. The setup cost is paid once. The recovery cost is paid once. Every output after the first is produced at the marginal cost of production, not the full cost of production plus transition.
Shingo and the changeover problem
The manufacturing world solved this problem decades before knowledge workers recognized they had it.
In the 1950s and 1960s, Japanese industrial engineer Shigeo Shingo confronted a bottleneck at Toyota's manufacturing plants. Changing a stamping press from producing one part to producing a different part required hours of downtime — disassembling dies, installing new ones, calibrating, running test pieces. The changeover time was so expensive that the rational response was to produce enormous batches of each part before switching, which created massive inventory costs and destroyed flexibility.
Shingo's solution was SMED — Single-Minute Exchange of Dies. Rather than accepting the changeover time as a fixed cost, he systematically analyzed every step in the changeover process and divided them into two categories: external setup (work that could be done while the machine was still running the previous batch) and internal setup (work that required the machine to be stopped). By converting as much internal setup as possible to external setup, Shingo reduced changeover times from hours to minutes — sometimes to under sixty seconds.
The result was revolutionary. When changeover time drops to near zero, the optimal batch size drops too. Toyota could produce smaller batches more frequently, reducing inventory costs and increasing responsiveness. The entire lean manufacturing movement grew from this insight.
Your cognitive changeover costs follow the same structure. When you sit down to write a lesson, there is external setup (gathering research, reviewing the skeleton, opening your editor) and internal setup (loading the phase context into working memory, finding the right voice and tone, warming up your editorial judgment). The external setup can be done in advance — you can prepare all your materials before the production session begins. The internal setup cannot be shortcut, but it can be amortized: once you have loaded the context and found the voice, producing the second, third, and fourth lesson in that voice costs far less than loading it fresh each time.
Shingo's principle, translated to knowledge work: separate your preparation from your production, do the preparation in advance, and then batch the production to amortize the irreducible internal setup cost across as many outputs as possible.
The maker's schedule
Paul Graham, in his 2009 essay "Maker's Schedule, Manager's Schedule," articulated a distinction that explains why batching is not just efficient but necessary for anyone who produces substantive work.
Managers operate on a schedule divided into one-hour blocks. Any block can hold any meeting or task. Switching between blocks is cheap because the work in each block is relatively self-contained — a decision, a conversation, a review. The manager's schedule is optimized for responsiveness and variety.
Makers — writers, programmers, designers, anyone who produces complex output — operate on a different schedule entirely. A maker needs long, unbroken stretches of time to enter the cognitive state required for production. The startup cost is high. The maintenance cost is low once you are in the state. But a single interruption can destroy an hour of productive momentum, because the recovery cost (Mark's twenty-three minutes, Leroy's attention residue) is catastrophic relative to the length of the work session.
Graham's observation: "A single meeting can blow a whole afternoon, by breaking it into two pieces each too small to do anything hard in."
Batching is the maker's production strategy. It takes Graham's insight — that makers need long, unbroken time — and applies it not just to protecting time from interruptions but to structuring production itself. Instead of scattering your outputs across the week, each one requiring its own startup, each one vulnerable to the interruptions that fragment the intervening time, you consolidate production into dedicated batch sessions. You protect those sessions the way Graham says makers should protect their afternoons. And you produce your week's outputs in a fraction of the total time they would have taken individually.
The economics of batching
The case for batching is fundamentally economic. Every output you produce has two cost components:
Fixed cost per session — the cognitive changeover cost. Loading context, finding your voice, warming up your judgment, configuring your environment. This cost is roughly the same whether you produce one output or five in the session. It is the price of entry.
Variable cost per output — the actual production work. Research, drafting, editing, polishing. This cost scales linearly with the number of outputs, but it often decreases slightly per unit within a batch because your pattern recognition improves and your decision speed increases as you progress through similar items.
When you produce one output per session, you pay the full fixed cost for that single output. The total cost is: Fixed + Variable.
When you batch four outputs in one session, the fixed cost is paid once and divided across four outputs. The total cost per output drops to: (Fixed / 4) + Variable.
If your fixed cost is forty minutes (context loading, environment setup, voice-finding) and your variable cost is fifty minutes per output, producing four outputs individually costs 4 x 90 = 360 minutes. Producing four outputs in a batch costs 40 + (4 x 50) = 240 minutes. You save 120 minutes — a 33% reduction — with no loss in quality. The savings come entirely from eliminating redundant setup.
In practice, the savings often exceed this estimate because the variable cost per output also decreases within a batch. By your third lesson in a session, you are not pausing to wonder about tone or structure. Your editing eye is sharper. Your phrasing comes faster. The production function is warmed up.
Cal Newport, in his work on deep work and time blocking, describes a related phenomenon: the "deep work session" as a production container. Newport advocates scheduling fixed-length blocks dedicated to a single type of cognitively demanding work. Batching is the natural extension of time blocking — instead of blocking time for "writing" and producing whatever comes, you block time for "writing three lessons" and produce all three in that block.
What is batchable and what is not
Not every output type benefits equally from batching. The key variable is cognitive similarity — how similar the outputs are in terms of the mental mode they require.
Highly batchable outputs:
- Blog posts or lessons on related topics within the same domain
- Social media posts derived from a single source (a lesson, an article, an event)
- Email responses of the same type (client updates, vendor communications, team check-ins)
- Code reviews within the same codebase
- Meeting agendas for a recurring meeting type
- Data analysis reports using the same methodology
These outputs share a cognitive mode. Writing the second one requires the same mental posture as the first. The changeover cost between items is minimal.
Poorly batchable outputs:
- A technical report followed by a creative pitch
- A performance review followed by a product specification
- A mathematical proof followed by a marketing email
- Any combination of outputs that require fundamentally different cognitive modes
When the cognitive mode shifts between outputs, you re-introduce the very switching costs batching was designed to eliminate. Batching dissimilar outputs is worse than not batching at all, because you suffer the switching costs in a compressed timeframe with no recovery breaks.
The diagnostic question: "Does the second output require me to load a different mental context, or can I stay in the same mode?" If the answer is "same mode," batch. If the answer is "different mode," separate.
Designing a batch session
A productive batch session has three phases, and the critical insight is that the first phase happens before the session begins.
Phase 1: Preparation (before the session). This is Shingo's external setup. Gather every input you will need during production. For a lesson-writing batch: have all skeletons printed or open, all research notes compiled, all predecessor and successor contexts loaded, all templates ready. The goal is that when the session clock starts, you produce. You do not search, gather, organize, or decide what to work on. Every decision about what to produce and what materials you need was made during preparation. The session itself is pure execution.
Phase 2: Production (the session). Start with the easiest or most familiar output to build momentum. Once you are warmed up, move to the more demanding outputs while your cognitive resources are at their peak. Save the most formulaic output for last, when your creative energy may be lower but your pattern-matching is still strong. Do not check email, messages, or notifications during the session. Every interruption re-introduces the switching cost that batching exists to avoid.
Phase 3: Quality review (after the session or after a break). Batch the review separately from the production. This is the draft-quality separation from First drafts are for content final drafts are for quality applied to batch production: first drafts are for content, final drafts are for quality. Produce all outputs in the batch, then review all outputs in the batch. The review itself benefits from batching — your quality standards are calibrated by reviewing multiple outputs in sequence, and inconsistencies between outputs become visible in a way they never are when you review each output in isolation days apart.
A practical heuristic for session length: most knowledge workers can sustain high-quality production in a single cognitive mode for two to four hours. Beyond four hours, fatigue typically degrades output quality enough to offset the changeover savings. Start with two-hour batch sessions and extend only if quality holds.
The rhythm of batch production
Batching changes not just how you produce but when you produce. Instead of producing outputs on the day they are due, you produce them in advance during batch sessions and release them on schedule.
This is the content creator's secret. The YouTuber films four videos in one day and releases them weekly. The newsletter writer drafts four issues on Sunday and publishes them each Tuesday. The social media manager creates two weeks of posts in a single three-hour session and schedules them for daily release.
The production rhythm and the publication rhythm are decoupled. Production is batched for efficiency. Publication is distributed for audience consistency. The two schedules operate independently.
This decoupling provides a crucial buffer. If you get sick on Tuesday, your Wednesday output was already produced last Saturday. If a meeting eats your Thursday morning, your Thursday deliverable was already done. The batch creates inventory — not the wasteful inventory Shingo fought in manufacturing, but strategic inventory that absorbs the variability of real life without disrupting your output cadence.
The connection to Ship early ship often is direct. Ship early, ship often — but produce in batches. The shipping cadence is daily. The production cadence is weekly (or whatever interval your batch sessions follow). These are not contradictions. They are complementary rhythms: one governs when the world sees your work, the other governs when you do the work.
Batching beyond content
While content creation is the most obvious application, batching applies to any recurring output.
Decision batching. Instead of making budget decisions one at a time as requests arrive, hold a weekly budget review where you evaluate all pending requests together. The context is loaded once. The criteria are applied consistently. And decisions made side by side are better calibrated than decisions made in isolation, because you can compare the relative merits of competing requests.
Communication batching. Instead of responding to emails throughout the day — each response requiring you to load the sender's context, formulate a response, and then recover from the switch — process all emails in two or three dedicated blocks. The cognitive mode is "correspondence mode." You enter it, clear the queue, and exit.
Review batching. Instead of reviewing team members' work as it arrives, batch all reviews into a single session. Your evaluative criteria sharpen across items. You notice patterns — "three people made the same error, which suggests a systemic issue, not an individual one" — that would be invisible if you reviewed each item days apart.
Planning batching. Instead of planning each day independently each morning, batch your weekly planning into a single session where you allocate the entire week's priorities at once. The strategic perspective is loaded once. The tradeoffs between competing priorities are visible simultaneously.
In each case, the principle is identical: group similar cognitive work, pay the setup cost once, and amortize it across multiple outputs.
The limits of batching
Batching has failure modes that are worth naming explicitly so you can avoid them.
Over-batching. Producing too many outputs in a single session pushes past the fatigue threshold where quality degrades. If output number seven in your batch is measurably worse than output number two, your batch is too large. The fix is not to push harder but to split into two smaller batches with a recovery period between them.
Staleness. Outputs produced far in advance of their release can become stale if the context changes. A social media post written three weeks before publication may reference information that has been superseded. A report drafted two months before delivery may no longer reflect current data. The buffer that batching creates is an asset, but the buffer has a shelf life. Match your batch lead time to your output's freshness requirements.
Rigidity. An extreme batching practice can make you unresponsive to time-sensitive opportunities. If you batch all your writing on Saturdays and a compelling topic emerges on Tuesday, you need the flexibility to produce outside the batch. Batching is a production strategy, not a production prison. The batch handles your predictable, recurring outputs. Your ad hoc capacity handles the unpredictable ones.
Preparation neglect. The most common failure. You schedule a batch session, sit down, and realize you have not prepared the inputs. Now your "production session" becomes a "scrambling session" — gathering research, finding templates, making decisions about what to produce. The changeover savings evaporate because you never did the external setup. The rule is absolute: preparation happens before the session, not during it.
Your Third Brain: AI-assisted batch production
AI transforms batch production from efficient to formidable. The combination of batching and AI assistance attacks both cost components simultaneously — the fixed cost of setup and the variable cost per output.
Batch preparation. Before your production session, ask your AI assistant to compile all the inputs you will need: research summaries for each output, predecessor context for each lesson, key points to cover, relevant quotes and data. What would take you thirty minutes of gathering per output takes the AI five minutes for the entire batch. Your preparation phase collapses. You arrive at the production session with everything staged.
Template-assisted drafting. For outputs that share a structure (and batchable outputs, by definition, share a structure), AI can generate structural scaffolds for each item in the batch. You are not writing from a blank page. You are writing from a scaffold that already has the section headers, the key transitions, and placeholder content for each section. Your job is to replace the placeholders with your thinking — which is the high-value, irreducibly human part of the work.
Momentum maintenance. Between outputs in a batch, there is a micro-transition: you finished one and need to start the next. AI can bridge this by pre-loading the context for the next output while you take a two-minute break. When you return, the next output's brief, structure, and key inputs are already on screen. The micro-transition cost drops to near zero.
Batch quality review. After producing the batch, AI can perform a first-pass consistency check: are the tone, depth, and style consistent across all outputs? Are there accidental repetitions of phrases or examples? Does the difficulty progression make sense? You still do the final quality review — the AI cannot judge whether your argument is actually sound — but the mechanical consistency check is handled.
The human-AI division in batch production mirrors the human-AI division in every lesson: AI handles preparation, structure, and consistency. You handle judgment, voice, and meaning. The batch session becomes a focused burst of the high-value work, with the low-value work either pre-completed by AI or deferred to AI for post-processing.
The bridge to the output pipeline
Batching is a production strategy. It answers the question: "How do I produce outputs efficiently?" But it does not answer the larger question: "How does an output move from idea to delivered artifact?"
That larger question requires a pipeline — a defined sequence of stages that every output passes through, with clear criteria for advancement between stages. The output pipeline formalizes what this phase has been building piece by piece: output types are defined (Define your output types), quality standards are set (Output quality standards), checklists enforce consistency (The output checklist), drafts are separated from finals (First drafts are for content final drafts are for quality), templates reduce startup friction (Output templates reduce startup friction), minimum viable outputs set the quality floor (The minimum viable output), frequency establishes the cadence (Output frequency matters), shipping early keeps the cycle fast (Ship early ship often), and batching makes the production rate sustainable (this lesson).
In the next lesson, we formalize these components into a single, end-to-end pipeline: the stages, the transitions, the quality gates, and the flow of outputs from raw idea to published artifact. Batching is one stage in that pipeline — the production stage. But the pipeline as a whole is what turns occasional output into a reliable production engine.
You now have the production strategy. Next, you build the system that contains it.
Sources:
- Leroy, S. (2009). "Why is it so hard to do my work? The challenge of attention residue when switching between work tasks." Organizational Behavior and Human Decision Processes, 109(2), 168-181.
- Mark, G., Gudith, D., & Klocke, U. (2008). "The cost of interrupted work: More speed and stress." Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 107-110.
- Monsell, S. (2003). "Task switching." Trends in Cognitive Sciences, 7(3), 134-140.
- Shingo, S. (1985). A Revolution in Manufacturing: The SMED System. Productivity Press.
- Graham, P. (2009). "Maker's Schedule, Manager's Schedule." paulgraham.com.
- Newport, C. (2016). Deep Work: Rules for Focused Success in a Distracted World. Grand Central Publishing.
- Ohno, T. (1988). Toyota Production System: Beyond Large-Scale Production. Productivity Press.
- Allen, D. (2001). Getting Things Done: The Art of Stress-Free Productivity. Viking.
Frequently Asked Questions