In 2018, Nintendo handed children sheets of pre-cut cardboard and invited them to build working controllers for a video games console.
One kit became a piano they could actually play, another included a steering wheel and accelerator pedal capable of controlling a car in Mario Kart. Children could also build controls for a submarine, plane, motorbike, fishing rod and robot, then connect their cardboard creations to a Nintendo Switch.
Some builds took several hours. They involved scores of parts, folds, tabs, moving mechanisms and precise sequences. Yet Nintendo designed Labo so that children of around six could take part with an adult, while those around ten could often complete a build independently.
Nintendo had not found a way to give six-year-olds exceptional memories or advanced spatial reasoning - it had designed a better way to communicate the work.
The result offers a useful lesson for any OEM asking technicians, production teams, dealers or customers to understand a complex product.
A conventional instruction booklet or manual would have made Nintendo Labo considerably harder.
The builder would have needed to identify a cardboard component, match it to a printed drawing, work out which way up it should be held, interpret the required movement and remember the previous steps. A small misunderstanding early in the process could remain hidden until several other parts had been added.
Instead, Nintendo delivered the procedure through the Switch.
The on-screen instructions showed one manageable step at a time. Builders could move forwards or backwards, pause when they needed to, rotate the model and zoom in on an area. An action could be animated repeatedly until the movement, orientation and intended result were clear.
The user did not have to imagine how a flat drawing translated into a physical fold. They could see it happen.
Nintendo also designed the experience to accommodate different working speeds. Its guidance explicitly encouraged people to take breaks and build at their own pace. The instruction moved when the builder was ready by clicking and holding down the Forward button, rather than forcing the builder to keep up.
The Vehicle Kit, for example, combined interactive build instructions with tutorials that allowed users to look inside their creations and understand how the physical mechanism worked with the Switch’s digital technology.
This was more than a manual placed on a screen. The product, procedure and interface had been designed as one experience.
Nintendo’s first attempt did not work nearly as well.
In a developer interview about Labo’s testing process, the team described taking what was already a near-final product to user trials in the United States and Tokyo. They expected children to complete most of the builds successfully.
The results were described as disastrous.
The most worrying problem was not that children made mistakes. It was that some believed they had completed the build correctly when they had not. One child even suggested that the kit might be too difficult for younger children, despite their own model being among the least successful in the test.
For a toy, that might lead to frustration when the finished creation does not work. In manufacturing or field service, the equivalent is much more serious. A technician can miss a component, use the wrong orientation or skip a safety-critical step while remaining confident that the procedure has been followed correctly.
Nintendo went back into development.
The team reduced the number of parts, simplified difficult operations and prioritised ease of assembly over appearance. More than ten tab designs were tested to find versions children could insert reliably. Printed areas were added to distinguish the front and back of each cardboard sheet. Operations that appeared to require three hands were redesigned so that two were enough.
More than 100 children eventually took part in the testing. Because the cardboard designs and software were still changing, Nintendo created temporary instructions from sequences of photographs. Individual builds required between 1,000 and 3,000 images, all of which had to be retaken when a design changed.
The team watched where children stopped, misunderstood an action or carried on incorrectly. It then used that feedback to improve both the physical product and what Nintendo called its “touchable assembly videos”.
The developers believed the result addressed familiar problems with printed instructions, where a sequence can suddenly jump forward or a small but important change can be difficult to see. Video, rotation and zoom made the transition visible.
One developer concluded:
“We think it’s the coolest instruction manual in the world!”
(Quote translated from Nintendo’s Japanese developer interview.)
It is a playful quote, but there is serious design work behind it. Nintendo did not assume that adding video would solve the problem. It observed real people doing real work, found where their understanding broke down and redesigned the instruction around those moments.
The Nintendo Labo approach aligns closely with several established principles from cognitive psychology.
A technical manual often asks the user to convert a 2D representation into a three-dimensional action.
The technician must identify the physical assembly, find the corresponding drawing, recognise the relevant component, infer its orientation and then translate an arrow or written instruction into movement.
Every stage creates an opportunity for ambiguity.
Interactive 3D work instructions and SOPs (Standard Operating Procedures) can do more of that translation in advance. The user can view the product from a familiar angle, rotate it, isolate an assembly, identify the component and watch the required action in context.
Research into the design of assembly instructions has found that people respond better to guidance that explicitly depicts the required attachment or action. While not a complex build for some, in experiments involving the assembly of a TV stand, highly rated instructions helped participants work faster and make fewer errors. Instructions were clearer when each significant operation was shown as an individual step, while enough of the existing assembly remained visible to provide context.
A structural drawing shows where the parts eventually belong. An action-led instruction shows how to get them there.
Working memory is the part of cognition used to hold and manipulate information while we complete a task, but its capacity is limited.
A technician may already be thinking about the fault, surrounding components, safety conditions, tools, fasteners and the consequences of making a mistake. If the instruction also requires them to retain a paragraph of text while searching for the corresponding area of a drawing, some of that limited capacity is spent navigating the documentation.
Richard Mayer and Roxana Moreno’s work on multimedia learning, often used in Learning Management Systems (LMS). explains that people process verbal and pictorial information through partially separate channels, but both have limited capacity. Cognitive overload occurs when the processing demanded by the instruction exceeds what the user can handle at that moment.
Well-designed and contextual visual instructions can externalise part of the thinking therefore. The sequence, orientation and intended result remain visible instead of depending entirely on recall.
As researchers Barbara Tversky, Julie Morrison and Mireille Betrancourt put it:
“Graphics externalize internal knowledge… reducing the burden on memory and processing by off-loading.”
Their research also carries an important warning: animation is not automatically superior. If it moves too quickly, includes too much information or gives the viewer no control, it can create further cognitive demand - which is precisely why anyone learning to fix a product via YouTube will find themselves watching, re-watching, rewinding and fast forwarding through the video and inevitable adverts.
Nintendo avoided this by letting the builder stop, replay, rotate and inspect each action. The value came from purposeful animation combined with interaction and control.
A 45-step repair is still a 45-step repair. Digital instructions do not remove the underlying complexity, but they can stop the user having to process it all at once.
Mayer and Moreno describe this as the segmentation principle. People understood complex multimedia explanations better when the content was divided into learner-controlled sections rather than presented as one continuous sequence.
This is exactly what Labo did. The child completed the action in front of them, checked the outcome and then chose when to continue.
For a technician, the same approach could mean:
The complete process remains governed by the OEM, but the technician only needs to concentrate on the current action and the information required to perform it correctly.
Traditional manuals can split attention between text, drawings, reference tables and the physical product. The user repeatedly looks away from the work, searches the document and then attempts to relocate the same area on the machine.
Cognitive research calls this the split-attention effect.
Clear labels, highlights, arrows, warnings and short instructions can reduce unnecessary visual searching when they appear close to the relevant component or action. In multimedia learning research, this is supported by the principles of signalling and spatial contiguity.
For manufacturing work instructions and SOPs, that could mean showing a torque value at the fastening step, highlighting the connector to be released or displaying a warning immediately before pressure is removed from a system.
The information appears where it influences the decision. It does not remain buried in a general precautions section or a table several pages away.
Classroom training remains valuable because it develops technical principles, safety awareness, diagnostic ability and professional judgement.
The problem comes when training is expected to carry every detail of every procedure indefinitely.
A field technician might learn about a product during an initial course and not encounter a particular repair for another six months. They may also support several OEMs, multiple product generations and hundreds of possible configurations.
A 2025 meta-analysis of procedural skill retention and decay, covering 1,344 effect sizes from 457 reports, found that longer periods without use were associated with greater skill decay. Across the research, approximately half of the initial performance gains measured through accuracy were lost after 6.5 months without use.
That does not make classroom training ineffective, but shows why training and task-level support should work together.
Training gives the technician the capability to do the job. Product understanding in the flow of work gives them the current, product-specific knowledge needed to apply that capability to the asset in front of them.
Nintendo Labo gives us an engaging example, but similar principles have also been tested specifically in mechanical assembly and maintenance too.
A 2026 experiment published in Applied Sciences compared paper documentation with a sequential digital procedure for assembling an unfamiliar mechanical subassembly.
Image taken from the Multidisciplinary Digital Publishing Institute (MDPI) website
The digital procedure combined text, reference images, video, visual highlights and checking steps. Average assembly time fell from approximately 18 minutes to 10 minutes, a reduction of 44.75%. The researchers recorded 56 errors with paper documentation and nine with the digital procedure. Interventions from another person fell from 29 to one.
The sample was small, with five assembly cycles in each condition, so these figures should be treated as evidence from one controlled task rather than a universal performance promise. Even so, the scale of the differences is worth attention. The largest improvements came from reducing interpretation, keeping the sequence clear and helping inexperienced users work more independently.
The lesson from Labo is not that every manual should be converted into a spinning model.
A confusing 3D instruction is still a confusing instruction. Too many labels, unnecessary movement, long animations and a poorly chosen viewing angle can overwhelm the user as easily as dense text.
Effective digital work instructions need to combine several elements:
Nintendo succeeded because it combined interaction with relentless user testing and thoughtful physical design. It simplified the procedure without removing the detail needed to complete it.
OEMs need to apply the same discipline to their own products.
Most OEMs already possess the knowledge required to support their products.
It exists across CAD models, engineering data, technical publications, bills of materials, service bulletins, SOPs and experienced engineers. The problem is turning those sources into guidance people can understand and apply during the task.
A 2D drawing may be technically accurate while remaining difficult for an unfamiliar technician to interpret. A 300-page manual may contain the correct procedure but still leave someone searching for page 116 while a machine remains out of service. A classroom course may be excellent, but it cannot prepare every attendee to remember every product variation months later.
Digital 3D work instructions provide a different route.
With DRVEN, OEMs can turn CAD models, engineering information and existing documentation into guided, task-level procedures. Users can understand the product in 3D, follow the correct sequence and access supporting information through the device available to them.
The same approach can support:
This helps OEMs improve first-time fix by reducing uncertainty around the product, the component and the process. It also allows experienced support teams to focus on genuine exceptions instead of repeatedly explaining standard procedures.
For a broader look at the operational case, read Digital Work Instructions: How OEMs Turn Knowledge into Faster Repairs.
Nintendo Labo set itself a demanding standard.
Could a person with limited experience take an unfamiliar collection of parts, follow the approved sequence and produce a functioning result without having to memorise the process or continually ask for help?
Nintendo did not lower the ambition of the finished product. Children still built working pianos, steering controls, pedals and mechanical systems that interacted with video games. It lowered the cognitive barriers between the user and the task.
Complex machinery will always require trained, capable people. Digital work instructions do not remove the need for skill, experience or judgement.
They do remove the assumption that skilled people should also have to memorise every procedure, decipher every 2D drawing and search every manual before they can put that expertise to work.
If a six-year-old can help build a functioning steering wheel and accelerator from flat sheets of cardboard, imagine what a trained technician can achieve when the OEM gives them equally thoughtful guidance?!