R01
Digital product management
What change in behaviour are we betting on?
Work / Guides / Introduction to Craft Guides
Craft guides for educational technologists
Twenty-two short, sourced guides to the practices behind learning products, written for the people who fund, build and run them in schools. Each takes a practice borrowed from technology or design, explains the reasoning behind it, and shows where it breaks in education.
All guides
Why the series exists
General guides stop where education starts
There are good general guides to product management, service design, Scrum and data governance. Almost none of them deals with the things that decide whether a learning product works: annual procurement, the school calendar, children’s data, teachers who choose whether anything reaches a student, and learning outcomes that take years to show.
This series fills that gap. It follows a learning product through its life, from why build it and what to build to how to build it, the data and AI inside it, and whether it works and lasts. Every guide in the series follows the same pattern.
1
Correct a misconception
Each guide opens with a belief people commonly hold about its practice, and shows why that belief fails.
2
Teach the reasoning
Each guide rests on one central idea and five habits, with the logic behind each, so readers can apply them to cases the guide never mentions.
3
Show where it breaks in education
Every habit has a “where it breaks” slide on what changes when the users are teachers and students and the buyer is a school system.
Five constraints
Five constraints recur across the series. Each guide shows them all and marks the two or three that matter most for its subject.
The calendar
Terms, holidays and exams decide when anything can change, be tested or be measured.
Procurement
Annual budgets, tenders and contracts fix scope long before users are consulted.
Children
Minors cannot consent alone, cannot opt out of a national platform, and are owed a higher duty of care.
The gatekeeper
Nothing reaches a student unless a time-poor teacher chooses to use it.
The lag
Learning outcomes take years to show, long after the decisions that shaped them.
Who it is for
Three readers who see the same product from different seats
The habits in each guide apply to everyone. What differs is which part of the product each reader controls, so every guide closes with one slide per reader: what they own, what to ask for, and the mistake they are most likely to make. Each reader has a symbol rather than a colour, so the marks survive black-and-white printing.
Ministry product and policy staff
They set mandates, fund the work, write the rules and own the central systems. Their decisions reach every school at once.
Vendors
They build one part of a larger service and report on how it performs. They are best placed to produce evidence, and have the strongest reasons to present it favourably.
Inside a guide
One file, three views, one skeleton
| Part | What it does |
|---|---|
| Opening | The title, the series shelf, who the guide is for, the five constraints, and a question the reader answers before reading on |
| I · The correction | The common belief and why it fails, then the one central idea the rest of the guide depends on |
| II · The habits | Five sections, each giving one habit, the reasoning and a figure, where it breaks in education, then what to do and what to avoid |
| III · In practice | One slide per reader, then questions to ask of any plan and further reading |
| Reference | A glossary, a methods table, working templates, evidence notes and an APA reference list with links |
- Slides
- Present the guide in a room: about 55 slides that fit any screen, with a Read more panel for detail and print-ready pages.
- Longform
- Read it as an article, with contents, illustrations, appendices and references, printable as one document.
- Posters
- Ten A3 posters, two per habit. The reference poster carries the full habit. The statement poster reduces it to one mark and one line, for the wall and for chat.
- One file
- All three views come from the same content, so a correction made once appears everywhere. Each guide is a single HTML file that works offline.
What each guide contains
Twenty-two guides in five stages
Each card gives the guide’s central question, the belief it corrects and its habits.
Strategy and value
Why build
2 guides
WrittenOpen (opens in a new tab)
R02
Service design and systems thinking
What does the whole experience look like, beyond our screen?
Discovery and design
What to build
6 guides
R03
Design thinking, used honestly
Are we solving the right problem?
R04
Discovery research: interviews, observation and context
What do people actually do, and why?
R05
Evaluative research: usability testing, surveys and UX metrics
Does what we made work for the people using it?
R06
Design methods I: making sense of research
What did we learn, and what does it mean for the design?
R07
Design methods II: sketching, prototyping and critique
What is the cheapest way to find out if this idea works?
R08
UX foundations and design systems
Why do interfaces help or hinder?
Delivery
How to build
5 guides
R09
Agile and Scrum in practice
How do we deliver value in small, safe steps?
R10
Lean product development
What is the smallest thing we can do to learn whether we are right?
R11
Software development for non-engineers
What is actually happening when engineers build?
R12
DevOps and continuous delivery
How do we release often without breaking a school day?
R13
Experimentation and product analytics
Did this change do what we hoped?
Governance, analytics and models
How to improve
4 guides
R14
Data strategy and governance
What data do we need, and who is responsible for it?
R15
Learning analytics
What can logs tell us about learning?
R16
AI in learning products
Where does AI help learning, and how do we know?
R17
Privacy, security and ethics
What could go wrong with student data, and for whom?
WrittenNot yet online
Evidence, change, communication and support
How to measure & scale
5 guides
R18
Basic statistics for educational technologists
What can a number tell you, and how far should you trust it?
R19
Adoption and change in schools
Why do good tools go unused?
R20
Communicating change to schools
Does the right person know what changed, why, and what to do?
R21
Professional learning that changes practice
What kind of training actually changes what teachers do?
R22
Support and feedback loops
When something goes wrong, how fast do we hear and fix it?
How the series was designed
The decisions behind the series, and what each one costs
01Organise by the life of a product
Decision
The guides are grouped into five stages: why build, what to build, how to build, how to improve, and how to measure and scale.
Why
Readers can find the guide for the decision in front of them, and the order shows how early choices constrain later ones.
Trade-off
Some topics span stages. Privacy, for example, matters from the first discovery interview, so guides point to each other.
02Colour identifies the stage, and shade identifies the guide
Decision
Each guide has its own tinted background: yellows for strategy, corals and orchids for design, blues for delivery, violets for data and AI, and greens for evidence. Each colour has a light and a dark version.
Why
A reader holding any guide can tell its stage at a glance, and the series reads as one set on a website or a shelf.
Trade-off
Fitting 22 colours around the colour wheel leaves neighbouring guides only just distinguishable. Accent colours are darkened guide by guide so text stays readable, with every contrast ratio measured rather than assumed.
Why buildWhat to buildHow to buildHow to improveHow to measure & scale
03Only the colours change from guide to guide
Decision
Every guide keeps the same typefaces, blue and orange highlight colours, and colours for do, don’t, evidence and contested points. Only the background and one signature colour change.
Why
A reader who learns the visual language in one guide can read all of them. One shared template carries every guide, so a fix made once reaches all of them.
Trade-off
Individual guides have less room for their own visual identity.
04Five habits per guide
Decision
Each guide rests on one central idea and five habits. The original outlines had three.
Why
Five habits cover the topic without becoming a checklist, allow a full set of posters, and leave room for the habits that are specific to education.
Trade-off
Some topics would split more naturally into four or six.
05Symbols rather than colours for the three readers
Decision
Ministry staff, vendors and school leads are marked ■ ▲ ●, in ordinary text colour.
Why
Colour already signals the guide, the section and whether a point is a do, a don’t, evidence or contested. A fourth colour code would overload it, and symbols print cleanly in greyscale.
Trade-off
Symbols are less striking than colour, so the reader’s name always appears beside the symbol.
06The same five constraints in every guide
Decision
Every guide shows the same five constraints and marks the ones that matter most for its topic.
Why
It makes “where it breaks in education” a consistent thread rather than a loose aside, and lets readers compare guides.
Trade-off
Other factors, such as language, equity and infrastructure, are handled inside individual guides rather than as a sixth constraint.
07Deck, article and posters in one file
Decision
Every guide is one self-contained HTML file with three views and its own print settings.
Why
A guide has to work in a meeting, as a read and on a wall. Generating all three from one set of content keeps them from drifting apart.
Trade-off
The files are large and are edited through their content files rather than by hand. Tools that strip inline graphics need an image-only copy.
08Split research and design into four guides; fold DesignOps into another
Decision
The first plan had one user research guide. It became four: discovery research, evaluative research, and two on design methods. DesignOps became a section of the UX foundations guide.
Why
Research and design methods are where readers most often need specifics. DesignOps matters mostly as the practice of running a design system.
Trade-off
The design stage is the largest in the series, with six guides.
09Delivery and adoption carry more weight
Decision
Guides on lean product development and DevOps were added. Build, buy or partner, accessibility, and evaluating edtech impact were dropped. Adoption gained guides on communication, professional learning, and support.
Why
Adoption, not launch, decides impact, and much of it depends on how change is explained, taught and supported in schools.
Trade-off
Accessibility and evaluation are covered inside other guides rather than on their own, which gives them less depth.
10Public sources only
Decision
Every example and figure comes from public, documented sources. Invented cases are labelled as illustrative.
Why
The guides are published openly and must be safe to share with vendors and schools.
Trade-off
Some of the most instructive internal cases cannot be used.
Two series, two axes
This series explains how learning products are made; the context series explains the world they enter
CRAFT SERIES · 22 GUIDES
CONTEXT SERIES · 24 PRIMERSGuides for educational technologists
Organised by the life of a learning product, from why build it to whether it lasts. Each guide corrects a misconception, teaches five habits and shows where the practice breaks in education.
Understanding the Context of Education
Organised from the purposes of education outward. Written for a general reader, it supplies the knowledge the craft guides assume: aims, learners, pedagogy, curriculum, assessment, policy, research and reform.
Read the introductionThe two sets can be read together. Craft guide 14, on data strategy, reads differently once assessment systems (C14) has shown why any reported measure becomes a target. Craft guide 19, on adoption, is sharper after education reform (C23) has explained why system change is slow.
The evidence standard
Every claim is checked before it is used
Before drafting
Before each guide is written, every number, quotation and citation it will use goes to a separate verification pass. Corrections are made before drafting. A claim that cannot be verified is reworded, qualified or dropped. Popular figures that rest on weak evidence are named and set aside in the text rather than quietly omitted.
Inside every guide
Each guide ends with an evidence appendix that records vendor data, secondary quotations, differences in how figures were measured, and the limits of the studies it relies on. The fastest-moving topics, AI above all, date every claim and are due for regular revision. Content that touches law, such as the privacy guide, is checked against current policy before release.
Craft guides for educational technologists · All twenty-two written · September 2026
Singapore