Backward design is a three stage planning sequence in which a teacher first sets the learning outcome, then decides what evidence would prove learners have reached it, and only then plans the lessons. Popularised by Grant Wiggins and Jay McTighe, it replaces planning that begins with activities or with the next page of the coursebook.
The short version
Most planning runs forwards: pick a topic, pick some activities, teach them, then write a test at the end that fits whatever happened. Backward design runs the other way. You name the thing a learner should be able to do, you write down what would count as proof, and the lessons become whatever gets them there. Stage two is the one teachers skip, and skipping it is how a course fills up with good activities that nobody can show learners have learned from.
Part of the TEFL teaching methods and classroom techniques glossary.
The three stages
The framework is set out by its authors in three stages, and the order is the whole point. Each stage answers a different question, and each one has to align with the others: what you named in stage one is what you assess in stage two and what you teach in stage three.
| Stage | The question it answers | What it produces in a language course |
|---|---|---|
| 1. Identify desired results | What should learners know, understand and be able to do by the end? What is the transfer we actually want? | A small set of can do statements, usually two or three, written in the language of real performance rather than of grammar coverage. |
| 2. Determine acceptable evidence | How will we know they got there? What would we accept as proof? | One performance task that looks like the real situation, plus lighter evidence such as a quiz, a recording or a marked piece of writing. |
| 3. Plan learning experiences and instruction | What will equip them to perform well on that evidence? | The syllabus: the lexis, the structures, the skills work and the practice, chosen because the task needs them. |
The authors describe the framework as guarding against two specific failures. One is treating the coursebook as the curriculum rather than as a resource. The other is activity oriented teaching, in which lessons are engaging but carry no clear priority or purpose. Both are recognisable in language teaching, and the second is the more comfortable of the two, because a class that is enjoying itself rarely complains.
How it is used in practice
In English language teaching the first stage is less exotic than it sounds, because the profession already has a shared vocabulary for outcomes. The Council of Europe built the CEFR around can do descriptors precisely so that a level could be stated as performance rather than as a list of structures, and those descriptors make a serviceable stage one when a course has no other stated goal. A stage one written as learners can take and pass on a phone message containing a name, a date and a reason is usable. A stage one written as the present perfect is not, because it names content rather than a result.
Stage two is where the method earns its keep, and it is the stage that separates backward design from ordinary good intentions. The question is not what test exists but what evidence would satisfy you. If the outcome is a phone message, the evidence is a phone message taken under something like real conditions, not a gap fill on reported speech. The British Council’s guidance on planning lessons and courses makes the same alignment point in simpler terms: any activities you plan to do should directly link to that aim or learning outcome, and each stage builds on the stage before it until the learners eventually have all the skills and language they need to achieve the learning outcome.
Stage three then becomes a shopping list rather than a creative act. You work back from the task and ask what it demands. That is usually a short, dull and surprisingly specific list, and it tends to be shorter than the one a coursebook unit would have given you.
A worked classroom example
A six week evening course for eleven adults at B1 in a language school in Seville, two hours a week, all of them working in customer facing roles in a call centre that has just added an English speaking line. The coursebook unit on the shelf covered travel and holidays.
| Stage | What was decided | Why |
|---|---|---|
| 1. Desired result | By week six each learner can take a complaint call in English, establish what went wrong, state what will happen next and end the call politely. | It came out of a needs analysis in week one, not from the syllabus. Three learners said the part they dreaded was not vocabulary but not knowing how to close a call. |
| 2. Evidence | A recorded five minute role play in week six with the teacher as a difficult caller, marked against four criteria agreed with the group in week two. Plus two short recordings in weeks three and five, and a one page written follow up email. | The criteria were written before any lesson was planned. Agreeing them with the group in week two turned the assessment into a target rather than a surprise. |
| 3. Learning plan | Polite interruption and clarification language, the softened future for commitments, complaint and apology lexis, numbers and spelling under pressure, telephone listening, and a weekly controlled practice slot. | Every item on the list is there because the task needs it. Travel and holidays came off the plan, and so did two grammar points that would otherwise have been taught out of habit. |
The instructive part is what the sequence removed. Planning forwards from the coursebook would have produced six competent lessons on travel, and the group would have finished the course no readier for the call they were afraid of. The removal is the method working, not a shortcut around it.
One caution from the same course: the week six recording was the only piece of evidence that carried real weight, and that is a known risk of the framework. A single culminating task is fragile, because one learner with a cold or a bad night loses the whole record. The two interim recordings existed to guard against that, and they are the part of the design most often left out.
Commonly confused terms
| Backward design is not | The difference |
|---|---|
| Forward design | Forward design starts from content or a syllabus and moves through method to outcomes. Backward design starts from outcomes. Jack Richards sets the two out, with central design as a third option that starts from classroom methodology, in his 2013 account of curriculum approaches in language teaching. |
| A needs analysis | A needs analysis gathers information about learners before planning starts. Backward design is what you do with it. The needs analysis feeds stage one; it does not replace it, because a list of needs is not yet a statement of results. |
| Teaching to the test | Teaching to the test starts from an existing exam and works back from its format. Backward design starts from what you actually want learners to be able to do and then designs evidence for it. If the exam happens to measure that, the two coincide. Often it does not, which is the subject of washback. |
| PPP or any lesson shape | PPP, task based learning and the rest describe the internal shape of a lesson. Backward design describes the order in which you plan, and it is neutral about what the lessons then look like. A backward designed course can be taught through PPP throughout. |
| Writing a lesson aim | Almost every lesson plan has an aim written at the top. Backward design is the harder second step of writing the evidence before the activities, and it is that step, not the aim, that changes what gets taught. |
| Constructive alignment | John Biggs’s constructive alignment, from 1996 and in higher education, makes a very similar argument about matching objectives, assessment and teaching. The two are cousins rather than the same thing: backward design is a planning sequence, constructive alignment is a criterion you can apply to a course that already exists. |
Where it comes from, and what it does not claim
The idea that planning should start from ends rather than means is older than the name. Ralph Tyler set it out in 1949 in a four question rationale for curriculum planning that begins with the purposes a school seeks to attain. Wiggins and McTighe named it, gave it a template and made it usable at unit level in Understanding by Design, first published by ASCD in 1998 and revised in 2005.
Two parts of their framework are worth knowing and are often dropped in summaries. The first is the six facets of understanding, which is what stage two is supposed to assess against: the capacity to explain, interpret, apply, shift perspective, empathise and self assess. The authors are explicit that not all six are used every time, and that performance tasks built on them are culminating tasks for a unit rather than something a daily lesson carries. The second is that they do not recommend applying the full three stage process to an isolated lesson. The unit, not the lesson, is the design object, because an essential question is meant to be revisited over time rather than answered before the bell.
The honest caveat is about evidence. Backward design is very widely adopted and has a substantial practitioner literature behind it, including a peer reviewed account in CBE Life Sciences Education describing how graduate and secondary science teams used it to rebuild lesson plans under a National Science Foundation project. What it does not have is a body of controlled trials showing that courses planned this way outperform courses planned another way. It is best understood as a discipline for planning rather than as an intervention with a measured effect size, and a page that tells you otherwise is overselling it.
The second caveat is practical. The template can become the work. A three stage plan written for every unit of a general English course, complete with essential questions and transfer goals, is a large amount of paper for a teacher with four classes and no planning time. The part that survives contact with a real timetable is smaller: name the result, write the evidence, then plan. Everything else in the framework is optional scaffolding around those three moves.
References
Of the sources below, four were checked and are linked. The remaining five are books or journal articles that are not freely available online, so they are cited here without a link rather than left out.
| Source | What it supports |
|---|---|
| McTighe, J. and Wiggins, G. (2012) Understanding by Design Framework. ASCD white paper. Read it | The three stages, the seven tenets, the six facets of understanding, and the two failures the framework guards against. |
| Jones, K.A., Vermette, P.J. and Jones, J.L., in CBE Life Sciences Education (2007) Understanding by Design: A Framework for Effecting Curricular Development and Assessment. Read it | A peer reviewed account of the three stages in use, and the WHERE prompt for designing stage three activities. |
| British Council, Planning lessons and courses. Read it | The alignment principle in ELT terms: activities link to the learning outcome, and each stage builds on the one before it. |
| Council of Europe, The CEFR Levels. Read it | The can do descriptors that give a language course a ready made vocabulary for stage one. |
| Wiggins, G. and McTighe, J. (2005) Understanding by Design, 2nd edition. Alexandria, VA: ASCD. | The original full statement of the framework and its template. |
| Tyler, R.W. (1949) Basic Principles of Curriculum and Instruction. Chicago: University of Chicago Press. | The earlier ends first rationale that backward design descends from. |
| Richards, J.C. (2013) Curriculum approaches in language teaching: forward, central and backward design. RELC Journal, 44(1). | The forward, central and backward distinction as it applies specifically to language curriculum. |
| Biggs, J. (1996) Enhancing teaching through constructive alignment. Higher Education, 32(3). | Constructive alignment, the closest neighbouring idea. |
| Hutchinson, T. and Waters, A. (1987) English for Specific Purposes. Cambridge: Cambridge University Press. | Target needs and learning needs, the distinction that feeds stage one. |
In the classroom
- Write stage one as something a learner does, not as something you cover. If the sentence contains a grammar name, it is probably content in disguise.
- Write the marking criteria before the first lesson, and show them to the group. A criterion a learner has seen is a target; one they have not is a trap.
- Build two small pieces of interim evidence as well as the final task. One culminating performance is a fragile record of a whole course.
- Expect the plan to shorten. If backward design has not removed anything from your syllabus, you have most likely planned forwards and written the outcome afterwards.
Frequently asked questions
What is backward design in simple terms?
It is planning in reverse. You decide what learners should be able to do at the end, then decide what would count as proof they can do it, and only then plan the lessons that get them there. The name refers to the order of planning, not to anything about the teaching itself.
What are the three stages of backward design?
Identify desired results, determine acceptable evidence, and plan learning experiences and instruction. The three come from Wiggins and McTighe’s Understanding by Design framework, and the requirement that ties them together is alignment: what stage one names must be what stage two assesses and what stage three teaches.
Is backward design the same as Understanding by Design?
Not quite. Understanding by Design, often shortened to UbD, is the wider framework Wiggins and McTighe built, including the template, the six facets of understanding and the design standards. Backward design is the planning sequence at its centre. In everyday staffroom use the two names are treated as interchangeable, and little is lost by that.
How does backward design work in a language classroom?
Stage one is usually written as a can do statement of the kind the CEFR uses, because language courses already describe ability that way. Stage two is a performance task that resembles the real situation, such as a phone call, an interview or a written enquiry. Stage three is then the lexis, structures and skills work that the task turns out to require, which is often less than a coursebook unit would have covered.
Is backward design the same as teaching to the test?
No, although they can look similar from outside. Teaching to the test takes an existing exam and works back from its format. Backward design starts from what you want learners to be able to do in the world and then designs evidence for that. The two coincide only when the exam happens to measure the thing you actually wanted.
What are the criticisms of backward design?
Two are worth taking seriously. Its evidence base is adoption and practitioner report rather than controlled comparison, so it is a planning discipline rather than a measured intervention. And the full template is heavy for a teacher with several classes, which in practice means the paperwork gets abandoned before the idea does. Keeping the three moves and dropping the forms is a reasonable response to both.
How to cite this definition
tefl.ai (2026) Backward Design. TEFL Teaching Methods and Classroom Techniques Glossary. Available at: https://tefl.ai/backward-design/ (Accessed: 28 September 2026).
Related terms
Backward design sits in the assessment and course design family of the glossary, alongside needs analysis, which supplies stage one, and washback, which describes what happens when the evidence chosen in stage two starts to reshape the teaching. The full set of twenty terms is in the TEFL teaching methods and classroom techniques glossary.
For the qualification and career terminology that sits outside this glossary, see TEFL qualifications and terminology explained from The TEFL Institute, and its step by step guide to building a TEFL curriculum, which covers the wider mapping work this term sits inside. For teachers planning an accredited module inside an Irish school timetable, The TEFL Institute of Ireland’s guide to Transition Year TEFL programmes is a worked case of a course written around a stated outcome.
Last updated: 28 September 2026.
