What are the top 10 concrete steps to create a differentiated lesson plan for mixed-ability classrooms in 2027

The outcome you should expect

A differentiated plan is not ten different lessons. That misunderstanding is the single biggest reason teachers abandon differentiation after a term: they picture themselves writing three parallel plans every night, do the math on their prep time, and quietly go back to teaching the middle of the room. What you should actually expect from a well-built differentiated lesson is one lesson with a single objective and multiple engineered routes to it — usually one core task, two or three variants of that task, and a bank of scaffolds and extensions you can hand out in the moment.

Concretely, here is what the ten steps produce when the plan is done and sitting on your desk. First, a one-line learning objective written in student-facing language, identical for every learner in the room. Second, a pre-assessment result — a short entry ticket, a diagnostic quiz, a set of last week's exit tickets — that sorts the class into rough readiness bands, usually three. Third, a core task everyone attempts. Fourth, a tiered version of that task: a scaffolded route (more structure, worked examples, sentence frames, partially completed steps), the core route, and a stretch route (fewer supports, more abstraction, an added constraint). Fifth, a grouping plan that says which students start where and when the groups get reshuffled. Sixth, two or three formative checkpoints written into the timeline with the specific question you will ask at each. Seventh, a reteach plan for the 15–20% who will miss the checkpoint. Eighth, an assessment that lets students show mastery in more than one format without changing the standard being measured.

On paper this is typically two pages. That is a realistic target. If your differentiated plan is running eight pages, you have built something you will not sustain past October.

The behavioural outcome in the room is more specific than "engagement goes up." What you should see, within three to four weeks of running this consistently: fewer students sitting idle in the first five minutes of independent work, because the entry point is calibrated; fewer hands up for the same procedural question, because the scaffolds answer it in writing; a shorter tail of students who complete nothing; and a shorter head of students who finish in nine minutes and then decorate the margins. The middle of the class is largely unaffected — they were already being served. Differentiation is mostly a tail-management strategy, and it should be judged as one.

The outcome you should *not* expect is uniform outcomes. Differentiated instruction narrows the gap in access, not necessarily in results. A student three years below grade level in reading does not arrive at grade level because you gave them a graphic organiser. What the organiser does is let them participate in the same intellectual work as their peers today, while a separate intervention track — reading specialist, tiered support, whatever your school runs — does the longer remediation job. Conflating the two is how differentiation gets blamed for failing to do something it was never designed to do.

One more expectation to set honestly: the first three or four differentiated plans you write will take roughly double the prep time of a standard lesson. Teachers who track this typically report the overhead dropping sharply once they have built a reusable scaffold bank — sentence frames, worked examples, graphic organisers, extension prompts — because after that point differentiating a new lesson is largely assembly rather than authorship. Budget for the front-loaded cost and build the bank deliberately from lesson one.

What drives that outcome

Four mechanisms do nearly all the work, and understanding them tells you which of the ten steps you can afford to skimp on and which you cannot.

Mechanism one: accurate readiness data. Differentiation without pre-assessment is guessing with extra steps. The step that most determines whether the whole plan lands is step one — a short, targeted pre-assessment. Not a full diagnostic exam. Three to five questions, five minutes, aimed precisely at the prerequisite skills the coming lesson depends on. If tomorrow's lesson requires students to multiply fractions, your pre-assessment asks about equivalent fractions and multiplication of whole numbers, not about the new content. The output is a rough sort, and rough is fine — you are choosing a starting point, not issuing a verdict.

Mechanism two: task tiering against a fixed objective. The tiers vary the *support and complexity*, never the standard. This is the distinction between differentiation and tracking, and it is the one that gets violated most often. If your scaffolded group is doing vocabulary matching while the stretch group analyses theme, you have not differentiated a lesson — you have run two lessons and quietly lowered expectations for a third of the room. Tiering done properly means all three groups are analysing theme; one group does it with the passage annotated and a sentence frame ("The author suggests ___ by showing ___"), one does it clean, and one does it across two texts.

Mechanism three: flexible grouping. Groups must reshuffle. Readiness on fractions tells you nothing about readiness on data handling, and a group formed in September that still has the same six names in it in February has become a track. Practical rule: regroup at least at every new unit, and mix the basis — sometimes readiness, sometimes interest, sometimes learning-profile, sometimes deliberately mixed-ability for a collaborative task where the stronger student's explaining is itself the learning. Students notice group stability faster than teachers think they do, and the moment they perceive a permanent "bottom group," the motivational cost starts eating your gains.

Mechanism four: formative feedback loops inside the lesson. The checkpoint is what converts a differentiated plan from a static seating chart into a responsive one. Mid-lesson, you check, and students move. That movement — a student who was on the scaffolded route stepping up to the core task at minute 25 — is the actual proof the system works.

The loop closing back on itself is the point. A differentiated plan that runs once and never feeds its results into the next grouping decision is a one-off performance, not a system. The data review at the end is what makes step one of the next lesson cheap — you already know the bands.

There is an adjacent mechanism worth naming because it explains why some classrooms differentiate more easily than others: classroom routines. Differentiation depends on students being able to start work, retrieve a scaffold, seek help from a peer, and switch tasks without teacher direction. In a room where every transition requires the teacher's voice, running three routes simultaneously is genuinely impossible. This is why differentiation frequently fails in classrooms with unstable behaviour management, and why the honest sequencing advice is: establish routines first, differentiate second. Teachers new to a challenging group are often better served spending six weeks on transitions, materials access, and independent-start norms than on tiering tasks.

Benchmarks and realistic ranges

Numbers here should be read as planning heuristics from common classroom practice, not research constants. They exist to stop you designing something that cannot fit in the period.

Number of tiers: two to three. Three is the practical ceiling for a single teacher without support staff. Four is possible with a co-teacher or teaching assistant. Two — core plus scaffold, or core plus stretch — is entirely legitimate and often the right choice for a lesson where the prerequisite spread is narrow. Teachers who try five tiers spend the lesson doing logistics.

Pre-assessment length: 3–5 items, under 6 minutes. If it takes longer, it eats the lesson it was meant to serve. Run it the day before where possible, so you plan on real data rather than sorting on the fly.

Group size: 3–5 students. Below three, a group is a conversation that one student can dominate; above five, someone disappears. For a class of 28, that means roughly six to eight working groups, which is about the limit of what one teacher can circulate through meaningfully in a 50-minute period.

Teacher circulation: 4–6 minutes per group. In a 50-minute lesson with 15 minutes of instruction and 10 of wrap-up, you have about 25 minutes of circulation. That buys you five focused check-ins, or three focused check-ins plus one pulled small-group reteach. Choose deliberately; you cannot do both fully.

Formative checkpoints: 2–3 per lesson. One at roughly 40% through the independent work — early enough that a student on the wrong route can still be moved. One near the end as an exit ticket. A third mid-way if the task is long or the class is volatile.

Scaffold bank size: aim for 15–20 reusable assets by end of term one. Sentence frames per text type, a worked-example template per problem type, three or four graphic organisers, a vocabulary support format, and a standard extension prompt structure ("now do it with an added constraint / for a different audience / and explain why the alternative fails"). This bank is the actual productivity asset. Everything else is per-lesson work; this compounds.

Prep time: 1.5–2× a standard lesson initially, trending toward 1.1–1.3× once the bank exists. If you are still at 2× after a term, you are authoring rather than assembling — audit what you keep rebuilding and turn it into a template.

Proportion of lessons differentiated: not 100%. A realistic and sustainable target is heavy differentiation on the two or three lessons per unit that carry the most conceptual load, lighter differentiation (scaffold available, extension available, no formal tiering) on the rest. Whole-class direct instruction is not a failure state; for introducing genuinely new content, it is often the most efficient choice, and differentiation lives in the practice phase that follows.

Reteach group: expect 15–25% of the class at any given checkpoint. If it is consistently above a third, the core task is pitched too high, or the prerequisite gap is bigger than a lesson-level fix can handle and you are looking at an intervention question rather than a differentiation question.

Adjacent benchmark worth borrowing: in workplace training and onboarding design, the same tail-management logic applies — cohorts are pre-assessed, content is tiered, and the measured outcome is time-to-competency rather than uniform test scores. The transferable insight is that the metric should be spread, not average. Track the gap between your 25th-percentile and 75th-percentile outcomes over a unit. If differentiation is working, the bottom quartile's completion and accuracy rates climb while the top quartile's stay flat or rise — the average may barely move, which is why average-only reporting makes good differentiation look like nothing happened.

Risks, edge cases, and failure modes

Failure mode one: tiering becomes tracking. Covered above, but it deserves the top slot because it is the most damaging and the most common. Symptoms: the same students in the "support" group all year; different learning objectives per group; the scaffolded group never producing the same artefact as the core group. Fix: audit your group rosters monthly, and check that every route ends in the same demonstrated skill.

Failure mode two: the scaffold never comes off. A graphic organiser is a bridge, not a permanent installation. If a student has had the sentence frame for eleven weeks, the frame has become the skill. Build a fading plan: full frame → partial frame (first clause only) → frame available on request → no frame. Name the fade to the student explicitly, because unannounced scaffold removal reads as punishment.

Failure mode three: choice without structure. Offering students a menu of six formats sounds like differentiation and often produces paralysis plus a race to whichever option looks least demanding. Constrain choice: two or three options, all requiring comparable cognitive work, all producing assessable evidence against the same criteria. "Write an analysis, record a two-minute explanation, or build an annotated diagram" works because all three demand the same reasoning. "Write an essay or make a poster" does not.

Failure mode four: over-differentiating the top. Extension tasks that are simply *more* — ten more problems, a longer essay — are a punishment for finishing. Extension should go up in abstraction, not out in volume: add a constraint, remove a scaffold, ask for a counter-argument, ask them to write the worked example a struggling peer would need.

Failure mode five: the assessment un-differentiates everything. You run three beautifully tiered routes for four weeks, then assess with a single timed written test that advantages exactly one profile. If the standard is "can analyse causation," and the test measures "can write quickly under time pressure," you have measured the wrong construct. Vary format where format is not the skill being assessed; hold it constant where it is.

Edge cases. *Very large classes (35+):* drop to two tiers and lean hard on peer structures and self-checking materials — answer keys, self-marking digital tasks — because your circulation budget per student collapses. *Classes with wide language spread:* language scaffolding and readiness scaffolding are different axes and must be planned separately; a fluent thinker with limited English needs vocabulary and phrasing support on a full-complexity task, not a simplified task. *Practical and studio subjects:* differentiate by degree of independence and by the point at which you intervene, not by task complexity — everyone makes the same thing, some with a jig and a demonstration, some from the brief alone. *Students with formal support plans:* the plan's provisions are a floor and a legal obligation, not one of your three tiers; build differentiation around them.

The sustainability risk. The failure that ends most differentiation programmes is not pedagogical, it is workload. A teacher running five classes cannot build three routes for twenty-five lessons a week. Departments that sustain this share planning: one teacher builds the tiered version of a lesson, the whole team uses it. If your school has no mechanism for that sharing, the realistic honest advice is to differentiate fewer lessons more deeply rather than every lesson shallowly.

A practical rollout plan

Here are the ten concrete steps in the order you actually perform them, with what "done" looks like for each.

Step 1 — Pre-assess the prerequisites. Three to five items on the skills the lesson depends on, run the day before. Done when you have a sortable result, not a graded one.

Step 2 — Write one shared learning objective. Student-facing, single sentence, identical for everyone. Done when you could write it on the board and no group needs a different version.

Step 3 — Design the core task first. Build the middle route before the variants. Done when the core task alone would be a defensible lesson.

Step 4 — Tier it up and down. Produce the scaffolded and stretch variants by adding/removing support and complexity — never by changing the objective. Done when all three variants produce comparable evidence.

Step 5 — Build or pull the scaffolds. Sentence frames, worked examples, organisers, vocabulary supports. Done when a student could start the task using only the written scaffold, without asking you.

Step 6 — Set the grouping and the reshuffle trigger. Who starts where, and the specific condition under which they move. Done when the movement rule is written down, not held in your head.

Step 7 — Insert formative checkpoints into the timeline. Two or three, with the exact question and the exact minute. Done when the checkpoint has a decision attached: what you do with each possible answer.

Step 8 — Plan the reteach. A small-group worked example ready for the 15–25% who miss checkpoint one. Done when the reteach materials exist before the lesson, not after it fails.

Step 9 — Differentiate the assessment format, not the standard. Two or three formats, one rubric. Done when the rubric is format-agnostic.

Step 10 — Review the data and regroup. Exit tickets sorted the same evening or the next morning; groups adjusted before the next lesson. Done when step 1 of the next lesson is already answered.

Sequencing across a term. Weeks 1–2: routines and materials access only; no tiering. Weeks 3–4: add a scaffold and an extension to your existing lessons — two tiers, no formal grouping. Weeks 5–8: introduce pre-assessment and three tiers on the highest-load lesson of each unit. Weeks 9+: add the formative checkpoint discipline and the data review. Attempting all ten steps in week one is the standard way to abandon this by half-term.

What to share with colleagues. The tiered task variants and the scaffold bank travel well between teachers; the grouping decisions do not, because they are class-specific. Departments that split the work this way — shared materials, individual grouping — get most of the benefit at a fraction of the individual cost.

Related questions

How is differentiation different from personalised learning?

Differentiation keeps one objective and varies the route within a whole-class lesson, teacher-directed. Personalised learning varies pace and often objective per student, usually with technology managing the pathway. Differentiation is a planning practice; personalisation is closer to a system design.

Do I need three tiers for every lesson?

No. Two tiers is often correct, and some lessons — introducing genuinely new content — are best delivered whole-class with scaffolds available but no formal tiering. Reserve three-tier planning for the two or three highest-load lessons per unit.

How do I stop students noticing which group is "the low group"?

Reshuffle groups on varied bases — readiness, interest, mixed-ability — at least each unit, keep the objective identical across routes, and never label groups by level. Students infer status from stability and from differing expectations, not from seating.

What if my pre-assessment shows the whole class lacks the prerequisite?

Then it is not a differentiation problem. Reteach the prerequisite to everyone before the new content. Differentiation manages a spread; it cannot manufacture a foundation that no one in the room has.

How do I differentiate when I have 35 students and no assistant?

Drop to two tiers, invest heavily in self-checking materials and answer keys, use structured peer routines, and accept fewer teacher check-ins per student. Depth of differentiation scales down with class size — plan for that rather than against it.

FAQ

Does differentiated instruction actually raise attainment? The evidence base is mixed and depends heavily on implementation quality — much of what is labelled differentiation in practice is unstructured group work, which performs poorly. The components with the strongest support are the ones inside the ten steps: formative assessment, explicit scaffolding, worked examples, and flexible grouping. Judge your own implementation on those components rather than on the label.

How long should it take to write a differentiated lesson plan? Initially 1.5–2× a standard plan. Once you have a reusable scaffold bank of roughly 15–20 assets, expect it to settle near 1.1–1.3×, because most of the work becomes selecting and adapting rather than authoring from scratch. If the overhead is not dropping after a term, you are rebuilding things you should be templating.

Should learning styles drive the differentiation? No. The idea that matching instruction to a student's preferred "style" improves outcomes has not held up under testing. Differentiate on readiness, on interest, and on the format in which a student can demonstrate a skill — those are defensible. Assigning a student permanently to "visual" tasks is not.

What is the minimum viable version of this? One scaffold and one extension attached to your existing core task, plus one mid-lesson checkpoint. That is three additions, it takes about fifteen minutes to prepare, and it captures a large share of the benefit. Add pre-assessment and formal grouping later once the routine is stable.

How do I handle a student who is years behind the rest of the class? Differentiate for access so they participate in today's lesson — scaffolds, reduced reading load, partially completed steps — while a separate intervention track addresses the underlying gap. Lesson-level differentiation is not a remediation programme and should not be asked to function as one.

Does technology make this easier? It helps with two specific jobs: self-marking practice that frees your circulation time, and quick pre-assessment with instant sorting. It does not help with the hard parts — deciding what the tiers should be, judging what a student's work reveals, and choosing when to fade a scaffold. Treat it as logistics support, not as the plan.

Sources

• https://www.edutopia.org/differentiated-instruction
• https://educationendowmentfoundation.org.uk/education-evidence/teaching-learning-toolkit
• https://ies.ed.gov/ncee/wwc/
• https://www.ascd.org/
• https://www.readingrockets.org/classroom/classroom-strategies
• https://www.gov.uk/government/publications/send-code-of-practice-0-to-25
• https://www.cast.org/impact/universal-design-for-learning-udl
• https://www.oecd.org/education/

Related on PULSE

• [How to design a formative assessment loop that actually changes instruction](/knowledge.html)
• [Scaffolding vs. simplifying: where the line sits in classroom task design](/knowledge.html)
• [Flexible grouping strategies for large classes](/knowledge.html)
• [Building a reusable scaffold bank in one term](/knowledge.html)
• [Differentiated assessment: varying format without lowering the standard](/knowledge.html)