How to Study for the Exam

Most people who fail the SRE Foundation (SREF) do not fail because the material was beyond them — this is an entry-level exam, and its own syllabus caps the difficulty on purpose. They fail because they studied the way they studied at school: reading front to back, highlighting, re-reading the highlights. That method is almost perfectly mismatched to what this exam measures. SREF is a knowledge exam, delivered by PeopleCert on behalf of the DevOps Institute: multiple choice, no terminal, no cluster, no dashboard. Nobody watches your hands. What gets scored is whether you can hold four plausible sentences side by side and pick the one that is true by the syllabus's own definitions — roughly one every ninety seconds, for an hour. This page is the method, not the material. The eight blueprint pages are the syllabus, the SREF study plan is the calendar, and this is how to work through them so any of it is still there when the clock starts.

What kind of exam this is — and why that changes everything

☺ Like you're 10: One kind of test watches your hands. This one reads your mind. You practice hands by doing; you practice knowing by explaining, and by arguing with wrong answers.

Everything on this page follows from a single distinction, so get it straight before you plan an hour. A performance-based exam grades the state you leave behind. Nobody cares how you got there — if the service ends up healthy, you scored, whether you typed it, scripted it, or copied it from a runbook. A knowledge-based exam grades which option you selected. Nobody cares whether you could have built the thing; they care whether you can tell one correct statement from three that are also plausible.

SREF is squarely the second kind, and its published format leaves no ambiguity about that. Both official sources — the PeopleCert product page and the DevOps Institute certification page — describe the same paper: 40 multiple-choice questions, 60 minutes, 65% to pass, web-based delivery, no formal prerequisites. There is no lab component and nothing to build. (Those figures were current when this page was written. Confirm them on the vendor pages before you book — exam formats get revised, and this one has already changed platforms once.)

This distinction matters more on this site than on most, because the Certifications page in this course also covers exams of the other kind, and the performance-based exams in the sister Platform Engineering course are graded entirely on cluster state. If you are working toward both, do not run one study regime for both. They reward opposite skills, and a regime tuned for one shortchanges the other.

The syllabus tells you how hard to study — literally

Most exams make you guess at depth. SREF does not. The official SREF Exam Study Guide states that the certification is built on the Bloom taxonomy and that the paper contains Bloom level 1 questions (knowledge of SRE concepts and vocabulary) and Bloom level 2 questions (comprehension of those concepts in context). Those are the only two levels it names.

Read that as an instruction, because it is one. Level 1 and level 2 mean the exam asks you to define and to recognize in context. It does not ask you to design an SLO regime for a described organization, weigh two architectures against each other, or critique a postmortem process — those are levels 3 and up, and they belong to higher-level credentials such as SRE Practitioner, and to job interviews, not here. So the study target is not "be a better site reliability engineer." It is: be able to state, precisely and quickly, what every term on the syllabus means, and recognize it when it is dressed up in a one-paragraph scenario.

The arithmetic that drives every tactic on this page

Three numbers, and everything downstream of them. All three are derived from the published format above, so re-derive them yourself if the vendor revises the paper.

That third row is worth sitting with, because it cuts both ways. Fourteen is a generous allowance by exam standards, and it is why a candidate who has honestly covered all eight topic areas at a definitional level rarely fails. It is also exactly why skipping an area is fatal: the heaviest area is worth seven questions on its own, so writing one off spends half your entire margin before you have answered anything.

"Open book" is not the safety net it sounds like

Both official sources list this exam as open book — and PeopleCert's product page qualifies that in a way most third-party summaries drop: the permitted material is the Official Training Materials, whether supplied directly by PeopleCert or through an Accredited Training Organization. Not your notes. Not this site. Not a search engine. If you self-study without buying those materials, you have nothing admissible to bring, and you are sitting a closed-book exam with an open-book label on it.

Even if you do have them, the arithmetic above closes the gap: ninety seconds a question is enough to check a term you nearly know, and nowhere near enough to learn one you do not. Chris McLeod, who sat and passed the exam on 9 December 2023 and wrote it up on his own blog the following week, makes precisely this point from the other side: open book only helps if you have the official training materials to hand, and he did not, so he answered from memory. His account describes the exam as broad rather than deep — which is the Bloom 1–2 ceiling showing up in practice, from a candidate who had never read that table.

Retrieval, not re-reading

☺ Like you're 10: Reading a page again feels like learning because the words look friendly. Shutting the page and saying it out loud feels awful — and that is the one that works.

Here is the single most important sentence on this page: the act of pulling something out of your head with the page shut is what moves it from "recognized" to "owned." Everything else here is a wrapper around that move.

Re-reading fails for a specific and slightly cruel reason. The second pass over a page is fluent — the sentences arrive smoothly, nothing surprises you — and your brain reads that fluency as evidence of knowing. It is not. It is evidence that the page is nearby. Remove the page, and the fluency goes with it, which is exactly the condition the exam creates. Retrieval feels worse and works better: it is effortful, you fumble, and the fumbling is the part that does the work.

For SREF, retrieval has two units, and you need both.

Unit one — the definition

A single sentence you can produce cold, in your own words, that would survive a pedant. "An SLI is the measurement itself — a metric of some aspect of the service, taken at the boundary the user actually crosses." Not "SLIs are about measuring things." The difference between those two sentences is roughly the difference between passing and failing, because the second one cannot reject a distractor and the first one can.

The discipline: pick a term from the glossary or from the official study guide's concept list, say your sentence out loud before you look, then compare. Where you were vague, you have found a real gap. Where you were confidently wrong, you have found a much more valuable one.

Unit two — the discrimination

A knowledge exam is not really testing whether you know the right answer. It is testing whether you can reject three wrong ones that were written to be attractive. That is a separate skill, and it is trained separately: given four options, say why each of the other three is wrong, in words, before you commit. Not "this one feels right" — an actual reason.

This is where the self-check and the SREF practice bank earn their place. Three rules make them training rather than entertainment:

• Before you pick, argue down the other three out loud. If you cannot say why an option is wrong, you do not yet know the concept it is testing — you are pattern-matching on familiar words.
• When you get one right, read the explanation anyway. Right answers arrived at for wrong reasons are invisible unless you check, and they are the ones that fail you on a differently-worded question.
• When you get one wrong, classify it before moving on. Did you not know the fact, did you confuse it with its neighbor, or did you misread the question? Those three need completely different fixes, and the mistake log below is where the classification lives.

How to tell "I recognize this" from "I can do this"

Recognition is the counterfeit currency of exam prep. It is produced by every pleasant study activity — reading, watching, listening, highlighting — and it is spent by exactly none of the exam. Four tests separate them, and all four are cheap:

Run the neighbor test most often. On an exam capped at Bloom 2, near-neighbor confusion is the dominant failure mode — not ignorance. There is a whole table of the pairs worth drilling further down this page.

Read the syllabus as a checklist, not as prose

☺ Like you're 10: The exam publishes a list of everything it is allowed to ask, and even says how many questions each part gets. Turn that into a to-do list with ticky boxes instead of reading it once like a story.

The single most wasted asset in SREF prep is the official study guide, because almost everyone reads it once, nods at the topic list, and goes off to watch videos. It is not an introduction. It is the specification of the question pool: nothing outside it can be asked, and everything inside it can.

The eight topic areas, and exactly what each is worth

Here is the part most third-party study advice gets wrong by omission. The DevOps Institute's certification web page publishes no percentage split — but the official Exam Study Guide PDF carries a table headed Exam Topic Areas and Question Weighting, with a maximum question count for each of the eight areas. That table is reproduced below, alongside where each area lives in this course.

Two observations, and both are ours, derived from that table rather than claims the vendor makes:

• The maxima sum to exactly 40. 4 + 6 + 5 + 7 + 6 + 4 + 4 + 4 = 40. A set of "maximums" that adds up to precisely the paper length has very little room to vary — every area has to sit at its maximum for the arithmetic to work. Corroborating this, the 40-question sample paper inside the same official document is laid out in exactly that pattern: its own answer key tags questions 1–4 to area 1, 5–10 to area 2, 11–15 to area 3, 16–22 to area 4, and so on down to 37–40 for area 8.
• Three areas carry 19 of the 40 marks. Monitoring & SLIs (7), SLOs & error budgets (6) and Tools & automation (6) are, between them, worth nearly half the paper — and just seven short of the 26 you need to pass.

What this does not license is writing off the four-question areas. Four questions is 10% of the paper each, and the four smallest areas together are 16 of 40 — two more than the 14 wrong answers you are allowed. Write all four off and you have failed before the paper starts, however solid the heavy three are. The weighting decides order and hours, not what you can skip. Study everything; study the heavy three first and revisit them last.

Turn each topic area into claims you can defend

Syllabus lines are written as nouns: "Reducing Toil." Nouns cannot be ticked off, because there is no moment at which you have finished a noun. Rewrite each one as a sentence you can either say or not say:

Notice the shape. Every rewrite is a sentence you could say out loud, and most carry a second clause that proves you did not just memorize the first. If your rewrite has neither, you have written a mood, not a study item.

Build a confidence ledger

Now make it queryable. A plain text file beats any app, because you want to sort and count it without leaving the terminal you already live in. One line per topic area, carrying the published question count so the file can do the hour arithmetic for you:

mkdir -p ~/sref-prep && cd ~/sref-prep

printf 'area|title|maxq|conf|last\n' > ledger.psv

cat >> ledger.psv <<'EOF'
SRE-1|SRE Principles and Practices|4|0|never
SRE-2|Service Level Objectives and Error Budgets|6|0|never
SRE-3|Reducing Toil|5|0|never
SRE-4|Monitoring and Service Level Indicators|7|0|never
SRE-5|SRE Tools and Automation|6|0|never
SRE-6|Anti-fragility and Learning from Failure|4|0|never
SRE-7|Organizational Impact of SRE|4|0|never
SRE-8|SRE, Other Frameworks, Trends|4|0|never
EOF

Score each line 0–3 on a scale that has nothing to do with feelings. 0 = never studied. 1 = read the blueprint page once. 2 = wrote the area's terms out from a blank page, with gaps. 3 = wrote them out cleanly and scored 80%+ on that area's practice questions, cold, on two different days. Only a 3 counts as done — anything looser lets "I read that on Tuesday" masquerade as knowledge, and that is exactly the item that collapses at question 31.

Then the file answers the only question that matters at the start of a session: what should I work on today?

cd ~/sref-prep

awk -F'|' 'NR>1 {printf "%-6s %-46s %sq  conf=%s  exposure=%s\n", $1, $2, $3, $4, $3*(3-$4)}' ledger.psv | sort -t= -k3 -rn

awk -F'|' 'NR>1 {t+=$3*(3-$4)} END {printf "total exposure: %d (0 = ready)\n", t}' ledger.psv

That exposure column is the whole point: questions at stake multiplied by how far you are from owning the area. SRE-4 at confidence 1 scores 14; SRE-8 at confidence 1 scores 8. Both need work, and the ledger has just told you which one to open first — without any appeal to which topic you happen to find interesting. Total exposure starts at 120 and should be walking toward zero. If it has not moved in a week, you have been reading, not retrieving.

Let the counts allocate your hours

This course budgets roughly 20–25 hours in total for SREF prep, which is proportionate to a 60-minute foundation paper — see the study plan for the calendar version. Split a 24-hour budget by published question count and it lands like this:

The syllabus has moved since the free material was written

☺ Like you're 10: The free study guide you can download is from a few years ago. The real exam got updated since. Most of it is the same — but some new topics were added, and you would never know from the free file.

This is the trap that catches careful, thorough candidates, so it gets its own section. The free study guides that everybody finds — and that this page points you at, because they are genuinely excellent — are v1.0 (2020) and v1.1 (2021). PeopleCert's official training-material update notices list the current SRE Foundation syllabus as v1.2, updated October 2024, and describe that update as adding insights on modern DevOps practices including generative AI, AIOps, platform engineering, value stream management, observability and progressive deployments, alongside refreshed references.

Two consequences, and neither is optional:

• Study the v1.1 guide for structure, not for scope. Its eight topic areas, its Bloom ceiling and its sample paper are all still the best free map of how the exam thinks. Its topic list is one revision behind.
• Its administration details are stale. The v1.1 format table says Supervised: No — a leftover from the pre-PeopleCert delivery platform. The exam is now delivered through PeopleCert, and the online-proctored route — the one you get if you book the exam yourself rather than sitting it at a training centre — involves a live human proctor, an ID check and a webcam room scan; McLeod's December 2023 account describes exactly that, including panning the camera over and under his desk. Do not plan your exam-day environment from a 2021 PDF.

The good news for anyone studying here: the v1.2 additions map onto pages this course already has, so covering them costs you an evening rather than a rewrite.

Three passes over one syllabus

☺ Like you're 10: Go over everything three times — fast, then slow, then from memory. Not once, slowly, forever.

The instinct is to start at topic area one and go deep until it is perfect, then move on. That produces a candidate who is superb on error budgets and has never read the word "anti-fragility," because time ran out. Three passes over the whole syllabus beats one perfect pass over a third of it — especially on a paper where every area is guaranteed to appear and a zero anywhere costs the same as a zero everywhere.

Pass 1 — survey (about 15% of your time, so 3–4 hours)

Skim all eight blueprint pages and the official study guide's topic table. Take no notes. The only artifact you produce is the ledger, scored honestly. This pass exists to kill surprises: by the end there should be nothing in the syllabus you have never heard of. It will feel unproductive. It is buying you the ability to plan, and it is the pass people skip and then regret.

Pass 2 — read and produce (about 60%, so 14–15 hours)

The long middle, worked in exposure order. For each area: read the blueprint page, close it, write the area's terms and definitions out from a blank page, then run that area's practice set — principles, SLOs & toil, monitoring & tools, or anti-fragility & organizational impact. Finish every session by updating the ledger. An area reaches confidence 2 when the blank page comes out mostly right; it reaches 3 later, in pass three, when it survives a timed paper.

Pass 3 — compress (about 25%, so 6 hours)

Nothing new is allowed in pass three. You are doing three things only: retrieving from a blank page, drilling the lines your ledger and mistake log agree on, and sitting timed papers from the five SREF mock exams. If you find yourself learning a brand-new concept in pass three, you have mis-scheduled — bank it as a known gap, accept the four-question risk, and protect the revision time instead.

Spacing — review just before you forget

☺ Like you're 10: Things you learn leak out of your head. Grab them again just as they are about to leak and they stick much longer — and you have to do it less and less often.

Memory for anything you are not using daily decays, and it decays fastest right after you learn it. Reviewing something just as it starts to fade resets the clock and flattens the curve, so each successive review can sit further out than the last. That is the whole idea: review at increasing intervals, timed to just-before-you-forget. It is not a productivity fad; it is the most evidence-backed study technique there is, and it costs nothing.

It matters more here than it would on a longer, deeper exam, because SREF prep is short. Twenty-odd hours spread over four to six weeks means most of what you learn in week one has to survive three or four weeks of not being touched. Without spacing, it will not.

A schedule you will actually keep

Ignore the elaborate algorithms. A fixed 1 / 3 / 7 / 21 ladder captures nearly all the benefit and needs no software: learn something, revisit it tomorrow, then three days later, then a week later, then three weeks later. Anything you fumble on a review drops back to the start of the ladder. Anything you nail at +21 is done — trust it and stop spending time on it.

The cheapest implementation is a dated file. Add a line when you learn something; each morning, grep for today:

cd ~/sref-prep

item='SRE-2: 99.9% monthly SLO -> ~43 min error budget, from memory'
for d in 1 3 7 21; do
  printf '%s|%s\n' "$(date -v+${d}d +%F)" "$item" >> reviews.psv
done

grep "^$(date +%F)|" reviews.psv || echo "nothing due today — go take a fresh area instead"

That date -v+1d form is BSD/macOS; on Linux it is date -d '+1 day' +%F. Either way, the point is that scheduling a review must cost you one line, or you will stop doing it by Thursday.

The three drills on this site, and the order to run them in

☺ Like you're 10: Cards to find out what you have forgotten, questions to practice picking the right one, and a checklist to prove you can actually apply it. Different jobs — do not use one and skip the others.

This course ships three general-purpose drills plus a whole SREF-specific shelf. They are not interchangeable, and most people misuse the first one by treating it as training rather than as a diagnostic.

Flashcards — a diagnostic you have to schedule yourself

The flashcards here are 38 cards drawn from every module, laid out as a searchable grid you flip by clicking. Be clear about what that is and is not: it is a deck, not a scheduling engine. Nothing tracks which cards you fumbled or brings them back tomorrow — that bookkeeping is yours, which is exactly why the reviews.psv file above exists. Search the deck by topic, work the cards for the area you studied today, and log the misses into the ladder.

Two rules make the deck work, and both are about honesty. Answer out loud before you flip. Flipping first and thinking "yes, I knew that" trains recognition, which is the currency this exam does not accept. Use the five-second rule. Anything you cannot start answering within five seconds is not known, however familiar it feels — at 90 seconds a question you cannot afford a term that takes twenty seconds to surface.

Self-check — where discrimination is actually trained

The self-check is the closest thing here to the exam's own mechanic: pick a domain, take ten questions cold, commit, and get the explanation immediately. Options are reshuffled every session, so you are recalling the answer rather than remembering that it was the third one — and anything you miss is remembered on your device, so ↻ Past misses always has your weak spots queued up. That misses queue is the most valuable feature on the page and the one people never open.

Run it with the three rules from the retrieval section: argue down the other three options before committing, read the explanation even when you were right, and classify every miss. Ten questions done that way is worth an hour of questions done quickly.

The checklist — the one drill that is not about the exam

The on-call readiness checklist is not an exam-readiness checklist, and it would be dishonest to sell it as one. It is a phase-by-phase pass over SLO and measurement readiness, on-call and incident readiness, reliability engineering readiness, and reporting — with progress saved in your browser.

Use it anyway, for a specific reason. An exam capped at Bloom 2 still asks you to recognize concepts in context, and the fastest way to make a definition concrete is to try to apply it to a real service. Walking the checklist against something you actually operate turns "error budget policy" from a phrase into a thing your team either has or does not have — and concepts you have argued about are dramatically more resistant to a well-written distractor than concepts you have only read.

The SREF shelf, in the order it becomes useful

Definitions are the exam — so drill the near neighbors

☺ Like you're 10: Most wrong answers on this test are not things you never heard of. They are two things you know that live right next to each other, and you grabbed the wrong one.

On a paper capped at Bloom 1–2, the question writer's main tool is the near neighbor: an option that is a true statement about a different concept, or a true statement about the right concept in the wrong role. You defeat those by drilling pairs rather than terms. Cover the right-hand column, produce the distinction out loud, then check.

The arithmetic has to be reflexive

One item type deserves separate drilling because it is the only one where being slow costs you as much as being wrong: error-budget arithmetic. The official sample paper's question 7 asks for the error budget of a 99.9% monthly SLO, and the keyed answer is 43 minutes. That is not a hard calculation — 0.1% of a 30-day month — but doing it from first principles under a clock, in a room with a proctor watching, is where people lose ninety seconds they needed elsewhere.

So do not calculate it on the day. Know it:

30-day window = 43,200 minutes

99.9%  SLO  ->  0.1%   ->  ~43 min  of allowed downtime per 30 days
99.95% SLO  ->  0.05%  ->  ~21.6 min
99.99% SLO  ->  0.01%  ->  ~4.3 min

Each extra "nine" divides the allowance by roughly ten.

Sol's arithmetic drills in SLIs, SLOs & error budgets are where to build that reflex, and Know It Cold is where the finished numbers live.

Your production experience is an asset — and a specific liability

☺ Like you're 10: Knowing the job really helps. But your team probably uses some words slightly differently from the exam, and the exam only marks its own version right.

This exam assumes a working baseline — Linux, the basics of distributed systems, some monitoring exposure, and enough on-call experience to know what a page feels like. If you have that, most of the syllabus will read as things you already do, and you will be tempted to skip straight to the mocks. Two reasons that backfires.

First, the exam tests the syllabus's definitions, not your industry's. The official sample paper contains a question asking for the definition of latency according to the Catchpoint survey the course cites, and the keyed answer is the narrow one — the delay incurred in communicating a message — not the broader "total time from request to response" that many teams use day to day. Both are defensible in the wild. Only one is marked correct. That is not a flaw in the exam; it is what a vendor-neutral foundation certification is: a shared vocabulary, defined by the syllabus that issues it.

Second, the syllabus contains named artifacts your workplace probably does not use. The SLO VALET model is a good example — the official sample paper asks what its "T" stands for, and the keyed answer is Tickets. If you have never met the model, no amount of production experience will get you there, and it is a one-line fact to learn.

The upside of experience is real, though, and worth using deliberately: attach every abstract term to something you have actually seen. Which service in your estate has a genuine SLO rather than a dashboard? What was the last piece of toil you automated, and which of the six properties made it toil? When did you last see an alert that was cause-based and should have been symptom-based? Concepts anchored to a memory survive distractors far better than concepts anchored to a paragraph.

Keep a mistake log

☺ Like you're 10: Write down every mistake in one place. After two weeks you will notice you keep making the same three — and those three are your whole revision list.

Individual mistakes are noise. Patterns of mistakes are the highest-value information you will produce during the entire plan, and you cannot see patterns without writing things down. One file, appended to the moment something goes wrong, never edited afterwards.

What an entry contains

Five fields, and the second-to-last is the one everybody skips and the one that does the work:

cat >> ~/sref-prep/mistakes.md <<'EOF'

## 2026-08-24 · SRE-4 Monitoring & SLIs · question 17
- what was asked: which signal is the SLI in the described scenario
- what I picked: the internal queue depth metric
- what was really wrong: I do not have "measured at the user-facing boundary"
  as part of my definition of an SLI — only "a metric about the service"
- class: term-collision
- redrill: 2026-08-27
EOF

"What was really wrong" is a different sentence from "what I got wrong," and forcing yourself to write it is what converts a mistake into a lesson. "I picked C" is not actionable. "My definition of SLI is missing a clause" tells you exactly what to fix and exactly which card to make.

Classify it — the fix depends entirely on the class

Once entries are classified, one command each Sunday tells you what kind of candidate you currently are:

grep -h '^- class:' ~/sref-prep/mistakes.md | sort | uniq -c | sort -rn

grep -h '^- redrill:' ~/sref-prep/mistakes.md | awk -v today="$(date +%F)" -F': ' '$2 <= today'

A log dominated by knowledge-gap means you are early — keep reading and producing. Dominated by term-collision means you know the material and need the pairs table, not another pass over the blueprint. Dominated by misread means you need a ritual, not more study — and that is a thirty-second fix you would never have found by studying harder. The prescription is completely different in each case, which is why the classification is not busywork.

The traps

☺ Like you're 10: Some study activities feel productive and teach nothing. And some of the exam facts you will find on the internet are simply wrong, including for a different exam with the same name.

Trap one — re-reading the study guide

The official study guide is good, free, and pleasant to read, which is exactly what makes it dangerous. Reading it a third time produces fluency and no retrieval. It has two legitimate roles: a source for pass one's survey, and a reference you open with a specific question already in your head, after which you close it. The discipline that keeps both honest: never open a document without a question. If you cannot say what you are looking for, you are browsing, and browsing is entertainment.

One specific piece of that guide is worth knowing about: its "Value Added Resources" section — a long list of talks, articles and videos — is explicitly labeled by DevOps Institute as not examinable. It is enrichment. It is also the single easiest way to lose six hours that belonged to retrieval practice. Read it after you pass.

Trap two — third-party specs that are simply wrong

Search for this exam and you will land, overwhelmingly, on training-partner pages, practice-test vendors and content farms. Several publish exam specs that contradict both official sources. Verified examples worth inoculating yourself against:

Trap three — banking on the open-book label

This earns its own entry because of how badly it changes behavior. A candidate who believes they can look things up studies for recognition and stops at "I would know it if I saw it." Ninety seconds a question converts that into a fail. Study as though there is nothing in the room but you.

Trap four — dumps

Exam-dump sites dominate the search results for this exam more thoroughly than for almost any other on this site. They are worthless and they are a risk: memorized answers do not transfer to a re-worded question pool, and certification bodies generally treat their use as malpractice — read the candidate agreement you accept when you book rather than taking anyone's word for the consequences. There is a legitimate alternative, and those dump sites bury it — the official study guide contains a full 40-question sample paper with an answer key, written by the people who write the exam, laid out in the real topic distribution. That is the closest thing to a real form that exists in public, it is free, and it is allowed.

Trap five — expecting a community to study with

Worth saying plainly, because it surprises people. We went looking for first-hand accounts of sitting this exam and found one — McLeod's, cited above. There is no rich forum record here of the kind you find around the hands-on cloud-native exams, and the search results that look like experience reports are mostly vendor marketing. (We could not survey Reddit directly, so treat that as "we found none," not "none exists.") Practically: do not wait to find a study group or a curated video course. The official study guide, the eight blueprint pages here, and the practice bank are the material. That is not a gap in your research — it is the actual state of the public record.

Build a weekly cadence

☺ Like you're 10: Small amounts, most days, always at the same time. That beats one giant panic weekend, every time.

Study plans fail on scheduling, not on content. Three and a half focused hours spread across a week beats seven hours on a Sunday, because spacing is doing half the work and because a long block is where fatigue quietly turns study into transcription. Build the week around one session length you can actually protect — 45 minutes is the right unit for this exam.

The shape of a study week

That is about three and a half hours a week — roughly six weeks to the 20–25 hours this course budgets, or four weeks if you take the near-daily version in the study plan. The Sunday twenty minutes is the slot people cut and the one that makes the plan self-correcting; without it you will keep studying whichever area you enjoyed most last week.

A 45-minute session template

When you miss a week

You will. The rule is: do not restart, and do not try to make it up. Come back with a single 30-minute retrieval session on whatever the ledger says has the highest exposure, take the score hit honestly, and resume the normal week. Candidates who try to repay lost hours in one weekend usually abandon the plan within a fortnight. A study plan you resume is worth ten study plans you abandon.

How to know you are actually ready

☺ Like you're 10: "I feel ready" is not evidence. "I scored 34 out of 40 on a paper I had never seen, twice, on different days" is evidence. Book on evidence.

Confidence is not a readiness signal — it is a mood, and it tracks how recently you read something rather than whether you can produce it. This exam has an objective bar, so measure yourself against an objective test.

Signals that mean nothing

When the objective signals are green, book it. Waiting for the feeling is how people spend six months "almost ready" on an exam whose whole syllabus is twenty-odd hours of work. The registration walkthrough is on The SREF Exam.

The last 72 hours

☺ Like you're 10: The three days before the test are for polishing what you already have, not for learning anything new. And sleep counts as studying.

Whatever you do not know 72 hours out, you will not know on the day — so stop trying. The final stretch has one job: arrive rested, fast and calm, with the logistics rehearsed so none of your attention is spent on them.

One exam, one method. Now point it at something: open the eight topic areas next to a blank file, build the ledger, and let exposure pick tomorrow's session. When the objective signals turn green — two unseen papers clear, no area dragging, the mistake log gone quiet — work through the registration steps on The SREF Exam, confirm every number on the vendor's own page, and book it. Not before.