A final-week plan
Lesson 4 of 4 in Walk In Ready.
The final week is not for learning services. If a service is new to you seven days out, it will still be new on the day, and the hours spent on it come out of the hours that would have converted knowledge you already have into marks.
What the final week is for is conversion: turning four domains of knowledge into reliable reading, reliable discrimination between the pairs that compete, and reliable pacing. Those are three distinct skills, they are drilled differently, and the plan below is built around them.
One prerequisite: you need a cold, timed, full-length mock early in the week, because everything else is allocated from its result. A mock taken at the end of the week measures you and teaches you nothing you have time to act on.
The four levels of readiness — each one assumes the one below it
- Recognition — You can say each service’s one job
For every service you meet, you can finish the sentence "this is the one that…" in a single clause. Not its feature list — its job, and which station on the pipeline spine it stands at.
This is the floor, and it is cheap to repair: a service you cannot place in one clause is worth ten minutes, not an afternoon. It is also where the lightest domain pays best, because breadth at shallow depth is exactly what recognition is.
- Discrimination — You can name the separator for every competing pair
For each pair that genuinely overlaps, you can state the one constraint that decides it — not a comparison of features, a separator. "Does the requirement need the Kafka API?" "Do you need the cluster itself?" "Does the query reach into the warehouse?" "Is Airflow part of the requirement?"
This is the level most readers are weakest at, and it is also the level that most directly earns marks, because the two-both-fit item is the dominant shape on this exam. Drill it with the pairs table, not with documentation.
- Constraint reading — You find the deciding clause before the options
You read the question sentence first, name the constraint family out loud, and only then look at the options — and you run the four pre-commit checks without thinking about them.
This is a habit rather than knowledge, which means it is built by repetition under mild time pressure, not by reading. Twenty questions done slowly and deliberately in this style are worth more than eighty done at speed.
- Pacing under load — Two minutes per item, for 130 minutes, without losing accuracy
You hold the checkpoints — question 20 at minute 40, question 40 at minute 80, all 65 touched by minute 110 — and your accuracy on the last twenty items matches your accuracy on the first twenty.
That last clause is the one people fail. Pacing is not only about finishing; it is about finishing without the quality collapse that comes from realising at question 50 that you are behind. The only way to measure it is a full-length timed run, which is why the mock sits at the start of the week.
Seven days, allocated from one mock result
- Day 7 — full mock, cold and timed
All 65 items, 130 minutes, no notes, no pausing. Cold means before any revision, because the point is a measurement you can allocate from. Record the time you reached question 20 and question 40 as well as the score — pacing debt is the thing a score alone hides.
- Read the result by domain, not by total
Convert each domain to "marks available" — 34, 26, 22 and 18 per cent of the scored items — and multiply by how wrong you were. The product is where the week goes. A 60% score in the heaviest domain is worth more attention than a 40% score in the lightest.
- Comfortably above the pass line?
- Days 6–4 — close the two weakest domains
Two domains, not four, and by reviewing the modules you already did rather than by finding new material. Redo their module quizzes and read the explanation on every item you get right as well as every item you get wrong — a right answer for the wrong reason is a mark you will lose on a reworded stem.
- Days 6–4 — sharpen discrimination instead
If no domain is weak, your remaining marks are in the two-both-fit items. Spend the three days on the competing pairs and the reflex counter-cases, which is where a comfortable pass becomes a safe one.
- Day 3 — the pairs and the reflex table
One session, out loud, no notes: for each competing pair, state the separator and one clause that would flip it. Then walk the reflex table and say the counter-case for each row. This is the highest-value hour of the week and it is pure recall practice.
- Day 2 — a timed half-paper
Thirty-two items in 64 minutes, under the same conditions as the mock. You are testing the habit, not the knowledge: did you read the question sentence first, did you name the constraint, did you answer before flagging, did you hold the clock?
- Day 1 — nothing new
Re-read the exam guide end to end — it is short — and re-read your own notes on the two domains you closed. No new services, no new practice sets, no late-night documentation. Confirm the booking, the identification requirements and the start time, and sleep.
- The morning — three sentences
"Question sentence first." "Name the constraint." "Answer before flagging." That is the whole strategy, and it fits in the walk from the car park.
| Day | What you do | What it measures |
|---|---|---|
7 | The full mock: 65 items, 130 minutes, cold and uninterrupted. Record the clock at items 20 and 40 as well as the score. | Everything at once — and specifically whether your accuracy holds in the last twenty items. This is the measurement the rest of the week is allocated from. |
6–4 | Close the two domains where weight × wrongness is largest, by rereading the modules you already did and redoing their quizzes. If nothing is weak, spend these days on the competing pairs instead. | Knowledge, and only where it pays. Reading the explanation on items you got right is the part people skip and the part that finds right-for-the-wrong-reason answers. |
3 | Out loud, no notes: every competing pair and its separator, then every reflex and its counter-case. | Discrimination — the level that most directly decides two-both-fit items, and the one that decays fastest without recall practice. |
2 | A timed half-paper of 32 items in 64 minutes, under exam conditions. | The habit, not the knowledge: question sentence first, constraint named, answer committed before flagging, clock held. |
1 | Re-read the exam guide end to end and your own notes on the two domains you closed. Confirm booking, identification and start time. Nothing new. | Nothing — deliberately. The marginal value of a new service on day one is negative, because it costs sleep and confidence. |
0 | Three sentences: question sentence first, name the constraint, answer before flagging. | Recall of the strategy under mild stress, which is the only thing you can still improve on the morning. |
One last thing, and it is the sentence this whole module exists to deliver.
You will meet questions where two options both work. That is not a sign you have under-prepared — it is the exam working as designed, and it is the moment your preparation either pays or does not. When two survive, do not pick the better design. Find the clause that separates them. It is in the stem, it is usually a number with a unit or a named technology or a phrase about what must not change, and it is the difference between a coin flip and a mark.
Good luck. Read the last sentence first.
Key terms: Pacing (two minutes per question), DEA-C01, Deciding constraint, Keyword reflex, Compensatory scoring
Interactive checkpoint quiz (2 questions) — open this page in a browser to take it.
Tool: DEA-C01 mock exam — The DEA-C01 Exam Hall: a blueprint-weighted 65-question mock in 130 minutes — exactly two minutes each, which is all the real exam gives you — with a scaled-score estimate and the four-domain breakdown the real score report shows. Also a 20-question pace sprint and per-domain drills. The real exam scores only 50 of its 65 questions and never tells you which, so this mock scores all 65 and you should treat every question as live.