The honest answer: one platform as the base, specialists added only with a specific exam and date attached. A pure stack costs more than people notice, in money and in fragmented performance data; a pure single subscription occasionally leaves genuine capability on the table. The right question is what your base layer should be and what earns a slot on top of it.
Three sample stacks, costed
UK medical student, UKMLA year. Stack version: Geeky Medics Everything Bundle (£69.99 per year) for CPSA stations, plus a written-exam bank, plus ChatGPT for explanation. Consolidated version: iatroX (£99 per year) covering the AKT bank, adaptive revision, Tutor, planner and mocks, with Geeky Medics added for the clinical exam window. Either way you land near £170 per year; the difference is whether your written-exam analytics live somewhere that persists after graduation.
USMLE candidate. Here the specialist case is strongest: UWorld's bank is the standard, AMBOSS or Osmosis the library layer, Neural Consult ($224.99 per year) if your school's lectures drive the internal exams. That stack runs to several hundred pounds equivalent, and for Steps specifically it is worth it. iatroX serves this candidate as a lower-cost additional adaptive bank rather than the centre.
Working UK doctor sitting a Royal College exam. Stack version: a per-exam question bank subscription each sitting, a separate clinical reference, perhaps a CPD logger. Consolidated version: iatroX's £99 covers the MRCP, MRCGP, MRCEM and SCE banks under one roof, plus askiatroX for daily clinical questions and My CPD for reflective logging. For serial postgraduate exams, per-exam pricing is where stacks quietly become expensive.
The costs nobody itemises
Switching friction: every additional app is another login, another notification stream, another place your streak can die. Duplicated features: three subscriptions each bundling flashcards means paying three times for one capability. And fragmented analytics, the underrated one: if your cardiology performance is split across three products, no single system can see your weakness clearly enough to schedule against it. Adaptive learning is only as good as the data it sits on, and stacks shred the data.
Where specialists remain genuinely superior
Honesty requires the list. Geeky Medics for OSCEs and virtual patients: no consolidated platform matches it. UWorld for USMLE question quality. Neural Consult for turning your own lectures into a study environment. If one of those is your actual bottleneck, buy it without guilt.
The base-layer principle
The sensible architecture is a base layer that persists across exams and into practice, carrying your question bank, scheduling, tutoring and clinical reference in one performance model, with specialists rented for the seasons that need them. iatroX is built to be that base layer, which is a different claim from pretending it replaces every niche tool. Consolidate the spine; rent the limbs.
The audit: five questions before you subscribe to anything
First, what is my next assessment, and which single tool is closest to its format? That tool anchors the budget; everything else justifies itself relative to it.
Second, what does each candidate subscription duplicate? List the core objects, bank, flashcards, tutor, planner, reference, against what you already hold. Paying twice for flashcards is the classic stack leak, and it hides inside bundles.
Third, where will my performance data live? Adaptive scheduling is only as good as the history it can see. If the answer is "split across three dashboards", accept that you have traded away most of what adaptivity is worth, and make the trade knowingly.
Fourth, what is the exit plan per product? Specialists should have an end date attached, the week after the OSCE, the month after Steps. Anything without a natural exit is claiming to be your base layer, and should be judged by base-layer standards: does it persist across exams and into practice?
Fifth, what does the next three years cost, not the next month? Serial exams are where per-exam pricing compounds quietly. A doctor facing Part 1, then an SCE, then possibly an MSRA detour prices those sittings once under a subscription like iatroX's £99 annual, against a fresh purchase each time on the stack model; a US-only student faces the opposite maths, since UWorld's slot is non-negotiable regardless.
Run the audit annually. The honest outcome most years is boring: keep the spine, cancel one limb, trial one new thing free. Boring is what value looks like in this market.
Is there a case for deliberately over-buying?
One honest version exists: the month before a high-stakes exam, redundancy can beat elegance, and running a second bank purely for fresh calibrated volume is a defensible splurge. The dishonest version is subscribing to soothe anxiety, which buys notifications rather than marks. The test is whether the extra product has a job the spine cannot do and a cancellation date attached; if it has both, spend without guilt, and if it has neither, the money is tuition paid to your nerves.