Most people open a Bain SOVA practice page hunting for a predicted pass mark. That is the wrong object. A short numerical-plus-logical sampler is useful because it names the failure mechanism that will cost you points under soft time pressure, not because it clones Bain's full battery.
Bain's own digital-assessment FAQ says the test takes roughly 30–40 minutes, that your invitation decides Sova versus TestGorilla, and that the score is one input beside the rest of the application. London Associate Consultant recruiting points candidates at numerical reasoning, logical reasoning, and a personality questionnaire on Sova's Candidate Preparation Hub. Treat anything beyond that invitation as folklore until your email says otherwise.
After this piece you should be able to: (1) confirm which platform you actually face, (2) run a two-item diagnostic that classifies misses by mechanism, and (3) spend the next week fixing that mechanism instead of grinding random "SOVA packs."
The invitation is the syllabus
Bain states plainly that after you apply you get an email telling you whether to complete Sova or TestGorilla. Requirements vary by role and office. If the assessment is required and you miss the deadline, Bain says the application will not be considered.
That single line kills the most common prep error I see: candidates rehearse a Sova-style abstract pattern set for three evenings, then open a TestGorilla link with business-judgment and leadership modules. Or the reverse. The simulator in this article is for the Sova-style numerical and logical skills Bain London explicitly flags. If your email says TestGorilla, stop here and switch tools.
| Signal in your invite | What to practice | What not to waste time on |
|---|---|---|
| Sova (often EMEA; London AC names numerical, logical, personality) | Data exhibits, abstract patterns, honest personality responses | Memorizing a "Bain personality key" |
| TestGorilla | The modules named in that invite | Forcing Sova pattern drills onto a different battery |
| Unclear / office mix | Reply to the Bain recruiter email before you grind | Reddit format debates |
Sova Assessment's public materials describe a modular platform: employers pick from numerical, logical, verbal, situational judgment, personality, and other item types. Bain configures the mix. Your friend in another office is not your syllabus.
What a useful simulator should (and should not) claim
A useful Bain SOVA practice sampler does three things and admits three limits.
It should: present original numerical and logical items in the same cognitive family as Sova's published sample types; force you to choose under mild time pressure; and show worked reasoning that names the distractor logic, not just the correct letter.
It should not: claim to reproduce Bain's current confidential item bank; predict a Bain cutoff; or pretend two free questions equal a full multi-section battery. Bain does not publish section weights or pass rates. Anyone selling "the Bain SOVA score" is inventing a product Bain never promised.
CoachNed's live sampler at /bain-sova is built on that honesty: one numerical exhibit item and one abstract logical item, with review after each answer. Use it as a diagnostic loop, then leave.
Worked numerical example: pick the row before you punch the keys
Invented company: Northline Transit, a regional rail operator. Exhibit (values in millions of passengers):
| Corridor | 2023 | 2024 |
|---|---|---|
| Coastal | 12.4 | 13.1 |
| Inland | 8.6 | 7.9 |
| Express | 4.1 | 5.0 |
Question: What was the percentage change in Inland ridership from 2023 to 2024?
Options: A) −8.1% B) −8.9% C) +8.1% D) −7.9%
Work it once with me.
Required operation in words: Inland 2024 minus Inland 2023, divided by Inland 2023, times 100.
Inputs: 7.9 and 8.6. Change = 7.9 − 8.6 = −0.7. Ratio = −0.7 / 8.6 ≈ −0.0814 → about −8.1%.
Answer: A.
Now the traps that actually kill candidates in a screening setting:
| Trap | How it shows up here | Prevention |
|---|---|---|
| Wrong row | Using Coastal (12.4 → 13.1) and getting a positive growth distractor | Point to the corridor name before any arithmetic |
| Wrong denominator | −0.7 / 7.9 ≈ −8.9% (option B) | Write "change / start year" before entering numbers |
| Sign flip | Reporting +8.1% because the absolute change looks "about eight" | Keep the sign with the numerator |
| Unit theater | Treating 7.9 as absolute passengers instead of millions | Carry the unit in the margin even when it cancels |
Bain allows pen, paper, and a calculator on the Sova digital assessment (not in later interviews). The calculator does not save you from selecting Inland versus Coastal. When I score case math in live interviews, the same pattern shows up: people who can multiply accurately still lose the point when they grab the wrong line of the exhibit.
Reproduce the Northline item cold tomorrow. If you miss it, classify the miss with the table above before you open another question bank.
Worked logical example: one property at a time
Sequence of four frames (described, not drawn):
- Square, top-left black square in a 2×2 grid, other three white.
- Square, top-right black.
- Square, bottom-right black.
- Square, bottom-left black.
What comes fifth?
A candidate who invents a story about "rotating clockwise while the outer border thickens" will burn ninety seconds. The durable rule is simpler: the black cell walks clockwise around the four positions of the grid. Frame 5 returns to top-left black.
Test properties in a fixed order: count of objects, position, rotation of a single shape, shading or fill, then interactions. Confirm one property before you combine two. Eliminate any option that breaks a confirmed rule, then decide among what remains.
Ned's rule. On abstract items, if your first explanation needs more than two moving parts, you are probably wrong. Force yourself to state a one-sentence rule out loud; if you cannot, discard the story and re-check count and position before anything else.
That heuristic comes from watching candidates overfit. In a partner interview the parallel is structure: a four-branch MECE tree that you cannot defend loses to a two-branch tree that drives the math.
How to read a practice result (the part most guides skip)
After the sampler, do not write "numerical: weak." Write the mechanism.
| Miss type | Diagnostic question | Next session |
|---|---|---|
| Data selection | Did I use the named row, period, and segment? | Ten exhibits; point to inputs before calculating |
| Operation | Can I state the relationship in words first? | Write the formula line for every item |
| Unit / scale | Does the option match thousands vs units? | Carry units through every line |
| Pattern rule | Which single property changes consistently? | Thirty pattern items; one-property checklist |
| Pacing | Did one item eat the set? | Hard 60-second decision point, then move |
One wrong percentage does not prove weak numerical reasoning. Three denominator errors in a row does. Track categories across five sessions; that log is worth more than another unreviewed timed set.
A seven-day plan that respects Bain's own guidance
Bain tells Sova candidates to use Sova's Candidate Preparation Hub sample questions, find a quiet setup, and note that you need not finish Sova in one sitting so long as every section is done before the deadline. London AC recruiting again names numerical, logical, and personality. Build the week around that, not around folklore about secret cutoffs.
| Day | Session | Goal |
|---|---|---|
| 1 | Official Sova samples + CoachNed /bain-sova sampler | Baseline mechanisms, not a vanity score |
| 2–3 | Numerical only (exhibits like Northline) | Kill wrong-row and wrong-denominator habits |
| 4 | Logical only with the one-property checklist | Kill multi-rule storytelling |
| 5 | Mixed timed block (20–30 minutes) | Switch formats without losing method |
| 6 | Personality: answer honestly and consistently; no "Bain key" | Match Bain's stated point that they are not hunting one personality type |
| 7 | Technical dry run (Chrome, stable connection, calculator, paper) | Fail the setup at home, not on the invite deadline |
If you have more than a week, add timed arithmetic on /drills/quick-math rather than hunting longer fake batteries. Speed under soft timing is the scarce skill; screenshot memorization is not.
Personality and situational items without the folklore
Bain London says people at Bain have a diverse range of personality traits and thinking styles, and that the firm is not looking for one single type of person. That is the opposite of "learn the Bainie answer key" advice that floats around forums.
For personality questionnaires: read each statement, answer honestly, stay consistent across similar items. Gaming the inventory for what you think a consultant should say often produces internal contradictions that are more damaging than an imperfect but coherent profile.
For situational judgment, if your battery includes it: prefer responses that are proportionate, collaborative, and evidence-based. Escalate when the scenario actually requires it; do not pick the dramatic option for theater. The CoachNed sampler does not include these sections, so it makes no claim about them. Practice them on Sova's own hub if your invite lists them.
Practice this today
Open /bain-sova, answer the numerical item and the logical item, and write two lines: the mechanism of each miss (or confirm you got both). Then do ten Northline-style exhibit questions on paper with the "point to the row first" rule, and ten pattern items with Ned's one-sentence rule. Everything on CoachNed is open for seven days with no card; after that it is $120 for a recruiting season or $49 a month if you want the broader case and drill library.
If you also sit McKinsey or BCG screens this cycle, keep the same diagnostic mindset on those simulators rather than shopping for a universal "MBB test."
CoachNed is independent and is not affiliated with Bain & Company or Sova Assessment.
Frequently asked questions
Is there a free Bain SOVA practice test?
Yes. CoachNed hosts an original Sova-style sampler at /bain-sova with one numerical and one logical item plus worked review. It is a diagnostic, not a Bain score prediction or a full multi-section clone.
How long does the Bain digital assessment take?
Bain's careers FAQ states the assessment takes approximately 30–40 minutes. Your invitation email sets the completion deadline. For Sova, Bain notes you do not have to finish in one sitting as long as every section is done before that deadline.
What sections are in the Bain SOVA assessment?
It depends on the office and the invite. Bain London Associate Consultant recruiting points candidates to numerical reasoning, logical reasoning, and a personality questionnaire on Sova's hub. Other offices may use TestGorilla instead. Trust the email you receive after applying.
Can I use a calculator on Bain SOVA?
Bain says pen, paper, and a calculator are allowed for the Sova digital assessment. Calculators are for the digital assessment only, not for subsequent interviews. AI tools are prohibited during the assessment.
Is the digital assessment enough to get an interview?
Bain describes the assessment as one part of the review process alongside the rest of your application. Treat it as a mandatory screen when required, then put the bulk of remaining prep into case interviews.
Sources
- Digital assessment | Bain & Company — official time estimate (~30–40 minutes), Sova vs TestGorilla invite rule, calculator and AI policies, Sova may be completed in sections. Checked 2026-09-24.
- Bain London University Hiring — London AC assessment names numerical aptitude, logical reasoning, and behavioural preferences; points to Sova Candidate Preparation Hub. Checked 2026-09-24.
- Candidate Preparation Hub | Sova Assessment — practice-test types (numerical, logical, SJT, personality, verbal) and candidate setup guidance. Checked 2026-09-24.
- Assessments | Sova Assessment — modular platform description (employers configure cognitive and personality components). Checked 2026-09-24.