What the Bain SOVA test is
SOVA is an assessment platform that combines aptitude, judgement and personality questions in one candidate experience. Bain's careers page says its digital assessment runs on SOVA or TestGorilla, that the email after you apply tells you which, that it takes about 30 to 40 minutes, and that you can use pen, paper and a calculator. Bain also says the SOVA assessment can be completed across more than one sitting, which TestGorilla's can't.
Prep sites agree that SOVA is the platform reported for London and the UK, other EMEA offices and the Middle East. Bain doesn't publish an office list, so rely on your email. If it says TestGorilla, read the Bain TestGorilla guide instead.
The sections
Five prep sites that have collected Bain candidate reports agree on this core list. Not every candidate gets every section; the mix varies.
| Section | What you do | How settled |
|---|---|---|
| Numerical reasoning | Read tables and charts; percentages, ratios, growth, conversions | Agreed across sources |
| Verbal reasoning | Read a short passage; mark statements true, false or cannot say | Agreed across sources |
| Logical reasoning | Pattern series, matrices, odd one out; some reports add deductive questions | Core agreed; deductive part disputed |
| Situational judgement | Choose or rank responses to workplace scenarios | Agreed it exists; answer format disputed |
| Work-style questionnaire | Rate how much statements describe you | Reported in some versions |
Question counts aren't settled. Some sources say around 15 per section, others 25 to 30 across the whole test. One prep site says the test adapts its difficulty to your answers; we found no second source for that, so don't plan around it.
Timing: no clock on screen, but speed counts
This is the part that trips people up. Four prep sites and several candidate posts say SOVA shows no countdown per question, yet records how long you take, and that time feeds into your result. SOVA hasn't confirmed this publicly, but it's reported consistently enough to act on.
In practice that means two things. Don't treat the missing timer as permission to spend three minutes on a question. And don't rush so hard you misread. A steady pace of about a minute per numerical question and 30 to 45 seconds per verbal statement is a reasonable target, and our SOVA-style practice records your time per question so you can see where you drift.
Numerical reasoning: worked example
Expect a table or chart and a question that needs one or two steps. Currency conversion and percentage change come up often in reports.
Example: comparing prices across currencies
A client sells the same software licence in three markets. Exchange rates: $1 = €0.90 and $1 = £0.80.
| Market | Local price |
|---|---|
| United States | $120 |
| Germany | €99 |
| United Kingdom | £92 |
In which market is the licence cheapest in US dollars?
Germany: €99 ÷ 0.90 = $110. United Kingdom: £92 ÷ 0.80 = $115. United States: $120. Answer: Germany.
The trap is the direction of conversion. The rate tells you how many euros one dollar buys, so to go from euros to dollars you divide. Multiplying gives $89 for Germany and $74 for the UK, and makes the UK look cheapest. A quick sense check catches it: a euro is worth more than 90 cents, so €99 must be more than $99, not less.
Example: a share of a total
A retailer's sales by channel last year, in £ thousands: stores 2,400, website 1,100, marketplace 500. This year the website grew 40% and the other two channels were flat. What share of total sales came from the website this year, to the nearest percent?
Website this year: 1,100 × 1.4 = 1,540. Total this year: 2,400 + 1,540 + 500 = 4,440. Share: 1,540 ÷ 4,440 = 34.7%. Answer: 35%.
Two slips to watch for. Using last year's total (4,000) as the denominator gives 38.5%. And adding 40 percentage points to last year's share (27.5%) gives 67.5%, which fails the sense check immediately: the website can't be two thirds of sales when stores alone are over half.
Verbal reasoning: worked example
You get a passage of a hundred or so words and a statement. You decide whether the statement is true, false, or cannot say, based only on the passage.
Example: a passage and three statements
"The city's bike-share scheme launched in 2022 with 400 bikes. Usage grew every quarter in its first year, and the council added 200 bikes in 2023. A survey of members found that most use the scheme to commute. The operator has said it will not expand further until it finds a sponsor."
Statement 1: The council added bikes to the scheme in 2023.
The passage says so directly. Answer: True.
Statement 2: Usage fell in at least one quarter of the scheme's first year.
The passage says usage grew every quarter in the first year, which contradicts this. Answer: False.
Statement 3: Most bike-share trips are commutes.
The survey says most members use it to commute. That's about people, not trips; a minority of members who ride often for leisure could make up most of the trips. The passage neither confirms nor contradicts the statement. Answer: Cannot say.
The rule for this section: answer from the passage alone, even when your general knowledge says otherwise. "False" means the passage contradicts the statement. "Cannot say" means it neither confirms nor contradicts it.
Logical reasoning: worked example
Mostly inductive: a series of shapes or symbols follows a rule, and you choose what comes next or which item breaks the pattern. The shapes are hard to show in text, so here is the same reasoning with letters and numbers.
Example: two rules at once
Series: A2, C4, F8, J16, ?
Letters: A to C is +2, C to F is +3, F to J is +4, so next is +5: J to O. Numbers double each step: 16 × 2 = 32. Answer: O32.
Shape series work the same way, with two or three things changing at once: the number of shapes, their position, their rotation, whether they're shaded. Check one property at a time in a fixed order (count, shape, position, rotation, shading) and write down the rule for each before looking at the options.
Example: odd one out
Which doesn't belong: 16, 25, 36, 48, 64?
16, 25, 36 and 64 are square numbers. 48 isn't. Answer: 48.
Example: a deductive question
Some reports describe deductive items alongside the pattern series. "No junior analysts attend the steering committee. Priya attends the steering committee." Which follows?
If no junior analyst attends, anyone who attends isn't a junior analyst. Answer: Priya is not a junior analyst. Nothing follows about whether she's a manager, a partner or a client; the statements don't say. Deductive items reward stopping exactly where the premises stop.
Situational judgement: worked example
You read a short workplace scenario and respond. Reports disagree on the format: some describe picking the most and least effective response, some ranking all options, some rating each one. The thinking is the same whichever you get.
Example: a client asks for more
You're two weeks into a four-week project. At the end of a meeting, the client's head of operations asks if your team could "also take a quick look" at their supplier contracts. It isn't in the agreed scope, and your team is already stretched.
- Agree on the spot to keep the client happy.
- Say you'll check with your manager and come back to them, then do so the same day.
- Tell the client it's out of scope and can't be done.
- Ask a junior team member to look at the contracts in their spare time without mentioning it to your manager.
Most effective: B. It takes the request seriously, doesn't commit the team to work it can't absorb, and puts the decision with the person who owns the scope. Least effective: D. It hides the work, overloads a junior colleague and leaves the client with an informal promise nobody is managing.
A and C are the middle options. A is well meant but commits resources you don't control. C protects the team but closes a door that the firm might want open. Reasonable people could rank those two either way, which is normal for this kind of question.
What these questions tend to reward: taking ownership, being honest with clients and colleagues, involving the right person, and fixing the cause rather than the symptom.
The work-style questionnaire
Some SOVA versions include statements you rate for how well they describe you. There's no right answer to prepare. Answer honestly and quickly; questionnaires like this usually include similar statements in different words, and trying to present an idealised version of yourself tends to produce inconsistent answers, which is worse than an honest profile.
How SOVA differs from TestGorilla
| SOVA | TestGorilla | |
|---|---|---|
| Structure | One assessment with sections | Three or four separate tests |
| Clock | No visible per-question timer reported; time is still recorded | A visible timer on each test |
| Sittings | More than one allowed (Bain) | One sitting |
| Verbal passages | Yes, true / false / cannot say | Not in the reported Bain mixes |
| Calculator | Allowed (Bain) | Allowed (Bain) |
Common mistakes
- Treating "no timer" as "no time pressure". Speed is reported to count.
- Converting currencies the wrong way. Sense-check which currency is worth more before calculating.
- Using outside knowledge on verbal questions. Only the passage counts.
- Confusing "false" and "cannot say". False needs a contradiction in the passage.
- Guessing pattern rules from one pair. Confirm a rule across the whole series.
- Performing a personality. Answer the questionnaire as you are.
- Splitting sittings badly. If you use more than one sitting, stop between sections, not in the middle of one, and come back the same day while you're warmed up.
A one-week plan
- Day 1: confirm SOVA in your email. Do one short set of each section, untimed, to find your weakest.
- Day 2: numerical: currency conversion, percentage change, ratios, reading charts. Aim for about a minute a question.
- Day 3: verbal: twenty true, false or cannot say statements, then read every explanation, especially the ones you got wrong.
- Day 4: logical: series and odd one out. Practise the fixed checking order.
- Day 5: situational judgement: read the explanations and learn the reasoning, not the answers.
- Day 6: a full SOVA-style mock at a steady pace, and look at your time per question afterwards.
- Day 7: rest, check your setup, sit it fresh.
Bain practice on SolveForge has numerical, logical, verbal and situational judgement sections, with a SOVA-style full mock for pass holders. The questions are ours, written in the reported style, not SOVA's.
FAQ
What sections are in the Bain SOVA test?
Prep sites agree on numerical, verbal and logical reasoning and a situational judgement section, with a work-style questionnaire in some versions. Not every candidate gets every section.
Is the Bain SOVA test timed?
Prep sites consistently report no visible per-question timer, but say your response time is recorded and counts. SOVA hasn't confirmed this, so work at a steady pace rather than treating it as untimed.
How long does the Bain SOVA test take?
Bain's careers page says about 30 to 40 minutes for its digital assessment. Prep-site estimates for SOVA run wider, from about 30 to 75 minutes, depending on the sections included.
Can I do the Bain SOVA test in more than one sitting?
Yes. Bain says the SOVA assessment can be completed across more than one sitting. If you split it, stop between sections.
Can I use a calculator on the Bain SOVA test?
Yes. Bain's careers page says pen, paper and a calculator are allowed, and SOVA's candidate guidance also allows a calculator.
What does "cannot say" mean in the verbal section?
That the passage neither confirms nor contradicts the statement. Choose "false" only when the passage says something that directly conflicts with it.
Keep going
Back to the Bain online assessment overview. The other Bain platform: the Bain TestGorilla guide. More worked examples: numerical, verbal, logical and situational judgement. Applying elsewhere too: the MBB online assessments compared.
Sources we checked: Bain's digital assessment careers page, SOVA's candidate preparation hub, and prep-site reports from IGotAnOffer, CaseBasix, MConsultingPrep, MyConsultingCoach, MBB Gorilla and PrepLounge candidate posts. Where they disagree we've said so above.
SolveForge is an independent practice tool and is not affiliated with, endorsed by, or connected to Bain & Company or SOVA. Every example on this page was written by us; none is a real test item. "Bain" and "SOVA" are trademarks of their owners. Practising here does not guarantee any assessment result.