What's known about the video
Prep sites agree that the case ends with a recorded video recommendation of about one minute, in one take. They disagree on preparation time: some describe about 60 seconds, others a final section of about five minutes. Plan for 60 seconds; if you get five minutes, you'll have time to spare.
BCG doesn't publish how the video is assessed, who watches it, or any pass mark. Prep sites suggest clarity and a recommendation that follows from the case numbers matter most, which is plausible and unconfirmed. BCG says AI is never used to make hiring decisions, but not how Casey responses are reviewed.
A structure that fits a minute
Four parts, in this order, adding up to about 130 to 150 words: a minute at a normal speaking pace.
| Part | What to say | Words |
|---|---|---|
| The answer | What the client should do, and when, in one sentence | 20 to 25 |
| Reasons | Two or three, each with a number from the case | 60 to 65 |
| Risk and mitigation | The one risk most likely to break the answer, and what to do about it | 35 to 40 |
| Next step | One concrete action the client can start soon | 15 to 20 |
The answer comes first even when it's "no" or "yes, with a condition". Round the figures so they're easy to say: "about $800,000", not "$797,400". Pick one risk and attach a response to it.
Time yourself reading 140 words aloud before the day. If you finish early, slow down rather than adding content. We haven't found a reliable report of what happens if you run over, so treat one minute as a hard limit.
Preparing notes during the case
With 60 seconds to prepare, you can't start from scratch at the end. Keep a box in the corner of your paper and update it as the numbers come in:
- Answer: your current view.
- Reasons: the two or three figures that most support it, with units.
- Risk and mitigation: often the assumption where a small change hurt most, plus one way to reduce it.
- Next step: the test that would settle the biggest open question.
With a longer window, write the first sentence in full and keep the rest as numbers and short phrases. A full script pulls your eyes off the lens.
Delivery on camera
- Look at the lens, not at your own face on screen, which reads as looking away.
- Light in front of you. A window or lamp behind the screen, not behind you, and the camera at eye level.
- One take means keep going. If you stumble, correct it in a few words ("sorry, $1.9 million") and carry on. Restarting wastes seconds.
Two example recommendations
Both cases are invented for this page.
Example 1: a subscription launch (weak and strong)
Tidewell Pet Co., with 60 pet supply stores, is deciding whether to launch a monthly pet food delivery subscription. You worked out 25,000 subscribers in year one at $38 a month, $9 contribution per monthly order, $1.9m a year in fixed costs and a $2.4m build cost. Revenue: 25,000 × $38 × 12 = $11.4m. Contribution: 25,000 × 12 × $9 = $2.7m, so profit is $2.7m − $1.9m = $0.8m. Payback: $2.4m ÷ $0.8m = 3 years. Break-even: $1.9m ÷ ($9 × 12) = about 17,600 subscribers, roughly 30% below plan.
The weak version:
"Thanks. We looked at a lot of different factors in this case, like the market, the costs and the competition. The pet market is quite big and a lot of customers seem interested, so the subscription could be an interesting opportunity for Tidewell. There are some risks, for example competitors might react, costs could go up, and customers might cancel. Overall, I think it could make sense for Tidewell to consider launching, but it would probably need more analysis before deciding."
What's wrong: the answer comes last, hedged twice. No number from the case, three risks with no mitigation, "more analysis" as the next step, and at about 80 words it leaves a third of the minute unused.
The strong version, about 135 words:
"I recommend Tidewell launches the home delivery subscription next spring. We expect about 25,000 subscribers in year one, which at $38 a month is about $11.4 million in revenue. Each monthly order contributes $9, so the subscription covers its $1.9 million in fixed costs and makes about $800,000 in its first year. If that holds, the $2.4 million build cost pays back in about three years. The main risk is sign-ups. Break-even is about 17,600 subscribers, so if we land roughly 30% below plan, the subscription stops making money. To limit that, I'd use a fulfillment partner paid per order in year one instead of leasing our own warehouse, which keeps fixed costs low until demand is proven. As a next step, I'd run a four-week pre-sale in ten stores to test sign-ups at $38."
Why it works: answer first, a figure in each reason, a risk with a number and a response, and a next step that could start on Monday.
Example 2: an acquisition (a "no" with a counteroffer)
Calder Hotels owns 14 hotels in one region and has been offered Pinemoor Inns, six nearby hotels, for $48 million. Pinemoor makes $3.2 million a year in operating profit, combining booking and purchasing would save about $0.8 million a year, and Calder's board wants acquisitions to pay back within eight years. Its occupancy fell from 74% to 66% over two years. Working: $3.2m + $0.8m = $4.0m a year; 48 ÷ 4 = 12 years to pay back; an eight-year payback supports 8 × $4.0m = $32m.
The recommendation, about 145 words:
"I recommend Calder does not buy Pinemoor at the asking price of $48 million, and offers about $32 million instead. With $0.8 million a year of savings, Pinemoor would earn about $4 million a year for us, so $48 million takes 12 years to pay back, against the board's limit of eight. The trend makes that worse: Pinemoor's occupancy has dropped 8 points in two years, from 74% to 66%, so its profit is more likely to fall than rise. At an eight-year payback, the price that works is about $32 million, a third lower. The risk in walking away is that a competitor buys Pinemoor and adds six hotels to our region. To manage that, I'd put the $32 million offer in now rather than stepping back. As a next step, I'd ask Pinemoor for its booking data to find out why occupancy fell."
A "no" is still an answer, and it's stronger with the price at which it becomes a "yes".
What to avoid
- Recapping the case. Twenty seconds of background leaves forty for the answer.
- Hedging the answer. Put conditions in the risk and next step, not the first sentence.
- New numbers. Quote figures from your notes. One you haven't calculated invites doubt.
- "More analysis" as the next step. Name the analysis, or better, the test.
FAQ
How long is the BCG Casey video recommendation?
Prep sites agree it's about one minute, recorded in one take.
How much preparation time do you get before the Casey video?
Sources disagree: some prep sites describe about 60 seconds, others a final section of about five minutes. Plan for the shorter version by building notes during the case.
Can I re-record the Casey video?
Prep sites describe a single take, so correct any stumble briefly and keep going.
How is the Casey video recommendation assessed?
BCG doesn't publish how the video is assessed or any pass mark. Prep sites suggest clarity and a link to the case numbers matter most; that's unconfirmed.
What should the Casey recommendation include?
Your answer in the first sentence, two or three reasons with numbers from the case, the main risk and how to reduce it, and one concrete next step, in about 130 to 150 words.
Keep going
The whole case, worked through: the BCG Casey guide. The numbers you'll be quoting: Casey math, worked examples. Back to the BCG online assessment overview. Dedicated BCG practice isn't built yet; that page explains what's planned.
Sources we checked: prep-site reports from Hacking the Case Interview, MConsultingPrep, Management Consulted (its HireQuotient page), Road to Offer and MyConsultingCoach, and BCG's page on responsible AI in hiring. We couldn't open BCG's own online case user guide.
SolveForge is an independent practice tool and is not affiliated with, endorsed by, or connected to Boston Consulting Group or HireQuotient. Both example cases on this page were written by us; neither is a BCG case or based on one. "BCG" and "Boston Consulting Group" are trademarks of their owner. Nothing here guarantees any assessment result.