Issue Tree Example - Step-By-Step Approach with Takeaways
ฝัง
- เผยแพร่เมื่อ 19 มิ.ย. 2024
- If you've found this video helpful, you're gonna LOVE our free course at www.craftingcases.com/freecourse -- click the link for more info.
**
Check out our guide to Issue Trees at: www.craftingcases.com/issue-t...
**
Issue Trees are one of the most important techniques to create structures and frameworks for case interviews.
Consultants from McKinsey, BCG, Bain and other consulting firms use this tool as a "map" to solve both simple and comples problems. For this reason, it's one of the most widely used problem solving tools.
Yet, few people from outside of these firms can do this.
In this video I'll show you an example of an Issue Tree created to solve a fairly tough problem. I'd say about 95% of candidates who get interviews with top consulting firms wouldn't be able to structure this problem.
I'll also walk you through my thinking process as I build the tree and share my key takeaways at the end of the video.
**
If you've liked this video, make sure you click the Like button and leave a comment below! (This lets TH-cam know this video's good and will make it show it to more people).
If you want to hear more from us, click Subscribe (and click the Alarm Bell button if you want to get notified whenever we release a new video).
If you have any questions, ask it in the comments below - I'll read and answer each one of them and may even make a video about it, who knows!
And if you know someone (or someones) who might benefit from our videos, make sure you share it using the Share button next to the Like button.
Hope the best for your case interview preparation and I'll see you on the next video :)
Bruno - แนวปฏิบัติและการใช้ชีวิต
![ทัศนศึกษาโรงงานผลิตโค้ก [THAI/ENG SUB]](http://i.ytimg.com/vi/M9y-TJBNem0/mqdefault.jpg)
Want to improve your structuring skills? Join our free course! You'll learn our best techniques to create structures and frameworks from scratch, brainstorm ideas in a MECE way, and generate and test hypotheses. Get instant access at: www.craftingcases.com/freecourse
Phenomenal. I wish I had seen your work 10 years ago. It would have helped me so much that I could be retired by now! You have a talent for teaching.
This video is absolute gold. I can't tell you how much this video has helped me. Thank you so so so much. This is absolutely amazing. I've watched so many other videos and they are so bad. Not only is your video free but I can clearly tell that you have thought deeply about how to teach all of this.
The video had thoroughly guided me to create an issue tree. I personally really appreciate your sharing.
Bruno, thank you so much. My 2nd round of interviews is next week and I think that you had the biggest impact on my case preparation. The second best thing was I Got an Offer video guide.
Hey Daniyar, I'm glad to hear our resource has helped you so much. Tell me how it goes on your 2nd round! (You can just comment here once you get the news).
Excellent videos guys!!! Going through a process at McKinsey and am really anxiuos about the case interview.
very enlightening! nice work Bruno and Team, I personally really appreciate your hard work on designing this guide
Thank you a lot guys for such useful videos!
The best case prep channel i've ever seen
I and my friends appreciate your efforts a lot!
That's great to hear! Good luck to you and your friends in your interviews.
Thank you Bruno! This video was extremely insightful!
Super helpful! Thanks for walking us through a problem
Thank you guys for your effort. I believe this is the best channel on TH-cam about case interview preparation.
Thank you so much for your increadible course and your videos 🥰
It was an excellent content with a specific approach to the same topic, I haven't seen this before.
Thank you very much, already subscribed to your channel. ;) look forward to your new videos soon.
Thank you so much for the amazing videos
Thanks Bruno !!
Hi Bruno, thanks for the video, I am learning a lot from here and your web. One Question regarding your takeaway #1. I try to create the issue tree as well, when I have not really learnt the "textbook" structure. My issue tree is based on origin and operational. Origin is related to the fuel (price change, fuel type change) and operational is anything I can think of causing cost changes. Second attempt I make a tree using driver and math trees (as you suggested), I compared the two of them (with yours) and I miss more sub-issues on the second framework then the first one. Considering my first tree is more effective on me yet contradicts takeaway #1 (I mix up quantitative and qualitative issue), kindly asking for your opinion, should I learn to change my views to be like your example? Will it be bad if I keep on mixing up Quantitative and Qualitative issue?
@Craftingcases which software do you use to design the tree?
I would branch out average distance per trip has risen to explain further that the distance has risen due to new routes. Would that be ok or not needed ?
Thank you so much Bruno!
You're welcome, Wiktor! Glad you like it.
Hi Bruno, I have a question. Your reasons for the "avg. distance/trip has risen" branch seem to overlap somehow with the "routes" branch to me. How will you clarify this confusion?
Hi Bruno, I'm not sure I understand why routes is a driver for liters of fuel per km. You suggested if they run shorter routes then lift off and landing (maybe sitting on the runway) increase fuel usage. But if they're running shorter routes isn't that a driver of average distance per trip?
Hey Avinoam,
Taking off and landing spend a ton of fuel. In shorter routes, that amount of fuel dilutes over a lesser distance so the avg. fuel/km is greater.
In the 'Number of trips' bucket, why didn't we discuss about - the fall in price of trips * Increase in quantity demanded for trips?
Will that not effect increase in fuel cost if they are flying more frequently?
Bruno, I have a question: is this the issue tree that you expect a good candidate to actually create in the first 2 - 3 minutes of the case? I say this because I actually thought about these same main drivers for the problem, but I'd probably go as deep as this framework goes along the course of the case, not at all at the beginning. It's a doubt for me because probably it would take me more than 2 - 3 minutes to draw this issue tree and I'm really concerned with that awkward moment of silence while I keep drawing and there's a partner right there in front of me bored.
Great question Pedro!
Assuming you’re doing this case with a partner, it depends a bit on the exact phrasing of the question, but in general I’d give them my first two layers (ideally in less than a minute or so) and ask for data/if I should focus on a specific prt of the tree.
They’ll either tell you to focus on one part, and then you dig deeper there (ideally speaking as you build it) or tell you to generate hypotheses for the whole tree, in which case I’d take each bucket and go a bit deeper on it and generate specific hypotheses for each sub-bucket.
One thing you can always do is ask them - show your work and ask if you must go deeper.
Hi Bruno. Can i ask why didnt u divide 'same fuel from the same supplier' bucket?
amazing
Hi Bruno, brilliant content again.
Being perhaps a bit nitpicky here, but wouldn't it be better to know the type of aircrafts flown much earlier in the case ? For example, we can assume that indeed the fuel cost per KM has strongly risen for one type of aircraft. However total KMs flown could have decreased proportionally for this type of aircraft. The two elements cancel each other out, as in the mix effect you mentioned in a previous video. So the problem would not be here, but in another branch. The only way we could rule out that solution was if we had segmented early on in the case and had used the same framework, but for the two types of aircraft.
To come back to your segmentation video, this gives rise to a more general question. It seems to me from watching your video that to be perfectly MECE and avoid mix effects, we should segment very early on in the cases, but this also gives rise to the issue of how do we know this early on what is the adequate segmentation.
Looking forward to your answer :)
Hey Matthieu, good question.
The problem of segmenting too early in the structure is that you can't know if the specific segmentation criterion you're choosing is the relevant one. For the problem AS A WHOLE there is probably a bunch of criteria that could be relevant.
However, for each small piece of the problem, you'll find that there are fewer criteria that can be relevant. In this example, "aircraft mix" is highly relevant for the "fuel efficiency" driver. Changing the aircraft mix by itself will not (significantly) change $/liter of fuel, it will not change the # of trips you've flown and it will not change the avg. distance per trip.
If you identify that yes, the aircraft mix has changed significantly and that has impacted fuel efficiency in a relevant way, then you should indeed do the analysis you suggested to see if the relevance of the less efficiency aircraft have declined or not. However, this is an overly specific hypothesis for you to test early on, without any evidence that it happened.
Gents,
Excellent video - I have a question: what if the liters per km are intertwined with the average distance per trip?
For instance, suppose the CEO says:
"In order to cover new longer routes, we had to lease new aircrafts with higher fuel consumption"
Do you have any suggestion on how to untangle that problem? Is this something that makes sense?
How often do you see such "intricacies" and how do you deal with them in issue trees?
Thanks
Hey Daniel, this is a very good question.
Yes, this type of problem does happen often. But the best way to deal with them is not through issue trees, as I think it's apparent for you that issue trees aren't very good at handling this type of complexity.
Issue trees are great for isolating and identifying options, drivers, opportunities.
When you have to deal with tangled problems like this, though, the best thing to do is to separate them out into 2 or more "options" and run the analysis of which option is better. If you try to do it all in one issue tree, you're gonna get super confused.
@@CraftingCases My view is that Daniel's example is not an issue, since the issue tree in the video still allows us to acknowledge both factors and both of them have the same directional impact on costs? (the con is just that they could have the same cause which may be perceived as not Mutually exclusive?). What bothered me instead was a counter-intuitive interaction between # trips flown and Ave. dist/trip.
In the situation where direct flights between 2 cities changed to including 1 intermediate stop, the number of trips will increase from 1 to 2. The distance could change from 800km to 500km + 500km. Here we see ave.dist/trip dropping from 800km to 500km which looks like total distance flown has dropped (and thus total costs should decrease), but because the # trips also increased, the total distance actually went from 800km to 1000km.
This makes me think how can we discuss and present such interactions? (more commonly seen in price affecting demand).
I found that using a process structure also works for this problem and covers these complexities a bit easier, at least for me. You could use Purchase Fuel (market price, broker charges), Distribute Fuel (infrastructure use rates, change of volume distribution to more expensive airports), and Use Fuel (rate of burn, burn time, other losses). You can break down to another level for some of these, especially the Use Fuel category. Hope this helps somebody think about this differently!
@CraftingCases What is the name of the software you use to create the issue tree? Btw, thanks for your great contents
PowerPoint
Bernardo Silva
min 9