Complex McKinsey Interviewer Led Profitability Case in Pharma

welcome to the latest case from the firmsconsulting today's case will be a profitability case and it's quite a complex profitability case so the suggestion is that if you are going to be spending time going to this video you should dedicate about maybe 50 minutes to an hour maybe even a bit longer and to understand the details that sit behind this case it's a profitability case but a difficult one at that and what I would suggest you do is that before you begin the case maybe get a few clean sheets of paper there'll be a lot of calculations a lot of conceptual issues you need to understand as we talk to you through how to solve the case and finally when we present the case question it may be a good idea to print the case question and keep it next to you because throughout the case we will be referring to details that have been presented initially by the interviewer so it's a very interesting case and the reason we pick this case is because we wanted to teach you several concepts with regards to the pharmaceutical industry but we also wanted to show you that profitability cases are not as easy as they look too many candidates profitability cases usually the ones they practice right at the beginning and they just naturally assume that they are so simple they could ignore them and move on to more complex cases like competition strategy turnarounds and so on and what this case does is it shows you that things are not as obvious as they seem so we're going to run through the case as if it was in a real case interview and talk you through firstly how you should approach it and the rationale behind some of the things the interviewer is asking for and how you need to respond to those things so in any particular case the first thing you would get is you would receive the case itself now the case could either be read out to you so you're sitting across from an interviewer and the case could be read out to you or the case could be given to you as a printed sheet of paper either option could work right this particular case is in McKinsey case is that the difficult level is not something that we expect people to be able to grasp if they're beginning their training in fact this is the kind of cases that we would give to our candidates towards the end of the training usually when they've finished all of their training and we're going through drills to test them across more more difficult cases if you'd expect to take about 30 to 40 minutes up to an hour to do this case and in reality very few candidates would be expected to complete the case in its entirety and more likely the interviewer would point out different parts of the case that they want you to focus on and they want you to explore in more detail now this is the case I'm going to read through it and what I've done is I've split it up into two parts right the first part is the part in black and the other part is in gray depending on the interviewer they may give you all of the information upfront or they may withhold some of the information for example if I was running a case I would only give you the part in black you know the CEO of a client recently commissioned an internal benchmarking initiative the initiatives results which are not being questioned indicate the clients cause up Farb of competitors the CEO wants you to reduce R&D costs you want you to start here while growing revenue and all of the rest of the information which is very important is information that the interviewer may give you throughout the case and drips and drabs or they may give it to you up front now if they give it to up front it's going to be information overload and obviously if they give it to you later it depends on you asking for that information or taking a hint when they offer it to extract the information so let's go through the detail write all of the R&D facilities and staff are concentrated in the Boston area there are approximately 2,300 R&D employees and the total R&D costs are approximately 12 billion dollars per annum the client has traditionally received 80 percent of his total revenue which is 24 billion from drugs to treat heart disease blood pressure and liver disease eighty percent of these patents on these three drugs will expire by 2015 and the company therefore expanded research into adjacent areas in 2010 not getting satisfactory results the company again expanded the research portfolio in 2012 the clinical trials team is pushing for an expansion of R&D citing their ability to conduct 20% more trials than their current volume of every 10 drugs in the final stage there are six drugs in the final stage of the five stage trials process only three will go on to receive FDA approval two of the three will recoup their initial investments two them to market and this one of the two will generate more than 1 billion dollar in revenue per year and the first thing you need to do when you receive information like this is you need to take a few minutes to understand what you have been given it is impossible to solve the case unless you understand what I've been given to you so what I would do is and what I recommend people do is once the information is given to you ask for the interviewer to give you a few minutes and you highlight all those things in the case that are not really clear or the quiet clarification another way of saying it so for example the initiative results which are not being questioned does that mean that you you in the case enough men to question the study or does it mean that the study is not being questioned but it should be questioned patents you know 80% died expiring what do they mean by expanded the research portfolio what do they mean by the five still trials process what do you mean only three will receive FDA approval those are some of the things to me if I was doing this case a bit ambiguous and require clarification so I take a few minutes no more than three to four minutes go through the case and I would come up with a list of questions that would help me understand the case and clearly anything that is in the case question itself that is not clear should be part of those clarifying questions right so what happens next right you have this clarifying question clearly you will have to start by actually asking the interviewer the questions you've collected when you took those few minutes to yourself right but remember this if you've given a lot of information up front the odds are pretty high that the interviewer is not interested in hearing more clarifying questions you've been given so much data upfront the interview pretty much believes that because of the volume of information you've been given you should be able to come up with the structure at the beginning if you've been given less information up front like maybe just a two-line case then the odds are pretty high the interviewer would tolerate additional questions however that being said you should always ask the interviewer what they would prefer so I would do something like this I would after I'd finished my three minute pods I'd say okay I'm ready to begin now thank you for the information would you prefer to see my approach prior to my class fine question so what you're telling the interviewee is that you do have some things that are not clear but you know that you may want to see the approach first you're giving him an option so the interviewer may respond and say yes tell me you're clarifying questions or this is the McKinsey case most likely they would say yes please give me your give me your approach and let me offer you some you know tips now here when they talk about the five stages of probability figures it's a hint right take the figures and understand what they are useful there are very few cases where interviewers try to mislead in fact interviewers are not trained to mislead you they don't start a case by saying okay how can I deliberately lead you down a wrong path right interviewers don't mislead in fact due to time constraints in a case they try to provide guidance every single time they speak so every time an interview is speaking to you they're trying to guide you sometimes they try to test whether what you saying make sense but they're never misleading you every time they speak they're giving you more information in fact the more times you get an interviewer to speak the more times you are collecting useful information in the case right so whenever an interviewer speaks listen to the verbal clues very carefully take them down in fact in one of your sheets of paper you should have two sheets of paper in a case one to write out your structure another one for calculations in one of your sheets where you do calculations that create a little column and all the information the interviews giving me I'd writing I'd write it down there I might have not to use it immediately but I'd tip it because clearly it's useful information now we're going to do a little bit of a timeout Jim and there is but timeout is because although we've structured this case well but at least we've attempted to structure the case by understanding it I think about 99% of candidates don't know what to do with its information once they get it you know once you've asked these clarifying questions how do you actually build a structure or a framework for your case you know do you build the decision tree do you throw our hypotheses do a combination of both how do you move forward so what we're going to do is going to take a short time out and show you how we would typically structure a case and then we'll move back into the case there are typically three approaches to solve full case questions and I distinguish folkways questions from estimation questions brain teasers and market sizing cases for those group of cases that is different approach to solve them which are covered in our other videos right so what are the three approaches to solve full case questions well let's think about it for a second we have the famous hypothesis based approach or the hypothesis based approach which to be fair I mean many people refer to it and talk about it a lot but I think most people don't know at all how to use it and what it means finally we have the decision tree approach which many people consider to be different from the hypothesis based approach and then we have the no structure approach where you basically guess your way through okay so let's give you some examples of all three approaches in the apophysis approach you would based on the information you received in the case you tell the interviewer well I think that the increased clinical trials is leading to higher R&D costs and this is this will be something I want to examine that's an example of a hypothesis the decision tree is basically the left hand side of this chart right we're going to talk about this a lot more throughout the rest of the video so we're going to park this issue for now but basically a decision tree allows you to break down an issue into so I break down a question into either the drivers of the issue or the options to solve the issue we're not going to delve into this now because we're going to cover it extensively later and no structures where you just guess you know you say well I read this in the news that maybe salaries are too high and they're inefficient you're guessing no one solves the case by guessing and you're out undeniably going to be rejected at some point that interview and I'm pretty much going to piss off your interviewer now let's look at hypotheses what I want to do is show you why I policies are difficult to work with and then show you how you can bring structure to your hypotheses so let's come up with three hypotheses based on the information you've already seen in the case you may want to pause the video here and try to develop three hypotheses yourself and compare them to what we come up with or you can simply watch the video and practice later so that you've seen the first hypothesis let's look at the second hypothesis right and what we're going to do is as we structured all the hypotheses we're going to show you why they're difficult to work with and what exactly when we what do we mean when we say that difficult what makes them difficult and why that's a problem in solving cases well the second hypothesis is that we can say launching research in so many areas so quickly must lead to so many inefficiencies which are areas that we need to explore right that could be a plausible hypothesis and the third hypothesis which will we would end up with just three so I wonder if there are any synergies between the trials like procurement of material and we may be able to cut cost there as well now I want to ask you this now would you be proud of these hypotheses and consider them a good start to the case and ask yourself this know do they make sense to you I think about 90% of people would agree that these are great hypotheses because the first one looks at clinical trials the second one looks at inefficiencies and the third one looks at synergies you know broad enough areas to cover everything but the reality is that these are part that these hypotheses are not mutually exclusive right inefficiencies is the flipside of synergies think about that very carefully the opposite of inefficiency is synergy so when you tack when you measure inefficiency in hypothesis two you're actually measuring the opportunity for synergies which is covered in hypothesis three and I part this is three when you're looking for the benefit some synergies you're obviously looking at the opportunities to cut out inefficiencies so i potus's to and I put this is three are not mutually exclusive moreover it's clear some of the inefficiencies of synergies are going to rely are going to be related to clinical trials so we can see that I potus's two and hypothesis three are not mutually exclusive but we can also see that I pollicis one is is partly measured in hypothesis two and I potus's three so we are double counting in some cases triple counting here right and that's the problem when hypotheses because they tend to be this fairly large mass of text as you can see it's very difficult to know when they are mutually exclusive and this raises three problems with hypotheses the one is that they require prior knowledge right it's very difficult to develop a hypothesis or a set of hypotheses for a case unless you tend to have prior knowledge about how to do them or about the sector and the company despite what people may say that's one of the rules about hypotheses you need to have prior knowledge second is is that a very may see light what I mean by that is it's difficult to develop hypotheses that are mutually exclusive and collectively exhaustive because there are so many parts to an hypothesis it's very difficult to lay them up side by side and say what is in one what is in the second hypothesis what is in the third hypothesis and are they different and finally the purpose mean you've come up with these beautiful hypotheses but to solve what problem because at this point we haven't really figured out what the problem is in the case we're given a lot of data until the cut cost but the problem is not clear and if you have a lot of hypotheses but the purpose of their policies and I've clear how do you know you are solving anything meaningful and this obviously raises two obvious questions you know which approach should you use to tackle cases and then how do you build a framework so do you use the hypothesis based approach decision trees do you guess your way through it first question second one is once you've chosen approach how do you build your structure well what are we going to show you is that decision trees can drive hypotheses and in fact the best decision trees do Drive hypotheses and in reality when you're not an actual project with McKinsey & BCG and Bain and you are in a sector where you know nothing about it you are forced to build hypotheses first and then tack your hypotheses on to those decision trees so let's now continue with our time out and actually build what we have your line so what we're going to do now is we're going to show you how you build the structure to yoke and the most important thing we want you to understand here is that when you get a case you must understand the key question you are trying to solve in the case and to many of you the concept of a key question will be alien because it's not something that's taught formally in any case books around the world but it's an important question and I'm going to prove to you that it's an important question right let's look at the case we've been given earlier ask you to print it out and have a look at it so print it out have a look at the printout one interpretation of the case would be how to reduce costs and increase revenue simultaneously detain it sorry simultaneously that is the purpose of the case and if that was the purpose of the case the sub questions to solve that case to be how to increase revenue and how to reduce R&D costs but what if as the result of having a discussion with the interviewer he tells you to only focus on R&D costs and ignore revenue for now then the key Kastner change wouldn't it the key question would be how to reduce R&D costs because the interviewers told you to just focus on R&D costs and clearly the sub question is to reduce R&D costs would be the buckets of R&D cost which is labor equipment costs land and buildings inventory and clinical trials but what if the interviewer changed it even further what if he said okay I want you to focus on cost in this case but only costing it can be cut in the short term well clearly inventory and clinical trials are the only things I think can be cut in the short term labor costs given the legal requirements going to take a few months to fire people or euphemistically you know right size the business will take a few months to inventory equipment sell them and close the sale it will take a long time to sell land in buildings it's not something that happens overnight now what do you notice here well you should notice that depending on the way you read the case and depending on what information the interviewer gives you you can come up with one of three and possibly even more key questions to solve in this case but you have to agree the key question you are solving in the case right with the interview because if you solve the wrong key question or if you identify the wrong key question you will solve the wrong part of the case and as you can see that different key questions generate different sub questions and what you should know about your sub question is that is that they are your framework right so if you want to know how to build a framework for the case well the way to do it is identify a key question identify the sub questions and the sub questions become your framework and if we look at the middle scenario how to reduce costs right the first part of your framework could be these five areas for analysis and we could break down labor costs into may be fixed labor cost and indirect labor costs or specialized technical and administrative labor cost the point is if you identify your key question if you break down your sub questions you automatically start generating your framework and in fact what this approach does is it teaches you to build frameworks without the need to memorize frameworks and we get this a lot where people would come to us and say you know I both this approach but I don't but people tell me doesn't exactly solve the case because you're building approaches independent of the key question the case you're just building up you're just using an approach because it sounds vaguely familiar but your approach is not directly tied to the case you just step back for the manager we've been given all this information from the case that hold sheet at the beginning from that case we asked a few clarifying questions well we were allowed to ask one from there we identify the key question and we agreed with the interviewer and then we generate some questions which becomes our framework now what happens from here right now we're going to show you how to build iPod to see some a decision tree going forward it's not difficult to do sexy to do and the way we're going to do that is we're actually going to use the we're going to use one of the scenarios that we listed earlier and we can expand it even further so let's take the middle scenario where we had to reduce costs and we know there were five areas to analyze so we're going to build an hypothesis for just one of these five buckets of analysis right and the area we're going to analyze is how to reduce clinical trials now the initial hypothesis we developed much earlier which if you go back in the video you'd see is we said I think the increased clinical trials is leading to the higher R&D costs and this will be something I want to examine for those of you understand how to build AI policies you'd know that is not a hypothesis and AI practices let's see first explain what an iPod is this is you look at the definition of hypothesis in the dictionary it comes from a Greek word I believe and it has something to do with having an observable phenomenon so let's assume you have an observable phenomenon which in this case could be anything from increase R&D constituted do to decrease profitability you have an observable phenomenon something is causing that observable phenomenon to occur and that observable phenomenon has an effect on some other part of the business your hypothesis should consist of three it should say do 2x there has been an increase in Y which is the observable phenomenon which has led to Z now the reason we break and I potus's into three parts is so that we can test it when an eye potus's only has one part for example some people would say my potus's is that labor costs are going up it's like a no-brainer it is almost impossible to test and even if you can test it you cannot test cause and effect and that's the key thing to an hypothesis you must be able to test cause and effect and of course with this kind of structuring of your hypothesis you can collect data to have it verified now what I'm going to do is using this structure I'm going to take this hypothesis that's written on the top and rewrite it in a correct structure the codecs truck should be due to patents expiring there has been an increase in clouds to find new drugs which has led to an increase in R&D tests and this is verifiable you can look at the company data to see when patents expire tick there has been an increase in trials to find new drugs you can see if there's been an increase in trials to find new drugs and where there's a correlation with patrons expiring and obviously then you could clearly see whether R&D costs have gone up and the most important gift to understand here is that a key question structures the decision tree as we've shown earlier and if you want to you can use that decision tree to structure your Mesa hypotheses the other thing you need to understand is that you don't need hypotheses to solve the case as you can see here on the left hand side the decision tree gives you a lot of structure anyway and you can pretty much solve the case by focusing on that structure without the need to generate hypotheses so don't feel a need to generate hypotheses you don't really need it if you feel comfortable doing it that's great if you feel comfortable using decision trees that will also be the right way to proceed so what we're going to do now is that we're going to end off by tying up the timeout because we started off by showing you the different kinds of ways to solve cases hypotheses decision trees we then zoomed in a little bit further and showed you the importance of the key question to building your sub questions which become your framework we then showed you how that framework can be used to generate hypotheses and now I want to delve a little bit into how you actually build these sub questions which lead to a framework right so now we're going to use the same example this is one way to break out the sub questions or sub issues of the key question another perfectly plausible way is this if someone told me if we are doing a case and we agree that the key question is how to reduce R&D costs if someone broke down R&D costs into the five major cost buckets that's good if someone broke down the issue into these five areas customers competitors cost company and market structures also fine but which is correct well there is a difference between both the first one you are breaking down your sub questions into the Macy options in other words what are the independent options that can help you solve this question in other words if we reduce labor costs could we reduce R&D costs yes they are mutually exclusive because you can reduce labor costs independent of equipment costs just as you could reduce in the equipment cost independent of land and buildings inventory and clinical trials so in the first approach we've broken down the sub issues into the options in the second one into the areas of analysis approach is 1 & 2 are both correct but approach 2 is inefficient and why is that well let's think about it for a minute right think about the typical format of the case we you've given it you've been given a case in writing or verbally you've got to understand the issues you're going to ask clarifying questions and build out the context you then got to analyze the problem through your questioning and as you analyze the problem you start in your head start understanding the problem and once you've pinpointed the problem you develop options to fix the problem usually develop up to three or four options you then come up with a set of criteria to analyze the options pick one or two preferred options and then give you a summary and your recommended option right so it should be obvious to you by now that the singular purpose of analysis is to identify the options we do analysis to identify the options to fix the problem we don't do analysis because well you know every consultant does it do analysis to identify the options and then we analyze the options as well but if the options are apparent you should jump to them immediately because it cuts out some of the unnecessary analysis time upfront and when presenting the options first for those of you who are familiar with this technique I mean Bain likes it McKinsey likes that you go for the answer first approach that's what the answer first approaches where you present the options to solve the problem up front but of course that implies that you should know what the problem is if you're not clear what the problem is and you don't understand the case perfectly then it is okay to structure the issues around the key question around areas of analysis that will help you identify the options so with that in mind we're now going to jump back to the case using that theory and show you how we try it so this it's quite a simple process here where we simply going to use the information that we've been given to sketch out a key question and sub question and if you can't think about the issue of revenue growth and cost reduction I mean the interviewer tells you that cost reduction is the priority but you also want you to look at revenue growth so you need to think about that well the key question here is how to reduce R&D costs from 12 billion while growing revenue from 24 billion whenever you sketch out your key question is is vital you put in numbers numbers give you direction numbers tell you the scope of magnitude of what you're trying to achieve the key question void of numbers is not a very good key question in fact the key question void of numbers is probably not going to help you at all again you then break out your areas of analysis now the thing with this you could either break it down into fixed and variable buckets which is be fine but I picked the major buckets of analysis what you must always remember is that when you interview with McKinsey they always like it when you say these are the broad areas I'm going to analyze give a brief description brief description and tell them why you want to analyze it so for labor you could say that well you know we've been adding so much investment into R&D hiring so many people but the number of drugs we've been putting out is decreased so there seems to be some inefficiency here and then obviously if you expanding new areas of research you're probably acquiring new equipment especially branching into new area as mentioned like you know epi' Titus and so on they probably are opportunities to consolidate this or buy less all the staff are based in Boston they would seem to be opportunities to consolidate them because commuting should be much of a problem chemical inventory must be a problem the more clinical trials you run and the more compounds you are researching well clearly there must be opportunities here to either you know change procurement at least get some efficiencies in and then you know the aggressive push to find new revenues must be driving trial so we need to find a way to improve the hit rate and our trials now you just step back a second so you've given the interview your key question you very quickly at the same time went into your buckets of analysis but which one do you analyze well when you are not sure which bucket to analyze there are two ways to find out the first one is split the cost these are all costs so find out what contribution is labor make up of total R&D costs find out the percentage always convert percentages to absolute numbers it becomes a nightmare to work with percentages if you don't convert them to absolute numbers whenever you are working with the decision tree even if you continue building out this decision tree always carry your numbers across to prevent arithmetic errors but also by carrying your numbers across you can see the relative sizes of the branches and you can determine which one is more important than the other one now many people would say because trials and labor are the biggest buckets they are worth analyzing but the size is not the only thing you need to consider you also need to consider the trend if trials cost was 38% but was steeply going down then you may not want to analyze it but here we see labor and trials costs are large but going up so clearly that has to be the two areas we want to handle eyes right I mean there's the cost buckets are large and they're going up there's you know unless there's something seriously crazily unnatural about equipment and machines there's going to be this huge order coming down the line which is information the interview would have to give us it will make sense to analyze labor but obviously starting with trials the other thing you also have to understand is that this is the application of the 80/20 principle but looking at the size of the buckets and when looking at their trends we have determined that these two out of five major areas will most likely give us 80 percent of the answer and now we've got to continuously Bowl that and what most people would do is they'd say let's start with trials and let's just do what we did with with R&D costs which would break down trials into its overall area right that Baker and the preclinical stage 1 stage 2 stage 3 and awaiting approval these stages shouldn't be new to you they were mentioned initially in the case you may have had to ask for clarification with all this information is there and they would have done what was done earlier they'd split down trials into its component bucket so preclinical makes up 20% they would carry the numbers across and they'd work out the trend and most people will be pretty happy with themselves right they would say wow this is a fantastic we're doing everything right but the problem with this is you're following the numbers too closely without thinking very carefully about the information you've been given upfront right so think about what was given the upfront right you were given a lot of information about patents expiring and all kinds of things now we know that they want to increase revenue and they want to reduce R&D costs but actually the amount of R&D done directly drives revenue it may have be a lag effect but it definitely drives revenue R&D done today it leads to revenue three or four years down the line right we know that R&D costs are being driven by the number of clinical trials of the fact we know that the more clinical trials run increases the probability of finding a promising drug that's obvious a promising drug drives revenue growth you find a promising drug you drive revenue growth so since revenue growth requirements drive trials which drives R&D Krauss we need to calculate the revenue growth requirements because that is driving clinical trials which is driving R&D costs so in other words our only costs are not independent of greven your growth they are linked if you cut revenue growth if you cut deaqon so you cut trials you will end up cutting revenue but before you cut revenue you need to determine what your revenue targets are because if they are linked to your revenue targets would drive your clinical trials to requirements which drives your clinical trials costs which drives the R&D costs and that is the big insight here now all this data has been given to you it's just about putting the links together let's work with it so let's just do the math let's calculate the revenue gap we know the revenue is 24 billion dollars we know that 80% of the 24 billion dollars comes from three drugs or three classes of drugs and we know 80 percent of the patents for those three classes of drugs would be expiring in 2015 so 80 percent of 24 billion gives us the three classes of drugs 80 percent of that number gives us the amount of patents that are expiring which is 15 point four billion dollars right now if we assume that they will lose 50 percent of their revenue in the year they lose their patents which is 2015 then we can work out that the revenue gap they need to fill is seven point seven billion dollars in fact even if you ignore revenue growth just to keep the revenue wait is today they need to find seven point seven billion dollars in new revenue in other words they need to find about seven to eight billion dollar drugs now you can make a different assumptions in terms of the market that will be lost when the patent expires you could say ninety percent will be lost you could say 50 percent I would refer three percent it seemed like the right thing to do right but more important data has been given to you also at the beginning and you need to use it so let's just work with that at the beginning you were told that three out of ten drugs passed the final stage the 30 percent going to the next stage we are told two out of three from that stage pass HDA or through approval which is 66 percent off the two out of three one will become a billion dollar drug which is 50 percent you were also told six drugs and the final stage so just simply plugging in the numbers will tell you that you have a very very bad pipeline coming up in fact the probability says you only going to produce half of a billion dollar drug which is a 500 million dollar drug right now remember the interviewer offered to give you the probability numbers for each stages well let's assume he gave it to you and then you have to carry the six upwards do the math roll it upwards to figure how many compounds we're being put into the pipeline at the beginning of the funnel to generate that 500 million dollars worth of sales on the new drug or half of a billion dollar drug right so basically four thousand four hundred and sixty compounds are to be put into stage one to get point six of a billion dollar drug now this changes the case dramatically because we now know the minimum target we are chasing of new revenue is seven billion dollars and if clinical trials cost an R&D costs are linked then we have to figure out what the link is we now know what it is and we have to now figure out what we need to do on clinical trials to get seven billion dollars worth of new revenue I'm going to put it you in a different way because what we've now done is we've identified the problem the problem here is that our only constant at independent of revenue if you want a certain revenue target you have to put in a certain amount of compounds at the beginning of the R&D process in the preclinical stage right the more compounds you put in obviously the more drugs come out at the bottom so what are the options available when any well let's think about this for the second round we have a number of options available option one if you increase the number of preclinical compounds going through at the beginning you obviously have more chance of drugs coming out at the bottom right lot that's basic logic if 4460 compounds gives you a point six of a billion-dollar drug obviously more than four thousand four hundred sixty compounds will give you mobile into all the drug so that's one thing you could do the other thing you could do is not change the number of compounds going in but you could change their weighting so put in four thousand four hundred sixty compounds but rather than investing in areas like hepatitis and HIV with company does not have lot of expertise and where the probability of finding a billion-dollar drug is smaller where the market size may not be large enough you could weight that portfolio towards areas the company knows better and where the market size is large enough and finally ignore the portfolio but you could look at things to do to increase the probability of success for example the FDA may only want to improve this drug if the company committed to doing certain kind of advertising around the usage and the risks of using the drug and the company could agree to do that which would increase its probability of the drug being approved at FDA stage now I'm going to do a little time out here again because I think it's very important that you understand the way we came up these options are not the same they are profoundly different and just to recap the case we were given a lot of data you focused on costs because that's what the interviewer told you to do but you figured out that costs we're being driven or there's a link between constant revenue and you realize that the revenue requirements were driving R&D constant clinical trials costs we then worked out what the revenue requirements we're and you realize that given the revenue requirements you have to increase the number of clinical trials going through the process and you came up with three ways to increase the number of billion-dollar drugs in mind so either increase the number of compounds you find some way to produce more billion-dollar drugs and these were the three options they came up with to fix the problem and if you step back to the first time out we had remember we said you do the analysis to arrive at the options right so yeah we did the same thing we we did someone even though we listed some options upfront we still did more analysis to arrive at a much more succinct and much more specific options so let's understand the differences between these options right now we know the current portfolio equals four thousand four hundred sixty compounds in the preclinical stage and for the sake of argument let's assume it's foot it's split into four broad areas of research let's just in 56 percent of the compounds I've a heart disease 22 percent of a liver 12% for lung or whatever it is right now option 1 what is option 1 saying option 1 is saying that the ratio split in the areas of research would stay the same but we'd be sending more drugs down the funnel so for example let's assume we're going to double the amount of compounds going into preclinical research the ratio split between the four areas would stay the same so we basically increase the size of compounds going through we're going to research more compounds which would raise the associated labor equipment and material costs right robot option two well option two saying we're going to split that we're going to change the portfolio maybe we're going to drop one of the areas we're going to reset so we're going to change the portfolio right dropping a compound reduces associated labor company material costs so you can see the difference between one an option - an option one you increase the size of the portfolio but you don't change its rating an option two you don't change the size but you change the portfolio weighting and a little less look at option three in option three you will go outside the portfolio so the number of drugs stay the same the split or the weighting of the portfolio stays the same but you try to influence things that are not part of the portfolio like FDA approvals packaging distribution and so on right so just to recap option one will increase revenue but likely increase cost as well and we don't want that we try to cut costs while we're trying to raise to decrease revenue option two is going to reduce the size of the portfolio which will lower cost but it could also increase the revenue by switching to areas of analysis or compounds or themes of diseases where we have that expertise an option three well option 2d is not changing the portfolio so costs are unlikely to change there but it could lead to major changes in the way the drugs are marketed of the distribution which will lead to its own cost so revenue and cost go option option one and two so an option one and three but for option two costs are likely to go down while revenue goes up and obviously that becomes the option we want to analyze in greater detail which then takes us to the next part of the case because now that you want to analyze the portfolio clearly the most important thing you have to do is you have to get a breakdown of that portfolio in this particular case you need to ask the interview there's any data about the portfolio of compounds researched and the success rate and cost attached to each part of the portfolio and you have to assume for example that the interview provides the table below and what I'll do here is maybe take a time art pause the video to analyze the table below and see what kind of insights you develop from it before we go through the analysis now the most important thing about reading any graphic is that you're not looking for this one or two pieces of insight that will just solve the day you've heard of the concept of storyboarding I'm sure by now if you're doing this case you know you've done a few cases you've done some researcher management consulting so you know the concept of storyboarding storyboarding is where you take pieces of insight or pieces of analysis and you lay it out together to generate a story in analyzing exhibits exhibits it's no different in a table like this and it's quite a complicated looking table you're going to find more than one important insight do not treat each insight as an independent issue if you treat them independently you're not going to see the story out of the numbers what is more important to do is to capture all of the inside so write them down a piece of paper and then look at the story generated from them now I've done that as well and I'm going to show you how I've rearranged them to generate a story I've come up with four insights here and will show you how it was rearranged to generate a story well the first insight is that 47% of R&D spend goes to areas where the client is strongest right art blood liver and 53% is broadly in the infectious areas category where they have no skills and capability all right and they're spending 53% of their money where the markets are smaller there's a lower probability of success and what competitors are spending moment why would you cut down your spending where you are strong where the markets are larger and where your competitors have no hope in hell of beating you it just doesn't make sense to do that and beyond that look at the number of drugs that was generated despite the cost spent right so basically in these new areas a lot of money is being pumped in but the cost to produce one drug is much higher which means the yield for the drug is much lower obviously it costs more to produce a drug the time to recoup the profits takes a lot longer and considering the fact that patents for all drugs are the same drugs that take longer to have a payback have a smaller pedia to recoup profits which means their yields will automatically be lower and then finally there's no rationale for doing this because applying clients existing knowledge to the unrelated infectious disease market makes no sense right and what you can see is I built a storyboard for facts or for insights on the data a layered upfront gives me a storyboard now you have to use this information and think about how you could talk about the recommendations you are going to make you've done a lot of analysis now and by this point you have to come up with some steps forward right what would I recommend well this is what I would recommend from the case cutting the trial spend by 53 percent that's the total budget on new infectious diseases saves two point four five billion dollars right and since these infectious diseases have a low probability of finding drugs and will take a long time to come to market the immediate revenue impact is going to be low well would take the two point four five billion dollars and reallocate the majority of that to the clients traditional research areas of heart live and blood because the table earlier shows me that the markets they are a much more promising the probability of success is larger and they yield for those drugs is much higher let's do some calculations right we know that 4460 compounds are run through clinical trials and it costs 4.6 billion dollars for clinical trials favored costs just over a million dollars to run one compound through to the entire trials process right if we are going to be reducing our trials budget by two point five four five billion dollars what I'm going to do is I'm going to take all of the 2.5 4 or 5 billion dollars with the exception of 1 billion dollars and reallocated just to heart live and blood which means my new trials budget would be 3.6 billion dollars and I'm spending roughly 1 million dollar per compound which means that I am now going to be spending I'm now going to be analyzing 3,600 compounds I still have the billion dollars of savings which I have not used more than that we also know that the probabilities of a success are going to change because we are no longer analyzing drugs with a low probability of success so yeah you can assume the probabilities are going to go up anywhere from 50% 100% some people would say hundred percent because the data shows it but I would say put it at 50 percent because there are other things that could go wrong as you scale the process and now you can redo the calculations right these are the original calculations you can put in the new clinical trial numbers you got 3,600 drugs going through with new probabilities of success each one going up by 50% which means now even though you've reduced the number of drugs in preclinical stages from 4,000 for 4460 to 3600 you have about 4 billion dollar drugs coming out of the pipeline right so now that's the analysis now how do we tie this all together obviously the tie-up is the most important thing it's what counts at the end of the day so let's think about how we actually tie this all together and the important thing is to know is that how you explain the data you've put together is obviously far more important in the actual data you generate so let's think about this for a second right cutting infectious trials completely and reinvesting all that savings which is about two point four five billion dollars with the exception of one billion dollars in the company stronger research areas were just about increased billion-dollar drug output seven times right it goes from point six billion dollar drugs to three point seven billion dollar drugs these actions still save one billion dollars we haven't allocated one billion dollars but the increase the billion dollar drug output basically four times right it still leaves us short three of the 7 billion dollar drugs we need you could argue they need you know eight just to get a slight revenue increase but I went with seven here because the numbers are approximated anyway now if we assume R&D labor is directly proportional to the number of compounds in the trials which is a fair estimate to make we could assume the ratio decrease in labor costs is equal to the ratio decrease in trance now think about this for a second if you're decreasing trials by 22% so you're still saving a billion dollars right there's 22 percent of the total trials costs you can it's fair to assume that the labor is split equally across these areas of research right so labor that would have been used if you spent the billion dollars who are no longer going to be used and you can assume is roughly equal so if labor is three point six billion dollars 22 percent of that is 790 million dollars so what we can see by canceling infectious trials and reducing infectious labor research costs we can save one point seven nine billion dollars in fact you could probably save someone equipment as well but we didn't put that in here because we're given new equipment information but you could always ask for that as well you may really consolidate some buildings it's possible but no data was provided for that but you can mention that to the interview now you still have we saved one point seven nine billion dollars throughout this process but you've only produced four billion dollar drugs now all other things being equal right that one point seven nine billion dollars can be plowed back into R&D particularly trials right to generate two new billion dollar drugs how do we come up with two well if three point six billion dollars generates four billion dollar drugs then one point seven nine billion dollars which is half of three point six billion dollars should generate two billion dollar drugs but it still leaves us one billion dollar drug so no matter even if we reallocated the R&D budget and didn't cut it at all so we just you notice that we just reallocated at this point all the sabres now plant it back into research so R&D budget hasn't changed on trials or anything like that it's simply been refocused on heart and the traditional areas we still short of 1 billion dollars which means we need to raise our R&D budget by nine hundred million dollars just to produce that 1 billion dollar drug shortfall the reality is we could not grow revenue and cut R&D costs at the same time an acquisition may be the more realistic option given the time and funding limit and you can always ask the interview should I analyze this and the only way I'll tell you know maybe it's really up to you right but at this point you've pretty much answered the case it's may be worthwhile spend some time looking at some of the observations and lessons in this case the first thing that's important to understand is that this is a data intensive case but clearly if you just follow the data you will make mistakes in order you need to really think curve for you know what are the relationships between the data you can't just follow the data you've got to think about the relationships between them in cases that are this complex and with so much data floating around you cannot solve the case unless you have a very good structure build up front and without the structure what will happen is that you'll constantly find interesting data worth analyzing that has nothing to do with solving the case the structure guides you it tells you what's within scope and what's out of scope things may look really interesting but if they're not within scope you ignore it this case is too long to complete in one session there is no way that an interview is going to ask you to do this case in one session right it's most likely they're going to pick pieces of analysis and ask you to focus on them and of course because it's a McKinsey interview in that case that works out quite well because you know the individual pick areas you want you to analyze but we reason why this case works well is because people want to test your math have ample opportunity to do that and people want to test logic have ample opportunity to do that so it's a bit for everyone here I think the other thing you need to understand here is that many candidates try to solve cases by rote they they just throw out a structure because they've learned that this is the structure to use and they work with that will hurt you if you did that you'll end up analyzing cost will you blue in the face and maybe you'll figure out the link between revenue later but in other words you know you're going to be inefficient it's important to identify the key question build out the sub issues which become your framework and then understand the relation if there's any between those issues right what you realize here is that R&D costs a not mutually exclusive of revenue even though our end is not a variable cost right another important lesson here is that pharmaceutical cases invariably use the stage-gate approach you know preclinical step stage 1 stage 2 stage 3 and then the final stage the stage gate approach where you work with ratios and probabilities of success has been adopted by just about every industry from the oil and gas industry to business cases venture capital and so on so it's important you understand that process and you know how to deal with it in this case we present revenue and constantly present these two separate issues it is very unlikely that in a real case two issues will be so unlinked or unlinked at all mostly they will be linked in it's your job to find that linkage and finally I think it's important for you to understand the decision trees and hypotheses are not mutually exclusive problem-solving techniques we've shown you why they are the same technique and in fact if you read Kenny chioma's early book the mind of a strategist the guy who basically built McKinsey strategy practice when he was director of the Tokyo office and a greater Asia he explains it very well hypotheses and decision trees the link is just recently you know as consulting training has become a cottage industry people have tried to separate them because they haven't seen the linkage between both but as you can see in this case they are very much linked they are not different the question is whether you want to use the decision tree to solve the case throughout or whether you want to use the decision tree to generate hypotheses and then solve the case either option can work it's really up to you where you want to take this I think this is some of the very important lessons there are obviously other lessons I mean if someone weak in math is obviously going to take out more math lessons those weak in conceptualizing construction cases will also take out very different lessons but I think the most important thing here is that you have to understand how to structure a case and how to build it right you can always play back the video to understand the approach for me the most important thing is that people would look at the timeouts because the timeouts present the background information in terms of why we do things the way we do and why entries and hypothesis links so well together and why the key question is important why so many people ignore it and so on right now fine let's just present a case summary and what we've done is we've scored this case against other cases right on a scale of 1 to 8 and we've given you explanations of the different writing you can see that this case goes very highly on communication and case structuring you cannot solve this case unless you have the very strong ability to contact so that contact to be able to communicate with the interviewer extract information understand what they saying use that information and guide the interview in your thinking if you cannot do that you cannot solve this case this case calls for strong communication skills and very strong case structures because the mathematics skills need to be very stronger because there's just so many numbers involved and business judgment skills tend to be the weakest yeah now what we've also done is that we wanted to give some guidance in terms of when you should do this case this is not a case you do at the beginning this is a case you do with us once you've done about you know 12 hours of coaching with us you understand all the fundamentals really really well and you now want to apply those fundamentals right do not try this at the beginning you will get lost during this case at the beginning you have to learn many different skills which we teach in other cases and then you need to bring them all to bear in this one case also do not expect to solve the case in its entirety or identify all the analysis requirements no interview expects you to solve everything the point is solve what they ask you to solve and be able to communicate in areas they want you to communicate right secondly have the skills to use the interviewer for guidance and use the prompts provided that's the communication plant the interview will guide you in terms of what needs to happen here finally this is an MBA level case pre-mba levels business analysts are not going to expect this case or associates at BCG and so on right the complexity of this case means that if you are weak on communication structuring the problem will appear very quickly so if you week at those areas right at the beginning of the case the interview will pick it up because this case forces those issues to come up the other thing we've done is that based on all people we screen not the people that join our program but the people who screen who want to join our program we have given a general rating of how MBA candidates would perform on communication business judgment arithmetic and case structure if they did this case the average MBA candidate would do poorly on communication and business judgment pretty well in an arithmetic and okay and structuring the average pre MBA candidate undergrad or master's student someone without an MBA would do roughly the same on communication roughly the same on business judgment we find that pre MBA candidates though it is the ones we pre-screen tend to be very strong at arithmetic but very weak on case function because they don't have strong enough business skills which makes sense right those people doing an MBA usually come to us for the screening to join the program after they've done a few months of the MBA so the way start thinking in business terms but undergraduates haven't started doing it so the burgesses structure isn't that clear for them and then finally the areas where canvas typically would fail this case is data overload they would die with the data they would struggle to extract the key information that's embedded in text in the data there's a lot of important relationships that's embedded in the text and they would not be able to either extract it if they did be able even if they were able to extract that they wouldn't know what to do with it second the structure wouldn't be obvious I think most candidates would build a revenue cost tree and they were tried to solve them separately they would if they even got that far they probably wouldn't have a key questions they wouldn't know what they were chasing and finally I think my personal feeling from what I've seen is the communication would kill people the ones who do this case very well the ones are very comfortable having a discussion with us they see the interviewer a partner level interviewer as a peer and they able to have a discussion if this case came up in the final round with the partner the math wouldn't be so heavy would be my guess because as a partner I would lead this by analyzing the conceptual thinking that said I've seen what in the final rounds and Harvard final rounds where a partner comes up and it gives you a lot of analysis as well so it really depends on the partner you face right I think the most important thing you want to do here is go through the case understand the underlying logic if you cannot solve the cases on the problem no one's going to be able to solve this case in its entirety but you should be able to identify the areas of analysis and with prompting and guidelines be able to move the case forward and if you can do that you'll be very very successful and suppose in closing I wanted to point out that you know this is one of the more interesting cases we do pharmaceutical cases tend to be very interesting as the competitive strategy cases and as always you know if our clients have any questions you're welcome to simply contact us arrange a phone call or if you would like we could do the case with you or you could do a similar case with you you know there are similar cases in the private equity industry that mirror some of these attributes or even in the electricity industry as well

Channel: firmsconsulting

Views: 173,630

Rating: 4.902555 out of 5

Keywords: Case interview, Bain & Company (Organization), Boston Consulting Group (Organization), McKinsey, profitability case, pharma

Id: pF1HTYg8Vhg

Length: 58min 38sec (3518 seconds)

Published: Mon Dec 02 2013