Management 2nd Part PDF

STRATEGY WHAT IS STRATEGY? Michael Porter, professor at Harvard Business School, explains how firms competes in a market and when you have to take decisions on how to compete you have to look at your firm comparing it to your competitors’ average There are two different way...

STRATEGY WHAT IS STRATEGY? Michael Porter, professor at Harvard Business School, explains how firms competes in a market and when you have to take decisions on how to compete you have to look at your firm comparing it to your competitors’ average There are two different ways of thinking of competing: - competing to be the best: run the same race of other firms, making similar products and try to make them better than others’ - competing to be unique: run a different race, differentiate your product from other firms’ ones We define strategy as the set of goal-directed actions a firm takes to gain and sustain superior performance relative to competitors (=competitive advantage) Good strategy consists of three elements (that reflects the three parts of the AFI framework): - diagnosis (Analysis): the ability of identify the competitive challenge by analyzing the firm’s external and internal environment - guiding policies (Formulation): the ability of top managers of addressing the competitive challenge deciding how to answer it by formulation of firm’s corporate, business and functional strategies - coherent actions(Implementation): the ability of lower managers and employees to implement firm’s guiding policies through actions that implement the strategy chosen So to craft a good strategy. 3 main steps are needed: - define an organization’s competitive challenge through a critical and honest assessment of the status quo - provide a game plan for dealing with the competitive challenge identified and provide clear guidance for all employees - effectively implement this game plan through a coherent and consistent set of actions Tesla secret strategy In 2006, Elon Musk, co-founder of Tesla, explained the start-up’s master plan: - step 1: build sports car - step 2: use that money to build an affordable car - step 3: use that money to build an even more affordable car - step 4: while doing above, also provide zero-emission electric power generation options - step 5: don’t tell anyone How they implemented it: How Tesla developed its strategy: Tesla’s mission: “Accelerate the world’s transition to sustainable energy” Then 1. diagnosis of the competitive challenge: Tesla’s goal is to build zero-emission electric vehicles that are attractive and affordable. So its competitive challenge is the it needs infrastructure for electric vehicles and network of charging stations, while keeping all of this as qualitative and low-cost as they can 2. guiding policy: Tesla’s current guiding policy is “to build a cost-competitive mass-market vehicle such as Model 3”. In order to follow it, Tesla backed with strategic commitments such as investment in lithium-ion battery in Nevada, in order to be self-suppliant for battery demand and to reduce costs 3. coherent actions: Tesla actions are aimed to reduce costs and to increase production volumes. These actions are for example retool existing facilities with cutting-edge robotics or secure uninterrupted supply of lithium-ion batteries thanks to their gigafactory What strategy is not Strategy is not grandiose statements or goals (such as “we want to be the best firm in the market” or “we want to make a better world”) Strategy is not failure to face a competitive challenge Strategy is not operational, effectiveness, competitive benchmarking or other tactical tools COMPETITIVE ADVANTAGE We define competitive advantage as superior performance relative to other competitors in the same industry or to the industry average. Competitive advantage is always relative and never absolut, you have to compare the firm’s performances with a benchmark If a firm is able to outperform competitors over a prolonged period it has a sustainable competitive advantage If a firm underperforms its competitors it has a competitive disadvantage, while if two firms have similar performances we call it competitive parity  Firms gain competitive advantage by combining value and cost through strategic positioning by: - delivering superior value while containing the cost to create it, that is creating a product that offers a higher value than competitors’ (differentiation) - offering similar value at lower cost, that is offering a product similar to competitors’ but priced less (cost leadership) The key to a successful strategy is to combine activities to stake out a unique strategic position in an industry. This includes trade-offs: while deciding what to do you are also deciding what not to do and while focusing on specific market segments you are losing all other possible customers. Trying to be everything to everybody will probably result in inferior performances Example: Walmart vs Nordstrom Walmart: low cost retailer. It offers an acceptable level of service by low-skilled employees, it is a big-box retail outlet and offers everyday low prices. You don’t go there expecting an high-level service or high quality products, you go there because of the low prices and the big variety offered Nordstrom: upscale retailer. It offers a superior customer experience by professional sales people and it has a luxury setting. You go there because of the experience that you’ll have and the quality and professionalism they offer, accepting the compromise for higher prices These two companies are important retailers in the US market, they compete in the same industry but they are not direct competitors. Indeed they offer different values at different prices and they are differentiated. For this reasons, their customers have little overlap since people buying from Walmart are different from people buying from Nordstrom Nevertheless, Walmart has a clear strategic profile in order to serve a specific market segment. They perform activities differently than rivals and they reinforce one another, for example they have big retail stores in rural locations where their customers are mostly set, they have high bargaining power due to their size and profitability, they have a sophisticated IT system and low base salaries in order to have the lower prices possible Value creation A good strategy should create value for shareholders and for all other stakeholders We define value creation when companies with a good strategy are able to provide products or services to consumers at a price point that they can afford while keeping their costs in check, thus making a profit at the same time. Both parties benefit from this trade as each captures a part of the value created STUDY CASE: SOUTHWEST AIRLINES Southwest Airlines is an American airline company which has been making a profit for the last 43 years, even if analysts say that it could make more profit if it made some changes in its policies. Anyway, Southwest has always refused to adopt usual competitors’ strategy and maintained its uniqueness Southwest is different from others airline companies for many reasons:  - no hubs: Southwest has never charged fees for additional bags or for any additional service, remaining the airline company with the lowest fee policy in the world - no hub: instead of adopting the hub-and-spoke model other companies adopt, Southwest prefers a point-to-point model which allows to lower costs - prioritizing employees: the company has a policy that prioritizes employees - prioritizing customers: company’s decisions aim to make customers better off, having them as first priority even if this means losing possible profits - single aircraft model: the company only operates with a single aircraft model to reduce costs of training and maintenance Despite its unique strategy, Southwest has been making a profit for the last 43 years, being one of the most successful and consistent company in the world One of the main points of discussion between analysts and the company’s CEO during a period of crisis in which revenues lowered a lot was about the firm introducing some service fees. The analysts said this would have softened the effects of the crisis increasing profits, while the company wanted to stick to its traditional policy. The whole discussion can be summarized as To make a decision, it is always important to collect data about what you are analyzing and to conduct experiments to see how customers would react to some possible changes. In this case, the company could have introduced some fees on some “experiment” flights and see how the customer reacted. What could have been the consequences?  Conducting these experiments, the firm could have tested if these changes impacted customers behaviour and how its profit could change thanks to these innovations. After having seen this, it could have decided which strategy to follow DECISION ANALYSIS Why do we need a scientific approach to decision-making? When a manager makes decisions, he can always be influenced by his personal perceptions, his beliefs and an exaggerated self-confidence on his past experiences The scientific approach is needed in order to avoid the fact that decisions are influenced by personal biases Rely just on personal experiences and opinions or on hopes and ideas can lead to hugely misjudge some situations An example is given by FreeNow, a car-sharing company on whom BMW and Daimler invested a lot thinking this business could have substituted car ownership in big cities where customers may prefer to rent a car when they need it instead of buying one. The project ended up mostly failing and is now alive just in a few European cities: the wrong decisions made by those managers was caused to the limited amount of information they had and on the fact that they were under pressure while they were taking them because they had to make quick decisions The main point is that in decision-making nothing is certain, there is nothing that you can decide to do knowing exactly which consequences it will have. Scientific decision-making is based on probability, that underlines uncertainty: you decide what to do basing on how probable is that every possibility gives you an outcome you want, but without being completely sure To have a scientific approach to decision-making is about learning how to make predictions using probability and facts following three main goals: - be less certain: stop trusting our overconfidence on our abilities or sources of information. This overconfidence depends on culture, personality, expertise… - learn to make predictions: is about questioning how often an outcome typically happens - think probabilistically: is about having a theory to envision probability of events Basics on probabilities and distributions A random variable is a variable that can take on some values and each value is assigned a probability. Random variables can be: - continuous: variables that can take on a value in a range - qualitative/categorical: variables that do not come from measuring or counting - dichotomous/binary: variables that can take on only two values We define probabilities: - probabilities indicate the likelihood of the occurrence of a random event - probabilities are a sequence of numbers in [0,1] that have to sum up to 1 A random variable X has a certain probability of taking on a certain value, and that probability can be assigned before conducting the experiment Uniform distribution: when each value has the same probability, meaning that X has an equal likelihood of being any of the possible outcomes (e.g.the roll of a dice). Taking decisions based on uniformly distributed probabilities is not effective since you can’t in anyway predict which is the most likely outcome Normal distribution: in a normal distribution, the outcome tend to distribute around a central value (e.g. people’s height: there are less people that are very short or very short, most of them are average) A normal distribution can be left skewed when the values of the observation cluster more around the right side of the distribution. In the same way, it can be right skewed if the values cluster more around the left side of the distribution Mean and standard deviation probability distribution have summary statistics or moments, that are used to describe data or the sample analyzed. The two most important statistics are mean and standard deviation Mean tells you where the center of the distribution is positioned. It is used to have an 𝑛 idea of how the sample is set. The formula to compute the mean is µ = ∑ 𝑝𝑖𝑥𝑖, 𝑖=1 where x is every possible value in an observation and p is the probability of whether that value occurs The variance, and then the standard deviation, tells you something about how far from the mean values are distributed: if they are concentrated around it or dispersed 𝑛 2 2 away from it. The formula to compute the variance is σ = ∑ 𝑝𝑖(𝑥𝑖 − µ) , where x is 𝑖=1 every possible value the variable can take, p is the probability of whether that value occurs and µ is the mean. The difference between x and the mean is squared to avoid problems with negativity or positivity of the difference. The standard deviation reflects this value removing the square, to give a value more aligned with the value  that the variables can take on. The formula to compute standard deviation is 2 σ = 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = σ Conditional probability and joint probability = P(X|Y) Conditional probability is the probability of an event (situation, characteristic, condition,...) occurring given that another event has already taken place Conditional probability is useful in many situations to predict how an event should go knowing other events that already happened and can influence it Example: This frequency table reports the number of letters of the last name of some students, dividing them based on whether they originate from Italy, Asia or somewhere else. In this case our random variable X is the number of letter in the last name This type of table is called a “contingency table” because it shows the values of X contingent to some other variables The totals of rows and columns are marginal frequencies (that are the last row and column) In this case we may ask “what is the probability of a student having 5 letters in his last name given that he is italia? We can compute P(X=5|Italy)=9/63=14.3% The conditioning event is often called a condition, clue, data point or signal (in this case being italian) The conditional expectation is the expected value of a random variable, given that another event has already taken place. The formula to compute conditional 𝑛 expectation is 𝐸(𝑋|𝑌) = ∑ 𝑃(𝑥𝑖|𝑌)𝑥𝑖, where x is any possible value the variable can 𝑖=1 take on and P(x|Y) is the conditional probability contingent to Y. Multiplying every conditional probability by its respective value one can compute the conditional expectation, that is like the mean but restricted by the event that has already happened. For example, in the previous table, E(X|Italy)=7, that is the number of letters the last name will have on average given that the student is italian Joint probability is the probability that two events take place together at the same time. It can be computed by P(X,Y)=P(X|Y)*P(Y) that is the probability that Y happens times the probability that X happens once we know that Y happened. In the previous chart, P(X=%, Italy)=P(X=%|Italy)*P(Italy)=9.7%, that is the probability that the student is italian and the last name has five words It is possible that an event that happened does not help to predict another event because it doesn’t condition it. In this case we say that these two events are independent and, in this case, the signal is not informative. Computing the conditional probability is not useful in this case because if X and Y are independent P(X|Y)=P(X) Therefore when two events are independent the joint probability is P(X,Y)=P(X|Y)*P(Y)=P(X)*P(Y). For example, rolling a dice multiple times is a set of  independent event, since every roll doesn’t influence the followings, then the probability of rolling two 6’s is P(6,6)=P(6)*P(6)=2,78% Even if two events are not perfectly independent, a signal can be more informative or less. Recalling the example of the number of letters in the last name we can compute that P(X=%|Italy)=14,3% and P(X=5)=14.0% so in this case being Italian is not very informative. On the other hand P(X=3|Asia)=41,3% while P(X=3)=6,4% so in this case being Asian is very informative Prediction of events Probability is important to managers in order to predict events and decide actions based on these predictions Example Your company has recently bought a new robot for the factory, but it is giving some problems and doesn’t work on some days. You come up with an hypothesis saying that the robot fails when the temperature in the factory is low, and you collect some data to verify this The theory we want to come up with has to be an if-then statement, meaning that whenever there is low temperature the robot fails Analyzing P(low temperature|accident)=60% the two events don’t seem that correlated because accidents occur even when temperature is high On the other hand, analyzing P(accident|low temperature)=100% we can tell that the low temperature always cause an accident and we can build an if-then theory: “if the temperature is low, then the robot fails” and this is always true Therefore we define a theory as a series of logical steps linking antecedents to consequences: if P then Q, if Q then R, if R then S, …, if V then Z… and you eventually show that if P then Z Theories are important because they tell you what data and factors to examine: the factors that affect your outcome improve your ability to make accurate predictions because they provide signals for your conditional probabilities. Practically you follow some steps: 1. you theorize something like “if A, then B” 2. you collect all the cases with A and you measure B in them 3. you compute P(B|A) and it has to be near 100% 4. if needed, you repeat with a theory about A and so on Some essential steps to predict whether a target event will occur: 1. you need a theory: you have to decide which factors to use to build your theory, searching for signals you think can condition the happening of the event. Based on this first theory, you envision a certain probability that the event happens: this is called prior probability, that is the predicted probability of an event before data are collected. It is the best assessment of the probability of an outcome based on the current knowledge 2. revise your theory: you now look for more information and signals, even if this is usually costly and you can’t collect much signals. There are at least two ways of collecting signals: - ask others: you can ask others, for example experts, for their opinion in order to update your probability  - collect data: collecting data about the event you want to predict and conducting a data analysis you can update your probability After having updated you prior probability, you end up with a posterior probability, that is the revised probability of an event occurring after having taken new information into account Technically: if you had a prior probability P(Y) and you collect signals that tell you S1 then you update your posterior probability at P(Y|S1), and you can repeat it several times with several signals 3. define your expectations: you decide when to stop collecting data and signals and to finalize your probability, that is a posterior probability that you cannot improve further. The you can make decisions based on this probability Making decisions We need to predict events because an event is associated with the outcome of an action and then we want to decide whether or not to take that action Example Let’s say that you can choose to pay 1000€ upfront and then you receive 100€ per student who supports Inter Milan in the class. Therefore if at least 10 students support Inter you have a profit, otherwise you’re gonna end up losing money The situation you are facing then is: The set of actions is the set of options that are available to you and you decide which one to take. On the other hand we have scenarios, that are the states of the world you are going to face, what could happen and that you don’t control: this is what you can predict. The outcomes are all the possible combination of action-scenario, this is what will happen based on what you choose and which scenario events How to proceed? In this case, intuitively, the higher is the probability that at least 10 students support Inter the more likely is that you get a profit. This is why you have to predict and estimate the probability of different outcomes Once you have done your prediction, you have come up with a posterior probability and now have to decide whether or not to accept the deal. How do you decide based on your prediction, that you don’t know to be surely true? You need a criterion There are two main way of establishing a criterion: thinking of a threshold probability or assigning values to profits and losses threshold: you think of a threshold P* such that if your expected probability P(Y|S)≥P* then you go ahead with the action, while if P(Y|S) 𝐶𝑚𝑎𝑟𝑘𝑒𝑡 the solution should be to contract with market players Both options have advantages and disadvantages: There are also alternatives along the make-or-buy continuum:  Short-term contracts: - describes contracts to be awarded with a short duration, generally less than one year - when engaging in short-term contracting, a firm sends out requests for proposals to several companies, which initiates competitive bidding for contracts - this allows a somewhat longer planning period than individual transactions and the buying firm can often demand lower prices due to the competitive bidding process Strategic alliances: strategic alliances are voluntary arrangements between firms that involve the sharing of knowledge, resources and capabilities with the intent of developing processes, products or services. They can be: - long-term contracts: longer than a year and helps facilitate transaction-specific investments - equity alliances: partnership in which at least one partner takes partial ownership in the other partner - joint venture: two or more partners create and jointly own a new organization Parent-subsidiary relationship: - describes the most integrated alternative to performing an activity within one’s own firm boundaries - the corporate parent owns the subsidiary and can direct it via command and control Vertical integration along the value chain We define vertical integration as the firm’s ownership of its production of needed inputs or of the channels by which it distributes its outputs The industry value chain depicts transformation of raw materials into finished goods and services, and each vertical stage represents a distinct industry  A firm can vertically integrate in two directions: - backward integration: consists in moving ownership of activities upstream to the originating inputs of the value chain - forward integration: consists in moving ownership of activities closer to the end customer Vertical integration has both benefits and risks. Benefits: - lowering costs - improving quality - facilitating scheduling and planning - facilitating investments in specialized assets - securing critical supplies and distribution channels Risks: - increasing costs - reducing quality - reducing flexibility - increasing the potential for legal repercussions When does it make sense to vertically integrate? - The main reason to vertically integrate is the failure of vertical markets, when you can’t find what you need in any suppliers or retailers - Vertical market failure occurs when transaction within the industry value chain are too risky or when alternatives to integration are too costly Corporate diversification We define diversification as an increase in the variety of products and services a firm offers or markets and the geographic regions in which it competes There are three main kinds of diversification: - product diversification: when a firm is active in several different product markets - geographic diversification: when a firm is active in several different countries  - product-market diversification: when a company pursues both a product and a geographic diversification There can be different levels of diversification based on the percentage of revenue and relationship of core competencies across business units: - single business: when >95% of revenues derive from one business and we have a low level of diversification - dominant business: when between 70-95% of revenues come from a single business and the firm pursues at least one other business activity - related diversification: when