Management Strategy

SWOT analysis organizes matters into strengths, weaknesses, opportunities, and threats. Of these, strengths and weaknesses refer to the company's internal environment, such as technical capability and human resources, while opportunities and threats refer to the external environment, such as market growth, legal changes, and competitor entry. Therefore the combination consisting only of external factors is opportunities and threats; any combination including strengths or weaknesses mixes in the internal environment and is wrong.

3C analysis examines the business environment from the three perspectives of Customer, meaning the customers and the market, Competitor, and Company, meaning one's own firm. Cost, distribution, and advertising are elements handled in the marketing mix, the 4Ps or 4Cs, and are not components of the 3Cs. The 3Cs are characterized by matching up whom to sell to, whom to compete with, and what the company itself can do.

PPM divides products into four categories along the two axes of market growth rate and market share. Low growth with high share is the cash cow, which needs little additional investment and generates funds. The star has both high growth and high share and still needs investment, the problem child has high growth but low share, and the dog is low on both and a candidate for withdrawal.

The 4Ps are a framework for assembling a selling approach from the seller's viewpoint, consisting of Product, Price, Place, and Promotion. Plan, Do, Check, Act is the PDCA cycle, representing a procedure for operational improvement, not a marketing framework. Combinations including Profit or Physical are not components of the 4Ps.

The 4Cs restate the 4Ps from the buyer's viewpoint: Product corresponds to Customer Value, Price to Cost, Place to Convenience, and Promotion to Communication. Price is the amount the seller sets, but seen from the buyer it is the expense borne to purchase and use, so Cost corresponds. The other three correspond to Product, Place, and Promotion respectively.

Core competence refers to a company's central strengths that others cannot easily imitate, such as proprietary manufacturing technology or a customer base built over many years, and concentrating management resources there is considered the basic strategy. Fixed assets are an accounting category, an internal audit department is a matter of internal control, and complaint management is a matter of CRM and complaint handling; none of these is the definition of core competence.

The growth matrix combines markets and products, each as existing or new. Launching new products into an existing market is new product development. Market penetration is raising share in an existing market with existing products, market development is expanding existing products into new markets, and diversification is the most challenging strategy of launching new products into new markets.

Value chain analysis breaks a business into activities such as inbound logistics, manufacturing, outbound logistics, sales, and service, clarifies which activities create value and where costs arise, and identifies the activities to strengthen. Share price analysis is financial evaluation, totaling working hours is labor management, and purchase prediction is a matter of CRM and data analysis; each has a different purpose.

A strategy of concentrating resources on a specific narrow market, a niche, is called a niche strategy. A cost leadership strategy wins on low prices by exploiting scale and suits resource-rich companies. A diversification strategy is expansion by launching new products into new markets, and a homogenization strategy is one in which a leading company imitates and neutralizes a lower-ranked company's differentiation; neither is concentration on a narrow market.

M&A stands for mergers and acquisitions, a means of expanding the business by acquiring technology, people, and sales channels in a short time through mergers and acquisitions. Programs for employees to acquire company shares are employee stock ownership plans and stock options, outsourcing all manufacturing is the fabless model, and jointly operating hubs is an alliance; none of these describes M&A.

OEM refers to manufacturing under the client's brand: the contracted party manufactures, and the products are sold bearing the ordering party's brand name. The ordering party can expand its lineup without capital investment, and the contracted party can raise the utilization of its production facilities. Parts standardization is one cost-reduction technique, publishing drawings is a matter of openness, and moving production overseas is offshoring; none of these is the definition of OEM.

An alliance is a form of cooperation in which, unlike mergers and acquisitions, each company keeps its independence; examples include joint development and sales partnerships. Due diligence is investigating the actual state of a target company before an M&A, cannibalization is the phenomenon of a company's own products competing for the same customers, and restructuring is reorganizing the business; none of these denotes a cooperative relationship between companies.

BPR is an effort that does not take the current way of working for granted but questions the very necessity of each task and radically rebuilds workflows and the organization. Preparing spares is an availability measure, measuring working hours is labor management, and importing another company's procedures is imitation; none of them means fundamental redesign of work. It is also distinguished from incremental improvement activities.

BPO stands for business process outsourcing, a form in which business processes are entrusted as a whole, on an ongoing basis, to specialized external companies. The company can concentrate on core operations and also draw on the contractor's know-how. Visualizing workflows is business analysis, moving to the cloud is a matter of system infrastructure, and allocating targets is management by objectives; each is a concept separate from outsourcing.

CRM is customer relationship management: it centralizes customers' attributes and their purchase and inquiry histories, tailors responses to each individual, and aims to keep customers doing business over the long term. Optimizing the flow from procurement to sales is SCM, centralized management of core business data is ERP, and supporting sales activities is the purpose of SFA; their scopes differ.

SCM is supply chain management: it manages the flow of goods and information from procurement through production, logistics, and sales as a whole, including business partners, aiming to reduce inventory and shorten delivery times. Individual proposals based on purchase history are CRM, a package integrating core operations is ERP, and a system specialized in sales support is SFA; each manages a different domain.

ERP is enterprise resource planning: its aim is to integrate the core business data previously scattered across departments so that people, goods, and money can be allocated from a company-wide optimal viewpoint. Encrypting communications is a security measure, access analysis is Web marketing, and three-dimensional drawing is a CAD function; none of these is the main purpose of introducing ERP.

SFA is sales force automation: it shares deal progress, visit histories, and planned next actions across the organization to raise the quality and efficiency of sales activities. Attendance management is a personnel system, automatic control of production equipment is FA, and producing financial statements is an accounting system's role. SFA is sometimes positioned as part of CRM, which covers customer relationships in general, but its distinctive focus is on sales activities.

Knowledge management is the effort to turn the hard-to-verbalize know-how held by individuals, tacit knowledge, into visible forms such as case studies and manuals, explicit knowledge, and to share and use it across the organization. Subsidizing certifications is a personnel development program, restricting access rights is an information security measure, and low-inventory production is the just-in-time method; none of these aims at sharing and using knowledge.

KGI stands for key goal indicator, an indicator that quantifies the final goal, such as reaching 100 million yen in annual sales. What identifies the key factors is the CSF, the critical success factor, and what measures interim progress is the KPI, the key performance indicator. It is not a term for a deliberative body. Sorting them as goal, winning factor, and interim indicator helps avoid confusion.

A KPI is an interim indicator measuring whether progress toward the KGI is on track, so a figure like monthly visitor count, which can be measured regularly and ties to sales, is appropriate. Annual sales of 100 million yen is the KGI itself, not a KPI. The number of philosophy articles and the head office location do not express business progress and are meaningless as indicators.

The balanced scorecard is a technique that sets goals and indicators from the four perspectives of financial, customer, business process, and learning and growth, evaluating performance without bias. Strengths, weaknesses, opportunities, and threats are the perspectives of SWOT analysis, quality, cost, and delivery are elements emphasized in production management, and finance, legal, personnel, and public relations is a mere list of departments; none of these is the BSC's set of perspectives.

TQM is a management approach that makes quality the responsibility not of manufacturing alone but of the whole company, including sales and administration, improving continuously. Full inspection is merely one way of checking quality, and a system where only the quality assurance department sets standards actually runs against TQM's philosophy. Certification to the international standard is third-party attestation of a quality management system, not TQM itself.

Process innovation is innovation that changes manufacturing processes or ways of working to greatly raise productivity or quality. Developing a product with new functions, launching one using new materials, and renewing the external design all change the value of the product itself and count as product innovation. Distinguish by whether what is made changes or how it is made changes.

There are three walls before a technology becomes a business, and the wall at the stage of moving from development to commercialization is called the valley of death. The devil river is the wall between basic research and development toward productization, and the Darwinian sea is the wall after commercialization, surviving competition until the product takes root as an industry. The chasm is the gap when a new product spreads from the early market to the mainstream market, a concept of the diffusion process.

The chasm refers to the deep gap that stands in the way when a product spreads from the early market, centered on novelty-seeking buyers, to the mainstream market, centered on ordinary consumers. The barrier from research to development is the devil river, and post-commercialization competition is the Darwinian sea; both are walls in the process of technology becoming business but are distinct from the chasm. It has nothing to do with patent examination periods.

A technology roadmap is a plan charting along a time axis which technologies are to be developed to what level and by when, used so that the parties involved share the outlook for development. A skills list is skill management, a work instruction document is a manufacturing directive, and a patent list is an intellectual property management document; none of these shows the path of future technology development.

The API economy refers to the economic sphere that grows as companies publish their services' functions as APIs and other companies combine them to create new services. Services embedding map or payment functions are typical examples. Cost recovery methods, standardizing internal data formats, and execution speed benchmarks all have nothing to do with the idea of creating value through inter-company functional integration.

The lean startup is a way of developing a business that quickly releases a prototype with minimal functions, learns from customer reactions, and repeats improvement or pivots, avoiding the waste of overbuilding. Launching a finished product all at once is the exact opposite, and cost reduction through mass production or imitating competitors differs from the idea of repeatedly testing hypotheses.

Electronic commerce is classified by the combination of transacting parties. This is a transaction between individuals, consumers, so it is CtoC. BtoB is between companies, BtoC is sales from companies to consumers, and BtoE is provision from a company to its own employees. The flea market app operator merely provides the marketplace; the parties to the sale are the individuals.

The long tail is the phenomenon in which, by broadening the assortment, the combined sales of rarely selling items contribute substantially to overall sales; it is pronounced in online retail, where shelf constraints are small. Narrowing to best-sellers is the opposite idea, price variation with demand is dynamic pricing, and referral benefits are referral programs; each is a different concept.

Affiliate marketing is performance-based advertising: when a purchase or application is completed via the hosting site, a reward is paid to the site operator. Aiming for high search rankings is SEO, serving ads based on browsing history is targeted advertising, and member-referral discounts are referral programs; none of these is a mechanism of advertising rewards tied to results.

Omnichannel is the effort to integrate sales channels such as physical stores, online shops, and apps together with inventory and customer information, so that customers can buy without noticing the difference between channels, for example ordering online and picking up in store. Bundling communication lines is a networking technique, splitting price tiers is a pricing strategy, and joint logistics is inter-company collaboration; none of these means integrating sales channels.

The sharing economy is a mechanism in which individuals lend, borrow, and share underused assets they own, such as homes, cars, and skills, with others through online intermediary platforms, and renting out a spare room is a typical example. Discount sales are sales promotion, mergers are M&A, and shareholding associations are employee asset-building programs; none of these involves sharing idle assets.

Crowdfunding is a mechanism of publishing a plan on the Internet and collecting small amounts of funding from many unspecified supporters; types include reward-based, where backers receive products, donation-based, and investment-based with financial returns. The name derives from crowd, and it is a different word from cloud services. It has nothing to do with load balancing, pay-per-use billing, or public loan programs.

Fintech is a coinage combining finance and technology, referring to new financial services that use IT. Code payments and transfer apps, automatic household account books, and AI-driven investment advice are examples. Automatic control of production lines is FA, digitizing documents is document management, and growth monitoring is smart agriculture; none of these is a financial service.

The kanban method has each upstream process make only as much as the downstream process has used, signaled by kanban cards, thereby producing what is needed, when needed, in the quantity needed, and cutting intermediate inventory and overproduction waste. Assembly by one or a few workers is the cell production method, calculating material requirements is MRP, and creating drawings is CAD; each is a separate mechanism.

BYOD stands for Bring Your Own Device, using employees' privately owned PCs and smartphones for work. It saves device costs, but loss and private use create risks of information leakage, so usage rules and technical measures are needed. Devices lent by the company are the opposite arrangement, working outside the office is telework, and a prohibition on taking data out is one kind of security rule.

Level 3 is called conditional driving automation: the system handles all driving operations under limited conditions, but the driver must respond to the system's requests. Level 1 assists a single operation such as accelerating or steering, and level 2 assists a combination of operations; in both, the human remains the primary driver. Level 4 is the stage where, under limited conditions, no response from a driver is required.

RPA is a mechanism in which software robots take over screen operations and the like, and it suits automating clerical work with fixed procedures, such as transferring and matching data between systems. Welding robots and reception robots are physical machines, not the clerical automation RPA covers. Demand forecasting from data is AI analysis, whose purpose differs from standing in for routine work.

A digital twin is a technology that gathers the state of real equipment and products with sensors and builds a twin in virtual space, enabling failure prediction and process improvement trials without stopping the real thing. Duplicating data is backup, two-person development is pair programming, and an always-running standby is hot standby; none of these is simulated experimentation through reproduction of reality.

A deepfake is video or audio produced by generative AI that makes a real person appear to say things they never said, and it risks being misused for impersonation scams and spreading disinformation. Inflating training data is data augmentation, human verification of output is an operational safeguard, and outputting memorized training data is a separate problem of memorization and overfitting; none of these refers to generating fake video and audio.

Generative AI creates new text, images, audio, programs, and more in response to user instructions, based on what it has learned from large amounts of data. A mechanism that merely finds and returns prepared answers is a conventional search or FAQ system, encrypting data is a security technology, and taking over routine clerical work is RPA; none of these generates new content.

The phenomenon of generative AI outputting untrue content plausibly is called hallucination. It can fabricate nonexistent sources and figures, so outputs should be fact-checked by humans before use. A deepfake is fake video or audio made with AI, overfitting is fitting the training data so closely that performance on unseen data suffers, and a filter bubble is the state of being surrounded only by information matching one's tastes; each is a different phenomenon.

Prompt engineering is the practice of crafting the instructions given to generative AI, such as concretely specifying roles, conditions, output formats, and examples, to obtain output close to the goal. Its hallmark is refining how the AI is used rather than rebuilding the AI itself; designing learning algorithms is developer-side work, server upgrades are an infrastructure matter, and fee negotiation is a procurement matter, so none of these applies.

Information entered into an external service may be transmitted to and stored by the provider, which could lead to leaks of confidential or personal information, so the range of information that may be entered needs to be defined in internal rules. Generative AI output can be factually wrong, so checking is essential, and the handling of input and the rights to generated content are set by each service's terms, so reviewing them is indispensable.

Generative AI output can end up resembling existing works, and publishing it as is risks infringing copyrights, trademark rights, and the like, so a person must check similarity and rights before publication. Being AI-made grants no exemption from rights checks, and crediting a source alone does not make anything free to use. The handling of generated content is also set in the terms of service, so checking them is necessary.

The Social Principles of Human-Centric AI call for respect for human dignity and for ensuring fairness, accountability, and transparency; the key point is not leaving everything to AI but keeping final judgment and responsibility with humans. AI judgments are not always right, and biased training data produces biased results. Transparency of use and accountability are also required by the principles, so the idea that no explanation is needed is wrong.

Using a cloud vendor's AI functions has the advantage of quick adoption without preparing large training datasets, computing resources, or specialist staff in-house. On the other hand, freely modifying the model's internals is normally impossible, and since it presupposes communication, it cannot be used when the line is down. Fees generally accrue according to usage, so it is not necessarily free.

Because AI learns from past data, bias in that data risks repeatedly disadvantaging applicants of a particular gender or age. In uses affecting people's rights, the training data and results must be verified continuously, the grounds for decisions must be explainable, and the final decision should be made by a person. Hiding the fact of AI use is inappropriate from the standpoint of transparency.