Self Help

Data Strategy - Marr, Bernard;

Author Photo

Matheus Puppe

· 55 min read

“If you liked the book, you can purchase it using the links in the description below. By buying through these links, you contribute to the blog without paying any extra, as we receive a small commission. This helps us bring more quality content to you!”



  • Data, artificial intelligence (AI), and the Internet of Things (IoT) are growing at an astonishing rate and transforming businesses and society. We are generating more data than ever before and AI capabilities are rapidly advancing.

  • This data-driven world is enabling companies to gain valuable insights about their customers, create more intelligent products and services, improve business processes, and make smarter decisions. Data and AI have the potential to solve major global issues as well.

  • AI is not yet at the level of “true” or general artificial intelligence that can match human abilities, but neural networks and deep learning are advancing AI capabilities.

  • The rise of data, AI, IoT and other technologies is fueling the Fourth Industrial Revolution or Industry 4.0, marking a new era of automation and data exchange in manufacturing and business.

  • Other major technologies like quantum computing, 5G, augmented/virtual reality, and advanced robotics are also evolving rapidly.

  • Virtually every business must embrace data and become a “data business” to remain competitive. Data provides valuable business insights and differentiation.

  • Companies need solid data strategies to capitalize on these opportunities. This involves identifying use cases, governing data ethically, developing analytical capabilities, building technology infrastructure, and creating a data-driven culture.

In summary, we are entering a new data-centric era that will transform business and society. Companies must leverage data and AI strategically to succeed and drive innovation. A data strategy is essential for competing in the 21st century.

  • The amount of data being created is growing exponentially - 175 zettabytes are expected by 2025. This is more data than has ever existed before.

  • Much of this data growth comes from unstructured sources like photos, videos, text, etc. This unstructured data has huge potential value if analyzed properly.

  • Big companies like Google, Facebook, retailers, etc already collect vast amounts of data about individuals and use it to target advertising, make predictions about preferences, relationships, intelligence, etc.

  • Data analytics now influences many aspects of life beyond just advertising - elections, healthcare, space exploration, etc.

  • The Obama 2012 campaign made highly effective use of data analytics to micro-target messages and influence voters. This has become common practice.

  • There are exciting possibilities but also concerns around privacy and data being used to manipulate or exert control over populations.

  • Overall, data is becoming a critical strategic asset for all organizations. Those that can utilize it effectively will have big advantages. But the term “big data” is simplistic - it’s about having the right data and using it successfully.

  • Data is being generated at an exponential rate from many sources like smartphones, social media, Internet of Things devices, etc. The amount of data created globally is expected to grow massively in the coming years.

  • Advances in artificial intelligence and machine learning are enabling us to analyze and draw insights from this huge amount of data like never before. AI and ML tools can process data and make predictions and recommendations much faster than humans.

  • Data analytics played a big role in major events like Donald Trump’s presidential campaign, Brexit referendum, and is being used extensively in fields like healthcare. During COVID-19, data and analytics have been critical to tracking the pandemic and developing treatments.

  • Technologies like machine learning allow computers to improve themselves and learn from data without explicit human programming. While true artificial general intelligence may still be far off, AI is already transforming many industries and our daily lives through applications like voice assistants.

  • The rise of the Internet of Things with billions of connected devices, wearables, self-driving cars etc. is generating massive new streams of data that can be utilized by AI.

  • We may be nearing an era where computers can not just analyze data but also understand and respond to human emotions and needs. The possibilities for using data and AI to transform business and society are immense.

  • Analyzing facial expressions, posture, gestures, voice, speech patterns, and typing rhythms can detect changes in a user’s emotional state. This technology has many potential applications such as computers recognizing user frustration, phones encouraging breaks during stress, and smart homes responding with soothing environments after a bad day.

  • Companies like Disney, BBC, Coca-Cola, Microsoft, and auto manufacturers are exploring these emotionally-aware technologies. We are nearing machines that can provide suitable emotional responses.

  • This explosion of data analytics capabilities marks the 4th industrial revolution or Industry 4.0. It combines automation, robotics, machine learning, and internet connectivity to create “smart factories.”

  • Benefits of Industry 4.0 include improved health/safety, supply chain control, consistent productivity, and increased revenues. But it also faces challenges like security, reliability, quality control, job losses, and hesitation to adopt new technologies.

  • Other key technologies shaping Industry 4.0 are genomics, blockchain, and extended reality. Combined, these advances in data, AI, automation, and connectivity are transforming business.

  • Data is becoming a key business asset for competitive advantage. Companies must become “data businesses” to understand customers, operations, markets, and products by leveraging data and analytics.

Here are the key points about using data to improve decision-making:

  • More data provides a better understanding of all factors affecting a business, leading to more informed decisions.

  • Analytics lets you analyze potential outcomes and be confident decisions are guided by data rather than gut feelings.

  • Humans make mistakes, so successful businesses use data to drive decisions where possible.

  • Data can augment human decision-making capabilities.

  • Automated decision-making takes this further by removing humans from some repetitive decisions.

  • Examples include using data to decide pricing, inventory levels, resource allocation, and more.

  • Overall, data-driven decisions tend to be more accurate and lead to better business outcomes.

Here is a summary of the key points about use cases for data:

  • Using AI and automation to improve decision-making - AI can analyze data and make automated decisions much faster than humans. This is happening across many industries like banking and HR.

  • Understanding customers and markets - Data analytics provides insights into customer wants, purchasing behavior, competition, etc. This informs business decisions in areas like product design and marketing.

  • Creating better products - Connected, smart products utilize data to provide more functionality and be more personalized. Examples include fitness trackers, smart home devices, etc.

  • Creating better services - Data allows companies to tailor and personalize services to customers’ specific needs and contexts, like Netflix recommendations.

  • Improving business processes - Data helps optimize operations in areas like marketing, logistics, R&D. AI and automation can handle repetitive admin tasks.

  • Creating revenue from data - Companies can monetize the data they collect by selling insights or analytics. An example is John Deere selling data from its farm equipment.

  • Telemedicine demonstrates use cases across all these categories, like using AI chatbots for triage, gathering patient data to improve services, developing connected health devices, optimizing operations, and monetizing data.

Here are a few key points on using data to improve business decisions:

  • Start by identifying your key business questions (KBQs) - the unanswered questions related to your core business goals and objectives. Focus on the top 10-20 most critical questions.

  • Good KBQs help you determine what data you need to collect and analyze to get meaningful insights. The right questions lead to better data and better decisions.

  • KBQs might include things like “Who are our most valuable customers?”, “What products have the highest profit margins?”, “Where are we losing customers and why?”

  • Data helps decision makers reduce uncertainty and make smarter, more informed choices based on facts vs guesses, beliefs, or assumptions.

  • Analyzing data to answer your KBQs can provide valuable insights into critical issues and help guide future strategy and decisions.

  • Data-driven decisions are proven to lead to better business outcomes when done right. But it starts with identifying the right questions aligned to your goals.

  • Examples could include a retailer using foot traffic data to decide where to open/close stores, or a spirits company reviewing sales and inventory data to optimize product assortment and pricing.

  • The key is to start small - focus on a few high-priority KBQs first. Let the data uncover insights that can drive incremental improvements. Then build from there.

  • Companies like Unilever and Google have used data analytics to gain insights into business challenges around supply chain shrinkage and manager effectiveness.

  • Asking the right questions is key to getting value from data. Questions should relate to overcoming specific business challenges.

  • Once data is collected, it needs to be analyzed and interpreted to extract insights. Dashboards are useful tools for visualizing and communicating insights.

  • There are two main types of dashboards:

  1. Curated dashboards - Created by experts to communicate insights to decision makers. Focused on key questions and metrics. Can create analytics bottlenecks.

  2. Self-service dashboards - Allow broader access to data exploration. Help build data literacy across the organization. Need proper ingredients/tools to be effective.

  • Best practices for curated dashboards include starting with key questions, keeping it simple, ensuring accessibility, and focusing on communicating insights effectively.

  • Self-service dashboards enable a “raclette grill” experience where everyone can engage with the data. This builds data literacy across the organization.

  • A self-service approach to analytics can lead to problems if not properly managed, like a free-for-all buffet situation. Guidance and training is needed to develop good data literacy.

  • Examples like Shell and Walmart show how creating internal data communities and translator roles can facilitate the democratization of data.

  • Data democratization involves spreading data skills across the workforce, not just to specialized teams. Translators can bridge the gap between data experts and business users.

  • Data storytelling is important to turn insights from dashboards into compelling narratives. Stories have a beginning (the problem), middle (the analysis process), and end (the insights and recommendations).

  • Overall, a more data-literate business is a more efficient and innovative business. Companies should invest in developing data skills for all workers to drive better decision making.

Here are a few key points to summarize the main ideas from this section on the future of data visualization and storytelling:

  • Virtual reality (VR) and augmented reality (AR) will enable new immersive ways of exploring and interacting with data in 3D environments. This can help us make connections and gain insights more naturally.

  • AI may play a role in keeping users focused on key business questions when interacting with large datasets in VR. VR can also help us visualize and understand complex AI systems.

  • Collaborative, multi-user VR environments will allow teams to explore data together.

  • Augmented analytics involves automating data analysis and reporting with no human intervention. This doesn’t replace data scientists but assists them.

  • Creative data visualization approaches like “data cuisine” (translating datasets into food) provide fun and engaging ways to present data.

  • Overall, emerging technologies will redefine how we experience and share insights from data, making it more immersive, intuitive, and collaborative. The key is finding ways to focus these technologies on answering important business questions.

Here is a summary of the key points about using data to understand customers:

  • Understanding customers is fundamental for businesses to meet their needs. Data and technology enables this today.

  • Market trend analytics examines broad trends, using sales data, economic indicators etc to see if markets for products are growing or declining. This informs business strategy and investment.

  • Analyzing niche groups and individuals allows personalized, tailored offerings and pricing. Customer loyalty programs evolved from simple metrics to provide personalized deals based on purchasing data.

  • Tech giants like Google and Facebook built power by aggregating user data to understand people’s interests and behaviors. Their platforms allow businesses easy customer segmentation and targeting.

  • However, these platforms can’t immediately tell you the right customer segments for your business. You need your own customer data strategy.

  • Collecting insightful customer data involves identifying valuable metrics and using tools like surveys, interviews, customer service logs and more. The aim is a 360 degree customer view.

  • Analytics converts this data into customer insights. This understanding of customers then informs business decisions at a strategic and tactical level.

  • Customer analytics involves using data to build an in-depth understanding of customers and potential customers. This goes beyond guesses and assumptions to provide real insights.

  • Customer data comes in three main forms: personal data, behavioral data, and attitudinal data.

  • There are two main sources of customer data: internal/first-party data collected by the business itself, and external/third-party data acquired from other sources.

  • Internal data offers proprietary insights into a business’s own customers, while external data provides information about the wider market and population.

  • Neither internal nor external data alone provides a full 360-degree view of customers. The most powerful analytics combine both data sources.

  • Internal data builds customer relationships but can lag behind changes in the market. External data enables prediction but may be available to competitors too.

  • The best customer analytics strategies use both internal and external data across personal, behavioral and attitudinal categories to gain a competitive advantage with predictive insights.

The key customer data types are personal data, behavioral data, and attitudinal data. Internal data reflects your own customers and processes, while external data reflects the wider market. Both data sources are needed for a 360-degree customer view.

Customer analytics started with direct marketing in the 1960s. Pioneers like Acxiom helped banks target promotions using data-driven customer segmentation. The internet allowed more advanced use of data, with Amazon and Facebook as major examples. Their recommendation engines are powered by analyzing user data.

Trust is critical - customers must feel their data is secure. Algorithms must not violate laws like the Fair Housing Act.

Netflix has mastered using data to create shows with high success rates. It recommends next shows and picks thumbnails based on each user’s viewing history. The goal is to keep viewers engaged.

Real-time personalization is the next wave, reacting to customer micro-moments. Speed is key - yesterday’s data has less value than today’s. Walmart only uses data from the past few weeks. The end goal is to serve each customer with relevant recommendations and offers at the right time.

  • Today’s most sophisticated customer analytics strive for real-time insights to capture ‘micro-moments’ when customers are ready to buy.

  • Companies can build tech infrastructure to track customers from initial interest to sale, like Disney’s RFID Magic Bands.

  • Online services like Netflix constantly gather user data to understand preferences and improve features.

  • Customer data enables tailoring of products and services, like tyre manufacturers analyzing tread wear.

  • Unilever acquired Dollar Shave Club for $1 billion largely for its vast customer data and insights.

  • Some companies are using data to rebuild personal customer relationships that were lost with the growth of impersonal retail banking.

  • By following a clear data strategy, combining datasets, and learning from other companies, businesses can better understand customers and identify their needs in real-time.

  • Services are becoming increasingly important in the modern economy. Tech companies have led the way in using data and AI to create smarter, more customized services.

  • Many tech giants started by providing one key service very well (e.g. Amazon with online shopping, Facebook with social networking). They then used the data gathered to develop new revenue streams and create additional connected services.

  • Services like Spotify use data on listeners’ tastes to create personalized playlists and recommend new music. Companies are even exploring using AI to automatically generate creative content like music.

  • Industries not traditionally associated with tech are adopting similar practices - e.g. StitchFix uses algorithms to send users customized fashion items.

  • Uber disrupted the taxi market by using data to match riders and drivers efficiently. Companies aim to build increasingly smart services by leveraging data.

  • Uber leveraged data and machine learning to improve customer service response times and increase satisfaction. Their automated system reduced the average handling time per ticket by 15%.

  • Chinese rideshare rival Didi Chuxing employs thousands of data scientists to offer smart transportation services beyond ridesharing, like traffic prediction.

  • Older companies are also using data and AI to evolve, like Iron Mountain applying computer vision to digitize documents and provide insights to customers.

  • Candy introduced an all-inclusive laundry subscription service, using data to predict detergent needs and machine repairs.

  • Banks are providing personalized, AI-powered budgeting and money management tools to customers by analyzing spending habits. These aim to rebuild customer relationships and counter fintech disruption.

  • Fintech startups like Monzo and Chip offer innovative, data-driven services to attract customers away from traditional banks.

  • In insurance, telemetrics and usage-based coverage leverage driver data to offer policies priced based on actual driving behaviour.

  • The Internet of Things (IoT) - the growing network of connected sensors and devices - is the foundation for new smart insurance services. By 2025 there could be 75 billion IoT devices capturing data to build more accurate digital models of the world.

  • Insurance is moving from a transactional model to an ongoing service built on data exchange and understanding the customer. Companies like Vitality Health use data to encourage healthy lifestyles, with rewards for activity that reduces medical risks.

  • Smart home devices like Leakbot detect problems early to reduce insurance claims. The focus is on prevention and avoiding crises through data-driven insights.

  • In healthcare, virtual providers like Babylon offer 24/7 access to doctors and AI-assisted diagnosis. They accumulate rich user data to constantly improve services. Controversy exists around claims of AI matching human doctors.

  • Other innovations include virtual hospitals to coordinate at-home care of patients using IoT monitoring devices. The pandemic accelerated adoption of telemedicine and remote healthcare.

  • Overall, increased data and analytics is enabling a shift to smarter, more customized and preventative services across insurance, healthcare and other sectors. But ethical considerations around intrusive data collection must be addressed.

  • Virtual clinics using telemedicine could help free up hospital beds by remotely monitoring patients with long-term illnesses, rather than needing them admitted. A key challenge will be determining which patients can safely stay at home vs needing in-person care.

  • Fashion retailers like Stitch Fix are using AI to recommend clothes to customers based on their measurements and style preferences. Other services help automate product descriptions and trend forecasting using computer vision.

  • Robots-as-a-service allows companies to rent robots and the expertise to run them via a subscription, rather than large upfront investments. This can greatly improve warehouse productivity. Security is another growth area.

  • Smarter education services are emerging to provide flexible, lifelong learning opportunities as jobs evolve. Services use AI to customize courses and provide automated feedback. AR/VR creates immersive training environments.

  • Though not replaced yet, teachers can utilize AI tools to improve student outcomes. Overall, ongoing education and human-AI collaboration will be key as jobs get smarter.

Here is a summary of the key points about using data to make more intelligent products:

  • Smart devices and the Internet of Things (IoT) are enabling more products to become “intelligent” through the use of AI, data, and connectivity.

  • Connectivity allows products to capture data about their environment and user behavior, and share it with other devices or cloud platforms. This data can be used to add new “smart” features.

  • Early smart products were mainly just connected to the internet, but true intelligence comes from using AI to analyze data and optimize performance. Examples are fitness trackers that learn about the user to offer personalized health advice.

  • The IoT has exploded thanks to billions of connected devices sharing data and communicating. This allows for new use cases like smart homes, smart highways, and smart cities.

  • For businesses, smart products present an opportunity to rethink offerings and provide more value through personalization, convenience, and automation.

  • Companies need an IoT data strategy to handle the scale of data created. Key considerations are security, privacy, storage, and identifying valuable insights.

  • Overall, connectivity and data-driven intelligence are reshaping products by making them more useful, customizable, and integrated into our lives. Companies that capitalize on this can gain a competitive advantage.

  • Autonomous vehicles like self-driving cars and drones are an exciting application of intelligent products. Companies are racing to develop fully autonomous personal aircraft and passenger drone taxis.

  • Autonomous flight is already being tested, but social and regulatory issues remain. Smaller aircraft may become autonomous before large passenger planes. Retrofitting existing ships with autonomous capabilities is a practical approach.

  • The Mayflower autonomous research ship demonstrates how removing the need to support human crews enables more efficient vessel design.

  • Autonomous technology is also being used to assist human operators, like the AI-powered electric bike that optimizes battery usage.

  • Vehicles are being fitted with intelligent systems to improve safety, efficiency, and driving experience, laying the groundwork for fully autonomous operation.

  • Development of autonomous mobility solutions involves overcoming technological challenges as well as winning social acceptance and regulatory approval.

  • Autonomous robots and drones are disrupting delivery services. Amazon’s Scout robot and Starship’s delivery robots are being tested for autonomous last-mile delivery. Drones have also been used by companies like Samsung for deliveries.

  • Many home products are becoming “smarter” by incorporating AI and automation. Smart home devices range from connected (e.g. Philips Hue lights) to intelligent (e.g. Amazon Echo, Google Nest).

  • Fridges from Samsung and LG use computer vision, voice assistants, and apps to suggest recipes, order groceries, etc. based on contents.

  • Other intelligent home products include: coffee makers that learn owners’ preferences, washing machines that optimize cycles, toilets with voice control and health monitoring capabilities, toothbrushes that use AI for brushing feedback, robotic vacuums that navigate intelligently using computer vision and LIDAR.

  • Home security cameras use facial recognition to identify intruders. Insurance providers may offer discounts for AI home security systems.

  • Key overall points are that more and more consumer products are incorporating meaningful AI to provide personalized, optimized, and automated experiences for users.

  • London-based startup Zobi has developed a device called the Hedgehog that uses AI to protect homes from cyber threats. It scans connected devices to identify vulnerabilities that could be exploited by hackers. This addresses the growing cybersecurity risks from poorly secured smart home devices.

  • Intelligent healthcare products like smart watches, contact lenses, and inhalers monitor patients’ vital signs and symptoms. They provide early detection of potential issues and allow personalized care. Data from these devices also enables research into preventative healthcare strategies.

  • Doctors have AI-enabled tools like smart stethoscopes and imaging analysis software to augment diagnosis. These technologies help address things like doctor shortages and improve detection of conditions like lung cancer.

  • In business and manufacturing, IoT and predictive maintenance reduce downtime and costs by fixing issues before failures occur. Robots are becoming more intelligent, moving beyond repetitive tasks to autonomous decision-making thanks to advances like computer vision.

  • Overall, the combination of AI, big data, and connectivity is leading to a new generation of smart, self-improving products across industries.

The passage discusses how products and services across various industries are becoming more intelligent and autonomous thanks to advances in AI and data analytics. It provides examples of autonomous vehicles like drones and ferries, AI-powered robots for delivery and cleaning, smart appliances and devices for the home, medical devices like insulin pumps and stethoscopes with AI capabilities, industrial machines and farm equipment equipped with computer vision and deep learning to optimize operations, and wearables and gear for sports that track performance metrics. The key idea is that integrating AI and leveraging data from sensors and cameras enables products and services to operate with more intelligence, less human intervention, and greater efficiency and effectiveness. This transformation is happening across transportation, logistics, healthcare, manufacturing, agriculture and consumer domains.

  • Data can be used to optimize operational processes across all business functions, from product design to sales and marketing. This involves implementing data-based decision making to streamline processes.

  • Digital twins - digital simulations of business processes - can identify inefficiencies and optimize processes. They use real-world data to accurately model systems.

  • Data is widely used in sales, marketing and customer service to predict customer satisfaction, prevent churn, personalize recommendations, enable dynamic pricing, and optimize online sales.

  • Logistics and distribution can be optimized through route planning, predicting demand, and automated inventory management. This reduces costs and improves service.

  • Data analytics assists with recruiting, training, productivity monitoring, and other HR processes. It provides insights to improve workforce efficiency.

  • Overall, automating operations via machine learning algorithms applied to business data boosts efficiency, cuts costs, and allows competitive advantage. The key is tying optimizations to strategic goals.

  • Amazon’s recommendation engine is considered the benchmark for using data to improve business processes, especially in ecommerce. It gathers extensive data on customers to fine-tune recommendations.

  • Other ecommerce companies like Alibaba are investing in automation and AI to improve processes like product descriptions and marketing. This is an example of robotic process automation (RPA).

  • Customer service chatbots are another application, like the one used by Marks & Spencer to replace call center staff. AI can also monitor and improve human customer service interactions.

  • In distribution and logistics, data can optimize stock levels, transport routes, warehouse operations, etc. Companies like Otto and Amazon use data to reduce waste and delays and improve efficiency.

  • Computer vision and sensors help retailers identify problems with fresh produce. Walmart used data analytics to cut down on “missed scans” and shoplifting in stores.

  • Overall, data can be applied across business processes like marketing, customer service, supply chain, and operations to remove inefficiencies, reduce costs, and improve customer experience.

Here is a summary of the key points about using data to improve business processes:

  • Delivery and logistics can be optimized through data-driven routing, scheduling, and autonomous delivery vehicles like drones and robots. Amazon’s Scout delivery robot navigates sidewalks autonomously to complete the “last mile” of delivery.

  • AI and data can improve product development by creating simulated users to test products, suggesting optimal designs and materials through generative design, and quickly creating 3D models and visualizations. This streamlines development and ensures products align with customer needs.

  • The industrial Internet uses IoT, data, and AI to connect and optimize manufacturing. Benefits include predictive maintenance, remote monitoring, and self-optimizing smart factories. Siemens used it to improve efficiency in their plants.

  • Robots and computer vision are automating more manufacturing processes like quality control. PepsiCo uses AI to analyze acoustics and images of chips to optimize production. Robots can even build other robots now.

  • Overall, data provides the feedback loop to drive automation, efficiency, and innovation in business processes ranging from R&D to delivery. This reduces costs and delivers better products and services.

Here is a summary of the key points about using data to improve business processes:

  • Manufacturing and production is being transformed by data-driven automation and robotics, such as ABB’s automated robot factory in Shanghai. AI-enabled ‘cobots’ can collaborate safely with human workers.

  • AI is used in IT operations for monitoring, automation, and security. It can detect cyberattacks and anomalies. Analysts have labeled this trend AIOps.

  • Accounting and finance tools like Xero use AI for automated data entry and short-term cashflow forecasting. Accountants are excited about how AI can improve their processes.

  • In HR, AI is used for resume screening and initial interviews to find top talent faster. Chatbots like Unilever’s Unabot help onboard new employees. Controversially, Amazon has used data to identify low productivity warehouse workers for termination.

  • Overall, mundane and routine tasks in business processes like manufacturing, IT, finance, and HR are being augmented and automated using data-driven intelligence. This allows human workers to focus on higher-value activities.

Here are the key points for how data can increase the value of a company:

  • Big tech companies like Amazon and Salesforce have made expensive acquisitions primarily to gain access to the target company’s data, not necessarily their products or services. The data helps them expand into new areas.

  • Companies with large volumes of high-quality data can often command higher valuations and sale prices because data is seen as a valuable asset.

  • Data can give companies competitive advantages and strategic insights that increase their value. For example, Netflix uses viewing data to create original content their subscribers want.

  • New data-driven business models and revenue streams also boost a company’s value. Some companies monetize data directly through selling access or insights.

  • Well-managed and governed data helps companies operate more efficiently and make better decisions, improving performance and value.

  • Investors and acquirers are willing to pay higher premiums for companies with strong data capabilities and assets. Data expertise is in high demand.

In summary, data is increasingly being seen as a core strategic asset that can significantly enhance a company’s overall value in various ways, from enabling better decision making to creating new revenue streams. Companies need to focus on collecting and managing the data that allows them to generate the most value.

  • Slack has built extensive datasets and models around workplace communication that are valuable assets. Microsoft buying LinkedIn gave it access to LinkedIn’s user data to help personalize Microsoft’s own business tools. When Google bought Nest, it was interested in Nest’s data on how consumers interact with connected home devices.

  • Data itself can become a company’s most valuable asset. Tesco’s Clubcard loyalty programme generated huge amounts of customer data managed by Dunnhumby. This data was so valuable that Tesco bought a majority stake in Dunnhumby. When Tesco later tried to sell Dunnhumby, its value dropped dramatically without the Tesco data.

  • For companies like Google and Facebook, user data is the core business asset. By gathering data on searches and social connections, they sell access to users via advertising. Though not new, it now happens at unprecedented speed and scale due to the internet.

  • Other companies like Experian and Acxiom pioneered data-driven marketing by combining public and private data sources. They have diversified into varied services to monetize their data assets. Newer companies like Cosmose AI offer real-time data insights to retailers.

  • The huge volumes of granular data these companies accumulate, combine, and derive insights from is their key differentiator and source of value. Their business models are built around monetizing access to this data.

  • Companies like Acxiom and Experian have built their business models around gathering and selling data, often by mining external data sources. While gathering massive datasets may not be feasible for all organizations, the ability to analyze data can boost a company’s value.

  • Companies can create extra revenue streams by selling access to their data or partnering with interested parties who can use the data. Examples include Tesco Clubcard data, John Deere agricultural data, and Apple’s health partnerships with IBM.

  • Credit card companies like Visa, Mastercard, and American Express have huge transactional datasets and sell anonymized insights to retailers.

  • Google profits from Nest smart home data by partnering with utility companies to manage energy demand.

  • Uber leverages its data on transportation patterns and behaviors to sell insights to urban planners and researchers.

  • In any industry, there are often opportunities to derive additional value from collected data, whether by enhancing services or establishing data partnerships. The key is identifying what data is valuable to outside parties and finding ways to monetize it.

Here are some tips for identifying potential data use cases through brainstorming:

  • Gather a diverse group of stakeholders - include people from different departments and levels of seniority to get a wide range of perspectives.

  • Frame the discussion around business needs and goals - don’t just focus on data capabilities, but how data can help solve key business challenges.

  • Encourage creative thinking - this is the time for blue sky ideas, don’t dismiss anything as unrealistic at this stage.

  • Leverage existing assets - review current data sources, analytics tools, and talent to spark ideas on how they could be better utilized.

  • Look outside your organization - research how other companies are using data and analytics in innovative ways that could inspire new ideas.

  • Categorize ideas into quick wins and transformational - quick wins provide faster returns with less investment, while transformational initiatives have bigger long-term impact.

  • Prioritize based on impact and feasibility - consider factors like resources required, implementation timelines, and potential business value when deciding what to tackle first.

  • Capture everything - record all ideas generated even if not selected for initial use cases, they may spark future initiatives.

The key is to set the stage for open, creative thinking about how data could transform any area of the business. The brainstorming session should produce a long list of potential use cases to then prioritize and evaluate for the initial data strategy.

  • The goal of this step is to identify specific opportunities to use data to meet strategic business goals across six key areas: better decision-making, understanding customers, smarter products/services, improving processes, creating monetizable data assets.

  • For each potential use case, fill out a template to detail how it links to strategy, the objective, how success will be measured, who will own it, who the customers are, and what data is needed.

  • Use a simple example of an ice cream shop owner wanting to better target marketing to “ice cream lovers” to illustrate the process.

  • Link use cases directly to strategic goals like increasing revenue or lifetime customer value.

  • Identify clear objectives, success metrics, data owners, and data customers.

  • Determine what data is needed from internal sources like transactions and external sources like demographics to accomplish the use case.

  • Careful planning of use cases this way ensures data projects achieve ROI and business impact.

Here are some key considerations for implementing the customer analytics use case for the ice cream shop:

  • Ensure proper data governance practices are in place before collecting or analyzing any customer data. This includes getting consent, having secure data storage, and only collecting necessary data.

  • Carefully evaluate different technology options like cloud services vs on-premise solutions based on factors like cost, speed, security needs, etc. Leveraging pre-built tools can be a good starting point.

  • Identify skills gaps in areas like data analytics, IT/networking, business intelligence, and communications. Consider upskilling current employees or hiring specialists as needed.

  • Set up clear processes for handing off insights to the appropriate business teams and ensuring they understand how to take action.

  • Monitor progress closely once implementation begins to identify any issues early. Be ready to tweak the approach if certain tech or skills don’t end up meeting needs.

  • Plan for how insights will be communicated on an ongoing basis as new data comes in. Dashboards, regular reporting, or similar routines will be important.

  • Make sure legal and compliance needs are continually addressed as the use case evolves, including regular reviews of data practices.

  • Overall, take an iterative approach to implementing the use case, starting simple and building on success. Be prepared to learn and adjust along the way.

  • Companies today have access to more data than ever before, but much of it is ‘dark data’ - information that exists but is difficult to access and analyze.

  • Dark data can include things like archived video footage without digitization capabilities, physical archives and records, and social media posts that are hard to decode.

  • The term ‘dark data’ is analogous to ‘dark matter’ in physics - we know the data exists and has an effect, but don’t fully understand it yet.

  • Organizations try to model the effects of dark data as best they can while also working to improve their capabilities to unlock and leverage it.

  • Sourcing and collecting the right data to answer key business questions is critical but can be challenging. Data may be internal or external, structured or unstructured.

  • Strategies like digitizing archives, using AI for sentiment analysis of social media, and consolidating siloed data can help transform dark data into usable insights.

  • Effectively leveraging more data requires considerations around data quality, integration, governance, security, and more.

  • Turning vast amounts of dark data into actionable insights is an ongoing process as technology and capabilities evolve. A thoughtful data strategy is key.

  • Data scientists analyze large amounts of data from different angles to uncover insights, similar to how physicists study particles under extreme conditions.

  • Iron Mountain helps companies extract value from ‘dark data’ like archives and documents by digitizing and structuring it for analysis.

  • Strategically using both structured and unstructured, internal and external data together can reveal the most valuable insights by showing connections between different elements like customers and processes.

  • Structured data like databases is easy to analyze but may not reveal deep insights. Unstructured data like text and video is harder to work with but can provide competitive edge.

  • Real-time data allows acting on ‘micro moments’. Collection frequency depends on strategic objectives.

  • Understanding different data types like structured, unstructured, internal, external is key to planning how to collect and use them based on their characteristics.

  • Structured data is organized in databases/spreadsheets with a predefined schema. Unstructured lacks organization and includes text, images, video.

  • Combining internal operational data with external demographic, social media, economic data provides a fuller picture for models and simulations.

  • Structured data represents only 20% of available data, while 80% is unstructured. As more activities move online, the proportion of unstructured data will grow.

  • Unstructured data is less “rich” than structured data and offers a more limited picture. Using other data sources alongside structured data can provide more insights.

  • Structured data is cheap, easy to store, and easy to analyze with basic tools, but advanced analytics tools can uncover more insights.

  • Unstructured data is messy and requires more complex, expensive systems to store and analyze. However, it contains undiscovered value.

  • Semi-structured data has some structure like tags but lacks database structure. It bridges unstructured and structured data.

  • Unstructured data is transformed into structured data by AI/ML techniques like computer vision and natural language processing.

  • The main advantage of unstructured data is the vast amount of rich, descriptive content it contains.

  • Internal data is proprietary data owned and controlled by a business, like sales, financial, customer service data.

  • Downsides of internal data include costs of maintenance and security, and it may not provide enough insights alone.

Here is a summary of the key points about external data:

  • External data comes from outside an organization and can be publicly available or privately owned. It can be structured, unstructured, or semi-structured.

  • Examples include social media data, government data, economic data, satellite imagery, and weather data. External datasets can be purchased if needed.

  • Advantages of external data include access to more data than what a company could generate internally, without the hassles of storing and managing it. This can benefit smaller companies especially.

  • Risks include reliance on external sources that can cut off access or raise prices. The costs and risks need to be weighed against the potential benefits.

  • New types of external data include:

  • Activity data like online behaviors, transactions, social media. Provides insight into what customers actually do.

  • Conversation data from calls, emails, messages. Can analyze sentiment and satisfaction.

  • Photo/video data from cameras, satellites. Used for monitoring, navigation, more.

  • Location data from mobile devices, wifi networks. Tracks real-world movements.

  • Sensor data from internet of things devices. Automated data collection.

  • External data expands insights beyond just internal data. Continually refer back to strategic goals to determine the most useful external data to collect and analyze.

  • Storage and management of data can be expensive, so it’s important to have a clear business need before collecting and storing large amounts of data. However, if you’re already collecting data like security footage, finding ways to use it may not cost much.

  • Sensor data from products like smartphones contains a wealth of information, but often needs context from other data sources to be truly useful. The ubiquity of sensors makes collecting sensor data easy.

  • Look internally first to see if your business already has useful data, or could easily start collecting it. This includes financial data, customer surveys, transaction records, video/photo data, and data from smart products and sensors.

  • External data sources like data brokers, government open data, scientific organizations, and social media platforms can supplement internal data. Though some is expensive, much is free. Tools like Facebook and Google Trends allow access to insights without exposing personal data.

  • The most powerful data strategies combine newly created internal data with external data sources. Internal data provides uniqueness, while external data provides context. Together they offer more value than either alone.

Here is a summary of the key points about the ethics of AI:

  • AI has the potential to greatly benefit society, but also carries risks if misused. Companies need to carefully consider the ethical implications of how they use AI.

  • Key ethical issues with AI include privacy, consent, bias, transparency, accountability, and the potential to negatively impact human lives and society.

  • Examples like automated firing of employees by Amazon raise concerns about machines making impactful decisions about people’s lives.

  • Consent and privacy are critical issues - collecting and using people’s private data without permission is unethical.

  • Legality and ethics are separate considerations - something may be legal but still unethical.

  • Powerful technologies like machine vision and natural language processing have dual use - they can be used to benefit or harm. Companies should think carefully about how their AI applications fit on the ethics spectrum.

  • Overseeing AI and data projects requires considering consent, potential harm, and whether use aligns with ethical values. Approaching AI ethically is an important responsibility.

  • AI and automation raise many ethical concerns, such as systems making life-impacting decisions about people’s employment, access to services, etc. Care must be taken even when using AI just to run processes more efficiently, as automated systems can have unintended consequences.

  • Replacing human jobs with AI is an ethical issue - while AI may create new jobs, organizations need to assess the impact on their workforce and try to reassign affected humans to new roles leveraging imagination and empathy.

  • AI techniques like generative adversarial networks (GANs) allow creating fake images/text of people, contributing to spread of misinformation. Legitimate business uses of GANs exist but ethical risks around misrepresentation must be considered.

  • Lack of transparency around AI decision-making (“black box” algorithms) raises ethical concerns when AI is used in areas impacting lives. Some believe AR/VR could help interpret and interrogate algorithms.

  • High energy usage of AI systems has environmental impact. While AI enables efficiencies, each application must be assessed for environmental ethics.

  • Initiatives exist promoting ethical AI principles around benefiting society, respecting rights and diversity, accountability, etc. Individual organizations should set up “ethics councils” to provide AI oversight.

  • Underuse of AI can also be unethical - if it could help address problems but is not used due to other ethical factors like transparency or environment.

The key points are:

  • Machine learning algorithms rely on the quality of the data they are trained on. ‘Clean’ data that is free from bias is essential for governance.

  • Data quality refers to metrics like consistency, accuracy, uniqueness, validity, timeliness and completeness. Auditing data against these metrics helps ensure it is fit for purpose.

  • Data bias occurs when the data does not fully represent reality. This can lead to issues like discrimination if left unchecked. Balancing bias while ensuring fairness is an ethical challenge.

  • Staying compliant with evolving privacy laws and regulations around data use is crucial. Users must actively opt in and consent to how their data is used.

  • Governance also covers respecting IP rights and ownership of algorithms and data.

The overall goal is to make sure the data used in initiatives is of high quality, free from bias where possible, and handled legally and ethically. Strong governance principles enable trust and prevent harmful outcomes.

Here are the key points from the passage on data governance, ethics, and trust:

  • Data governance is important for managing data quality, bias, legal/regulatory requirements, security, and ethics. A comprehensive data governance strategy is essential.

  • Data quality issues like inaccuracy, incompleteness, and lack of context can lead to poor decision-making. Strategies for improving data quality include data cleansing, establishing processes and standards, and monitoring data lineage.

  • Biased data and algorithms can lead to discrimination. Steps to reduce bias include diversity in data collection and algorithm design teams, auditing algorithms, and considering context in data analysis.

  • Legal and regulatory requirements like GDPR place obligations on companies to only collect necessary personal data, keep it secure, and allow people rights over their data. Fines for violations can be substantial.

  • Data security threats like breaches can be costly. Strategies to improve security include access controls, encryption, and addressing vulnerabilities in connected IoT devices.

  • Ethical concerns around data collection, usage, and algorithmic decision-making should be considered. Principles of transparency, fairness, accountability should guide data strategies.

  • Overall, a comprehensive data governance strategy accounting for all these factors is key for any company serious about getting value from data while maintaining trust. Regular auditing and review of policies is important as technology evolves.

Here is a summary of the key points about the evolution of analytics:

  • Businesses have been using structured data and tools like SQL and Excel for analytics since the 1970s/1980s to extract insights on things like forecasts, revenue, productivity, workflow, and customers.

  • Correlation analysis has been a popular traditional technique to understand relationships between variables.

  • In the 1990s, data mining emerged to analyze large datasets and find patterns using techniques like machine learning.

  • Big data analytics arose to handle the velocity, variety, and volume of unstructured data from sources like social media, mobile devices, IoT.

  • Cloud computing enabled scalable, on-demand analytics.

  • Today, advanced techniques like deep learning, neural networks, natural language processing allow for cutting-edge applications.

  • The field is rapidly evolving with new tools and methods emerging constantly.

  • It’s important for businesses to identify the analytics approaches that fit their strategic goals and use cases rather than jumping to the newest thing.

  • Wish lists of potential future analytics applications can inform data strategy. But firms should focus on what works for them now rather than always chasing the latest trends.

  • Correlation and regression analysis are traditional methods for analyzing the relationship between variables in structured data. Correlation measures the strength of the relationship, while regression analysis models how the relationship changes over time to make forecasts.

  • These methods don’t work well for unstructured data like images and video. Companies initially used manual processes to add structure to unstructured data, but new analytical methods can now automate this.

  • Cloud computing has enabled advanced analytics by providing increased storage, computing power, and easy access to analytical toolsets. This allows real-time analysis of massive, fast-changing datasets.

  • Artificial intelligence refers to machines capable of learning on their own like humans. Current business applications use “specialized” AI designed for specific tasks, not “generalized” AI.

  • Machine learning algorithms work by training on data, making predictions, receiving feedback on errors, and adjusting their models to become more accurate. Increased data and computing power have recently made machine learning more useful.

  • Deep learning and reinforcement learning are more advanced forms of machine learning gaining popularity. Overall, AI and machine learning are providing new capabilities to gain insights from all kinds of data.

  • Supervised learning algorithms are trained on labeled data, allowing them to assess their own accuracy and improve. They are commonly used for regression analysis and classification tasks. Examples include decision trees, random forests, Naive Bayes, and k-nearest neighbors.

  • Unsupervised learning algorithms find patterns in unlabeled data by grouping similar data points. This allows them to classify new data without labeled examples. They are used for clustering and association tasks.

  • Reinforcement learning uses rewards and penalties to train algorithms to maximize performance on a task. It is considered semi-supervised because there are still right/wrong answers. It is useful for problems requiring adaptable behavior.

  • Deep learning uses artificial neural networks modeled after the brain’s neural networks. The neural networks are very large, allowing them to model complex patterns. Deep learning has enabled major advances in computer vision, speech recognition, and natural language processing.

Overall, different machine learning approaches have tradeoffs between accuracy, computational requirements, and the need for labeled training data. Recent advances have expanded the power of techniques like deep learning by leveraging increases in computational power.

  • Image and video analytics involve extracting insights from visual data using machine learning and deep learning. Useful for tasks like facial recognition, analyzing brand presence on social media, and guiding self-driving cars.

  • Text analytics extracts meaning from unstructured text data like emails, documents, and social media. Uses natural language processing to categorize, cluster, summarize, and analyze sentiment in large volumes of text.

  • Sentiment analysis specifically looks at subjective opinions and attitudes expressed in text, audio, or video. Helps understand public or stakeholder reactions to things like marketing campaigns, new products, or government policies.

  • Speech analytics converts audio data into text and then analyzes it. Can identify speakers, transcribe conversations, and extract insights. Useful for improving customer service in call centers.

  • Predictive analytics uses statistical and machine learning models to make predictions about future outcomes based on historical data. Valuable for forecasting sales, detecting fraud, predicting equipment failures, etc.

  • Prescriptive analytics goes beyond predicting future outcomes to actually recommend optimal actions or decisions. Combines predictive modeling with business rules, constraints, and objectives.

  • In general, advanced analytics like machine learning can be applied to unstructured data like images, text, speech, and video to automate understanding and gain valuable business insights. The key is matching the right analytic techniques to your business problems and data assets.

Here is a summary of the key points about creating a technology and data infrastructure:

  • There are four key infrastructure layers to consider: data collection, data storage, data analysis/processing, and data communication.

  • Data collection - Challenges include accessing data from diverse sources and formats. Solutions include APIs, web scraping, IoT sensors.

  • Data storage - Challenges include storing large volumes of data cost-effectively. Solutions include cloud storage, data lakes, data warehouses.

  • Data analysis/processing - Challenges include managing complex analytics and AI. Solutions include business intelligence tools, data science platforms, machine learning APIs.

  • Data communication - Challenges include enabling data sharing and collaboration. Solutions include data visualization tools, dashboards, reports.

  • Cloud services and “as-a-service” options are making advanced data capabilities more accessible for organizations. But legacy systems may require integration.

  • The optimal infrastructure will depend on the specific data strategy and objectives of each organization. The key is choosing flexible, scalable solutions that can support evolving needs.

  • By leveraging the right mix of modern data infrastructure, companies can effectively collect, store, analyze, and share data to generate valuable business insights.

  • Consider what existing infrastructure and capabilities a company already has in place that could potentially be leveraged or built upon as part of their data strategy. This could include data collection mechanisms, storage, analytics tools, or communication channels.

  • Assessing existing infrastructure can help identify gaps that need to be filled with new investments or third party services. It’s unlikely everything needed will already be in place.

  • Data, analytics, AI, etc. offered “as a service” through cloud providers has become a huge industry and can allow even small companies to access powerful capabilities without massive upfront investment. These services can handle the technical complexities behind the scenes.

  • However, relying solely on external as-a-service providers may not be the best fit if a company needs tight integration with core operations or to monetize data. Some level of in-house infrastructure may still be preferred.

  • For internal data collection, IoT and inexpensive sensors have revolutionized capabilities. Smartphone apps can now provide tracking and monitoring that once required expensive dedicated hardware.

  • Open source software can also help reduce infrastructure costs compared to proprietary solutions. The best infrastructure approach depends on the specific data use cases and business needs. There is no one-size-fits-all solution.

  • There are various options for capturing the data you need, including sensors, customer apps, CCTV, beacons, and website cookies. Real-time, streaming data is becoming increasingly valuable.

  • External data can often be accessed through free public datasets or purchased from data brokers and marketplaces.

  • For storing data, main choices are traditional on-premises solutions like hard disks, solid-state drives, and tape, versus cloud-based storage.

  • Cloud storage offers benefits like scalability, reliability, and not having to maintain your own infrastructure. Main providers include AWS, Microsoft Azure, Google Cloud.

  • Often a hybrid approach is used, combining some on-premises storage with cloud.

  • Key considerations when choosing storage solutions are volume of data, frequency of access, speed and reliability needed, security and compliance requirements.

  • Cloud storage allows you to get up and running quickly with data storage and analytics. You simply sign up for a subscription and have access to storage and computing power immediately.

  • It is flexible, scalable, and often more affordable than on-premise solutions, especially for startups. You can create additional storage as needed.

  • ‘Cloud’ refers to storing data on remote servers operated by a cloud provider and accessing it over the internet. Data is distributed across multiple locations for redundancy.

  • Major public cloud providers include Amazon Web Services, Microsoft Azure, Google Cloud, Alibaba Cloud, and IBM.

  • Private cloud refers to internal data centers managed like a public cloud. Hybrid cloud combines public and private elements.

  • Avoiding data silos is crucial - data should be made available across the organization. Cloud makes this easier by centralizing storage and access.

  • Security is an important consideration with cloud. Though risks exist, large providers may offer robust security. Main threats are around access control and phishing.

Here is a summary of the key points about creating a technology and data infrastructure:

  • The infrastructure consists of four key layers - data collection, data storage, data analysis, and data communication.

  • Data collection involves identifying data sources, collecting and aggregating data from those sources. This can be done via APIs, web scraping, IoT sensors, mobile apps, etc.

  • Data storage requires deciding on a database structure, data warehouses, data lakes, etc. Cloud storage is increasingly popular. Future storage may use DNA, nanoparticles, etc.

  • Data analysis involves preparing, modeling, and drawing insights from the data using tools like Python, R, Spark, cloud analytics services, etc. GPUs and future quantum computing will provide processing power.

  • Data communication puts insights into the hands of stakeholders through visualization, self-service analytics, reporting systems tailored to different groups.

  • The infrastructure provides the foundation to execute the data strategy and use cases in order to achieve business objectives. Proper design of the layers and components is crucial for extracting maximum value from data.

Here are a few key points about the data skills shortage and what it means for businesses:

  • There is a huge gap between the demand for data skills and the available talent. By 2030, it’s estimated there will be a shortage of over 250,000 data scientists in the US alone.

  • Data analytics, machine learning, and AI skills are highly sought-after and well-compensated. Salaries for data scientists can exceed $120,000 in the US. Competition for talent is fierce.

  • Businesses of all sizes are looking to build in-house data teams. But many struggle to attract and retain qualified candidates due to compensation competition from tech firms.

  • The skills shortage impacts small and mid-sized companies the most. Unlike big tech firms, they often can’t match the salaries and perks offered by giants like Google and Facebook.

  • Outsourcing data work to consulting firms or freelancers is an option, but can be expensive over the long-term. Building some in-house capabilities is ideal for most companies.

  • Upskilling existing employees by offering data training and education is a cost-effective way to cultivate talent, but takes time and commitment.

  • Partnering with local colleges/universities on internship and apprenticeship programs is another smart talent strategy for businesses seeking analytics skills.

  • To attract scarce talent, companies may need to highlight benefits beyond compensation, like career development, work-life balance, and the chance to do meaningful work.

In summary, the data skills gap presents challenges for companies of all sizes seeking to leverage data analytics and AI. A combination of upskilling, smart recruiting, partnerships, and highlighting values may help secure the talent needed to execute a data strategy.

  • Job openings for data scientists are growing exponentially, far outpacing the supply of qualified candidates. According to, job postings exceed interested job seekers by a factor of three.

  • This skills gap poses challenges for companies looking to utilize data science and AI, who struggle to attract talent. Smaller companies in particular face difficulties competing with larger firms.

  • Potential solutions include upskilling existing employees, crowdsourcing, and automating parts of the data science workflow with AutoML tools like DataRobot and Alteryx. These “citizen data scientist” tools allow non-experts to deploy AI with little formal training.

  • Over time, more people are entering data science as it gains popularity, helped by its ranking as one of the best and most satisfying jobs. However, demand still exceeds supply.

  • Data scientists need a blend of technical skills like programming, math, and stats, and “soft” business skills like communication and problem solving.

  • Key skills include business acumen, analytics, computer science, statistics/math, creativity, and subject matter expertise. Well-rounded teams combine strengths across these areas.

The key skills needed to extract maximum value from data are technical skills like statistics, computer science, and mathematics; strategic thinking and business acumen; creativity and innovation; and communication skills. Organizations can build these competencies through recruiting new talent or training and upskilling existing staff. When recruiting, it may make sense to hire for a balance of skills rather than trying to find “unicorns” with all six competencies. Existing staff who show an affinity for working with data can be developed into “data ambassadors” through training. Free and low-cost training resources are available online from universities and companies to help build data science skills. Fostering a culture of data literacy and an understanding that working with data is everyone’s responsibility is also important. With the right balance of skills and a focus on continual learning, an organization can assemble effective data teams.

  • Online courses from providers like Coursera and Codeacademy can help build data skills within an organization. Courses are available covering topics like data visualization, IoT, and NLP.

  • Building a data culture means enabling a wide range of people across the business to analyze data and use it to inform decisions, not just relying on a few specialists. This ties into the concept of “citizen data scientists”.

  • An example is retailer Sears training 400 staff from its BI unit to do advanced customer segmentation, work previously done by data scientists. This created major cost efficiencies.

  • Data translators are important roles in data-driven organizations. They bridge the gap between technical data specialists and business decision makers, ensuring insights are translated into clear, actionable recommendations.

  • When in-house capabilities are insufficient, outsourcing data analytics is an option, either by partnering with a service provider or crowdsourcing via platforms like Kaggle.

  • When selecting a provider, recommendations and case studies are useful. Consider whether industry-specific knowledge is important. Ensure potential partners understand your key business questions and goals.

  • Having a draft data strategy in place before approaching providers is advisable, so you can clearly articulate what you want to achieve.

Here are a few key points to summarize how to put a data strategy into practice:

  • Attitude is key - leadership must buy into data as a vital asset and promote a data-driven culture. Challenge attitudes like “we are not a data company”.

  • Invest in change management - help people across the organization understand and adopt new data-driven ways of working. Provide training and support.

  • Start small - pilot projects with tightly defined use cases to demonstrate value and build confidence. Then scale up.

  • Communicate and collaborate - work closely with stakeholders and users to understand needs and get feedback. Share results.

  • Build in agility - be ready to tweak approaches based on learnings. Data strategies evolve over time.

  • Measure ROI - quantify the business impact of data initiatives. This helps justify further investment.

  • Plan for scaling - think ahead to how successful small projects can be grown for greater impact. Embed learnings across the organization.

  • Accept and learn from failures - some initiatives will fail. Analyze these to adjust approaches and accelerate learning. Failure is part of the process.

  • Executing a data strategy is similar to executing any other business strategy - it requires clear objectives, milestones, timelines, accountability, and monitoring of progress.

  • Many data strategies fail due to poor planning, lack of alignment with business goals, ineffective communication, lack of buy-in, silos between departments, management failure, and lack of necessary skills/resources.

  • To succeed, a data strategy needs involvement and understanding from all levels of the organization. Employees need to see how they fit into the bigger picture.

  • Regular communication across departments is vital so everyone understands their role. Data teams must build strong links with other departments and leadership.

  • Management failure, such as underestimating costs or not trusting algorithms, can doom a data strategy. Not having the right data skills at the right time is also a major risk factor.

  • With proper planning, communication, and buy-in, data strategies can deliver tremendous value, even for companies just getting started with data. The keys are taking a strategic approach and overcoming misconceptions about costs and competitiveness.

  • Implementing a data strategy successfully requires strong communication and buy-in across all levels of the company. A key goal is creating a “data culture” where data is recognized as a valuable asset and used to drive decisions and operations.

  • Leaders must lead by example in using data-driven decision making. Engaging key personnel as data advocates can help shift mindsets. Communicate openly on what data is being collected and why.

  • Change takes time. Use examples and focus on “pain points” data can alleviate to smooth adoption. Celebrate successes driven by data insights.

  • Revisit the data strategy regularly as business needs and the technology landscape evolves. Annually is a good baseline.

  • Changing business goals may require strategy adjustments. New technologies like machine learning and edge analytics may enable new opportunities. Costs tend to decrease over time.

  • The strategy must remain flexible and open to new data-driven opportunities and applications that arise. Data itself may reveal new revenue streams or business models.

  • We are living through an era of rapid technological change, perhaps the most exciting in history. Previous industrial revolutions brought great progress but also challenges.

  • The pace of progress is accelerating. In under 200 years we’ve gone from mechanical looms to AI systems that can see, talk, explore, and diagnose. This acceleration is likely to continue.

  • Key AI technologies like computer vision, NLP, ML, and analytics will become faster, more powerful, more efficient and cheaper. This will enable new applications.

  • Computer vision will augment human perception, allowing real-time video analytics for environmental monitoring, assisting firefighters, aiding the visually impaired etc.

  • NLP will enable natural conversations with machines, complex voice instructions, and seamless real-time translation.

  • ML analytics will become faster and more accurate, allowing automation of more decisions about health, lifestyle, finances etc.

  • More data sources will become available through proliferation of sensors and IoT devices. This “datafication” of the world will feed new AI applications.

  • But technology alone is not enough. We must direct it wisely to create positive change and improve standards of living, freedom, opportunities and life expectancy. The future will depend on us.

  • AI has the potential to help solve major global issues like poverty, hunger, health, education, gender equality, clean water, energy, economic growth, innovation, and inequality.

  • The UN’s 17 Sustainable Development Goals represent targets for improving the world by 2030. AI can contribute to achieving all of these goals.

  • Examples of how AI helps include: predicting natural disasters, improving crop yields, accelerating vaccine development, enabling remote education, identifying gender bias, managing power grids, stimulating economic growth, enhancing creativity and innovation, and reducing inequality by improving access.

  • The real value of AI lies not just in fun futuristic speculations but in solving these pressing global problems. We need to focus AI on the “why” - making a positive difference - rather than just the “how” and “what”.

  • By using AI to tackle major challenges, we can create a better future. But this requires focusing its development on worthy goals that improve life for all.

  • AI and data analytics have enormous potential to help achieve the UN’s Sustainable Development Goals and tackle major global issues like poverty, hunger, health, education, inequality, climate change, and environmental protection. However, they also pose risks if not managed responsibly.

  • There are concerns that advances in AI and automation could displace many jobs and exacerbate inequality if policies are not put in place to protect workers and ensure the benefits are shared. This could potentially lead to social unrest or even the collapse of the liberal world order.

  • Alternatively, some believe AI and automation could usher in a “post-work” era of luxury automated communism where machines provide for basic needs and humans focus on creativity and innovation. But this would require subordinating profits to human needs.

  • We may be entering an age of “digital feudalism” where tech elites control the means of digital production and hold disproportionate power over others. This raises concerns about consent, privacy, censorship, and centralized control.

  • The culmination of the AI/data revolution is uncertain - it could bring more equality or more concentration of power. Guiding it toward positive ends will require grappling with complex technological, economic, social and ethical questions.

Here are some key points regarding the index terms you highlighted:

AI Music Lessons - Online music education platform that uses AI to provide personalized lessons and feedback to students. Demonstrates use of AI in education. - E-learning platform that offers free online courses and skills training. Example of using data/AI to expand access to education and training.

Amazon - Major tech company known for extensive use of data and AI across its e-commerce operations, cloud services, logistics/fulfillment, recommendations, etc.

Augmented analytics - Approach that uses AI/ML to automate data analysis and provide insights. Making analytics more accessible.

Autonomous vehicles - Self-driving cars and trucks enabled by sensors, AI, and other technologies. Major area of development for AI and big data.

Babylon - Digital health startup that offers AI-powered diagnosis and health services via app. Example of AI transforming healthcare.

Bias - Algorithmic or data bias is a major concern with AI systems. Bias can lead to problematic outcomes. Mitigating bias is an important priority.

Blockchain - Distributed ledger technology enabling secure transactions and data sharing without central authority. Enabler for data sharing and AI.

Business needs - Critical to align AI/data initiatives with concrete business goals and needs. Avoid technology for its own sake.

Buy-in - Getting stakeholder and organizational buy-in is key challenge for executing an AI/data strategy. Requires clear communication and value.

Hope this provides a helpful overview of some of the key index terms and how they relate to central themes around AI, data, and business strategy. Let me know if you need any clarification or have additional questions!

Based on the summary, some key points are:

  • Machine learning involves various techniques like artificial neural networks, deep learning, reinforcement learning etc. These are used for tasks like image recognition, recommendation systems, predicting maintenance needs etc.

  • Data analytics skills like statistics, coding, communication are important. Organizations can develop in-house capabilities or outsource.

  • Ethics, bias and legality are important considerations in using AI and analytics. Data governance processes should be implemented.

  • Data is collected from various internal and external sources. It needs to be stored, processed and communicated securely.

  • The data strategy helps align data and analytics with business objectives. It should be executed and reviewed continuously.

  • Analytics is used in various sectors for applications like personalized recommendations, predictive maintenance, smart services etc. It can also help address global problems.

  • There are positive and negative potential impacts of technological progress like automation, inequality etc. Foresight is needed on how analytics and AI could shape the future.

Here is a summary of key points from the chapter:

  • Day-to-day business processes can be optimized and automated using data and analytics through digital twins - virtual representations that mirror real systems. This allows simulation and monitoring.

  • For sales, marketing and customer service, data enables highly personalized and tailored interactions and experiences for customers. Companies like Disney, Netflix and Starbucks are pioneers here.

  • Distribution, warehousing and logistics are being revolutionized through data-driven tracking, route optimization, predictive demand forecasting and automation.

  • For product development, data analytics facilitates customer-led design, simulations and digital prototyping, reducing costs and accelerating development cycles.

  • In manufacturing, smart connected machinery and real-time data create smart, flexible factories with early warning of potential issues. Data also enables customization.

  • Across support functions like IT, finance and HR, automation and AI are enabling greater efficiency, insight and decision making. Chatbots, digital assistants and intelligent processes are key applications.

  • Overall, data is transforming business processes through optimization, automation, customization and simulation - key foundations of the Fourth Industrial Revolution.

Author Photo

About Matheus Puppe