Data-driven or data-informed: why does a digital not replace a person?
- Джимшер Челидзе
- Apr 17, 2024
- 12 min read
Updated: Oct 2
Digitalization and digital transformation are inseparable from the use of data for decision-making. But how can you use data to make decisions?
Content
In total, there are now 4 strategies for using data for decision making:
No-Data
Data-driven
Data-informed
Data-inspired
And the most common request and thesis in Russia: “We are Data-driven.” But what is hidden behind this, and why are the world's IT giants moving away from this strategy as a single one for all situations and areas of activity?
No-Data
No-data - management and decision-making without taking into account data and analytics. This is the most common approach in small and medium businesses. It is used by a dominant number of managers and organizations.
Data-driven
Data-driven - management and decision-making exclusively quantitative based on data and metrics
First they get numbers/metrics, and then make decisions based on them. The resulting numbers are the first thing they look at when deciding where to move next. Great for justifying decisions to senior management/shareholders.
Data-informed
Data-informed - management and decision-making based on data.
Metrics provide additional information that can be useful in decision making. However, the final decision is made taking into account past experience, expertise, intuition, etc.
Data-inspired
Data-inspired is an approach in which analyzing the market and trends, searching for non-obvious connections in disparate data serves to make strategic decisions and search for new opportunities. The key thing is that we do not rely on experience and analysis of events, but on a vision of the future and research.
Allows you to build a strategy, but has the greatest risks
Let's look at what weak points does Data-driven have?
Lack of creativity
Data-Driven gives no room for creativity and different points of view
Narrow Focus
Data can only answer qualitatively those questions that were initially raised at the design stage
Elimination of Expertise
Data-driven approach does not take into account expert knowledge and experience
Bias
It is almost impossible to create an absolutely objective set of metrics; in any case, the manager’s view will be reflected in the set of metrics
The difficulty of digitizing people's reactions
Data-driven does not take the human factor into account: emotions and feelings, subjective opinion
Demanding on data quality
Data-driven is demanding on the quality of source data and the development of regulatory and reference information
At the same time, it is working with data and data quality that is one of the key reasons for most problems in IT projects
What problems exist in working with data, and what problems do they lead to?
Now, let's answer a few questions:
Is it possible for our companies to create solutions with ideal data?
Is it possible to create objective metrics without overloading managers?
How realistic is all this in conditions of constant change and uncertainty?
What matters most in times of uncertainty?
To do this, let's turn to the Kinevin model, which allows us to determine the order of actions depending on the complexity of the situation
As a result, Data Driven is poorly suited for decision making under conditions of uncertainty and low quality of source data.
And trying to create an ideal system of metrics can lead to an infinite cost of growth for each solution, overload of managers with data and critical errors
At the same time, Data-driven is great for making small daily decisions or decisions in stable conditions, in simple systems
Data-informed helps you navigate situations of uncertainty and combine expert qualities with data analysis and plan for the near future.
Disadvantages of the approach:
not suitable for making strategic decisions and searching for new opportunities, as it is based on experience and analysis of events and facts
It is more difficult to justify your decisions to management and stakeholders because they are not entirely based on quantitative data and largely depend on the point of view and picture of the world, experience and expertise
there is a risk of being influenced by cognitive distortions
the problem of multiple choice - there is too much input data, they will have to be prioritized, look for correlation, and the output information may be contradictory.
Data-inspired helps you build strategies and work for the long-term future, and find new solutions.
Disadvantages of the approach:
not suitable for operational and tactical control
decisions based on this approach are the most risky
forms abstract ideas and assumptions
it is even more difficult to justify your decisions to management and stakeholders depends entirely on the qualifications of the person analyzing the data
No Data Strategy (Intuitive solutions)
Successful examples
1. Southwest Airlines
The situation: In 1967, Herb Kelleher founded the airline on the principles of low cost without a detailed analysis of the air transportation market.
The approach: An intuitive understanding that people need low-cost flights without frills - food, designated seats, connecting flights.
The result: Southwest has become the largest budget airline in the United States, showing profits for 44 consecutive years.
The reasons for success: A proper understanding of the basic needs of the market, simplicity of concept, focus on the basic needs of customers.
2. Howard Schultz and Starbucks
The situation: In 1983, Schultz decided to transform the sale of coffee beans into coffee shops according to the Italian model, based solely on personal experience of a trip to Milan.
Approach: The decision was made based on an emotional perception of Italian coffee culture and an intuitive understanding that Americans would like it.
Result: Creating a global coffee empire with annual revenue of over $30 billion.
Reasons for success: Strong emotional connection with the product, creation of a unique customer experience, correct intuition about cultural needs.
Unsuccessful examples
1. The restaurant removes a popular dish
Situation: The restaurant owner decided to remove the dish from the menu, believing the waiters' opinion that it is rarely ordered.
Approach: The decision is made solely based on the subjective opinion of the staff, without analyzing sales data, profitability or submission time.
The result: Revenue dropped by 15%, regular customers began to leave, and the dish had to be urgently returned to the menu.
Reasons for failure: Ignoring objective data on sales and profitability, the HiPPO effect (opinion of the most influential person).
2. Quibi - streaming platform
The situation: In 2018-2020, a mobile video platform was launched for $1.75 billion based on the founders' intuition about youth behavior.
The approach: Investors and founders decided that short videos on mobile devices are the future of entertainment, without deep market research.
The result: A complete loss of investment, the closure of the platform 6 months after launch.
The reasons for the failure: overestimation of demand for a new content format, misunderstanding of the media habits of the target audience, ignoring competition with TikTok and YouTube.
Data Driven strategy
Successful examples
1. Netflix and "House of Cards"
The situation: In 2013, Netflix decided to invest $100 million in the creation of the first original series.
Approach: An analysis of data on the preferences of 23 million users revealed three key factors: the love of the British original "House of Cards", the popularity of director David Fincher and actor Kevin Spacey.
The result: The series received a 9.0/10 rating, Netflix attracted 3 million new subscribers, launched an era of original content.
Reasons for success: Accurate data on user preferences, correct interpretation of patterns, high-quality implementation based on identified insights.
2. Amazon recommendation system
Situation: Amazon has developed a personalized product recommendation system based on customer behavioral data.
Approach: Machine learning algorithms analyze the history of purchases, pageviews, search queries, time on the page and offer relevant products.
The result: Sales increased by 35%, and total revenue from recommendation algorithms exceeded $150 billion over 10 years.
Reasons for success: High-quality data, powerful algorithms, constant optimization based on feedback.
Unsuccessful examples
1. Target and the pregnancy case
Situation: Target's algorithm detected a minor's pregnancy by changing her buying behavior and sent an advertisement for children's products to her home address.
Approach: Automatic analysis of purchase patterns without taking into account ethical aspects and human control over the results of the algorithm.
The result: A loud scandal in the media, lawsuits for violation of privacy, and serious damage to the company's reputation.
Reasons for failure: Ignoring the ethical aspects of data usage, lack of human control over automated decisions.
2. Microsoft Tay chatbot
The situation: In 2016, Microsoft launched the AI chatbot Tay, which was supposed to be trained based on interactions with Twitter users.
Approach: Fully automatic machine learning based on user data without pre-filtering and content moderation.
The result: After 16 hours, the bot started posting offensive and extremist messages, and was immediately disabled.
Reasons for failure: Underestimation of social factors, lack of a system for moderation of training data, blind trust in the algorithm.
Data Informed strategy
Successful examples
1. Apple and the iPhone
The situation: The development of a revolutionary smartphone that was supposed to combine a phone, an iPod and an Internet communicator.
Approach: Steve Jobs combined market data analysis with a deep understanding of user needs, technical capabilities, and user experience.
The result: A revolution in the smartphone industry, with Apple's total iPhone revenue exceeding $365 billion.
Reasons for success: Perfect balance between data and intuition, focus on user experience, iterative development with constant feedback.
2. Procter & Gamble and Tide Pods
The situation: P&G sought to simplify the washing process and create an innovative product in a mature category.
Approach: The company conducted research on washing habits, analyzed data on consumer behavior, and applied a creative approach to detergent packaging.
Result: The creation of a completely new product category, increased sales of washing powder by 40%.
Reasons for success: Successful combination of quantitative research with qualitative insights, creativity in solving common problems.
Unsuccessful examples
New Coke (Coca-Cola, 1985)
The situation: Coca-Cola has conducted extensive research on the new taste, including 200,000 blind tests.
Approach: The company collected quantitative data (blind tests showed the superiority of the new formula), but incorrectly balanced them with qualitative factors - emotional attachment to the brand.
The result: Massive consumer outrage, falling sales, and a return to the classic Coca-Cola Classic formula after 79 days.
The reason for the failure: The classic mistake of overestimating quantitative data to the detriment of qualitative factors with a balanced approach.
Microsoft Kinect for Xbox One (2013)
The situation: Microsoft has decided to make the Kinect controller mandatory for Xbox One, despite mixed signals from research.
Approach: Research has shown a technical interest in gesture control and voice commands, but the team underestimated data on users' concerns about privacy and cost.
The result: Xbox One initially sold significantly worse than PlayStation 4, Microsoft was forced to remove the mandatory Kinect and eventually stopped developing it.
Reason for failure: Incorrect weighting of different types of data - technical capabilities versus social concerns of users
Data Inspired strategy
Successful examples
1. Tesla and electric vehicles
The situation: Elon Musk has decided to create mass-produced electric vehicles in an industry where previous attempts have failed.
Approach: Analyzing environmental trends, technological capabilities of batteries, and creating a revolutionary vision for the future of motor transport.
The result: Tesla's market capitalization has exceeded $800 billion, launching a revolution in the automotive industry.
The reasons for success are the correct vision of long-term trends, readiness for large-scale long-term investments, and a systematic approach to innovation.
2. Spotify Discover Weekly
The situation: Creating a personalized music playlist that is updated weekly for each user.
Approach: Analyzing music trends + studying user behavior + creative machine learning algorithms to find new music
The result: User engagement increased by 30%. Subscriber retention has improved significantly. A new standard of music recommendations in the industry has been created
Reasons for success: Spotify has not just improved recommendations - they have created a new way for people to interact with music, turning the discovery of new tracks into a weekly ritual.
Unsuccessful examples
1. Google Glass
The situation: Google invested more than $50 million in the development of AR glasses, which were supposed to change the way people interact with information.
Approach: A technological breakthrough in augmented reality without sufficient consideration of social perception and market readiness.
The result: The cessation of mass sales after 2 years, significant damage to Google's reputation in the field of innovation.
The reasons for the failure: Techno-centricity without taking into account the social context, privacy issues, high price and unpreparedness of the market.
2. Segway
The situation: In 2001, the Segway was positioned as a revolutionary means of transportation that would transform urban transportation.
Approach: Innovative technology of self-balancing transport without a deep understanding of the practical needs and limitations of users.
The result: The cessation of production in 2020 after 19 years of losses, without reaching payback.
Reasons for failure: Overestimation of the revolutionary nature of technology, underestimation of practical limitations (price, weight, infrastructure), incorrect assessment of the market size.
1. The paradox of equal effectiveness
All four strategies have successful and unsuccessful examples, which destroys the myth of the existence of a "better" decision-making strategy. Success is determined not by the choice of a specific strategy, but by the quality of its application in the context.
2. Contextual dependence of success
Each strategy demonstrates maximum effectiveness in specific conditions:
No Data: Time constraints + high expertise + simple tasks
Data Driven: Big Data + clear goals + operational processes
Data Informed: Complex context + Data + Experts
Data Inspired: Uncertainty + Innovation + Long-term
3. The criticality of the team's competencies - the lack of key competencies is the main reason for failures, regardless of the correctness of the strategy choice:
Data Driven requires analytical skills and ethical control
Data Informed needs synthesis skills and user understanding
Data Inspired involves creativity and risk-taking
No Data relies on deep expertise and intuition
4. At the same time, each strategy has its own blind spots
No Data
HiPPO effect - dominance of management opinion
Ignoring objective data
Incorrect assessment of the situation/ market, etc.
Data Driven
Blindly following data without context
Lack of ethical control
Underestimating social factors
Data Informed
Fear of cannibalization of existing business and conditions
Underestimating the social context
Reassessing technical superiority
Data Inspired
Reassessing the revolutionary nature of innovation
Underestimating the social context
Ignoring practical limitations
5. Key success factors for each strategy
No Data works with a deep understanding of the market, simple customer needs, and strong intuition based on experience.
Data Driven is effective when high-quality data is available, correctly interpreted, and ethical aspects are taken into account.
Data Informed is optimal for complex solutions that require a balance between data and expertise, with a focus on user experience.
Data Inspired is suitable for long-term innovation with a willingness to take high risks and a proper understanding of trends.
6. Critical mistakes to avoid
The use of Data Driven without data quality assurance and ethical control
Using the No Data approach for complex strategic decisions with high risks
Data Informed without sufficient expertise or with overestimation of quantitative data
Data Inspired without willingness to take high risks and long-term investments
Ignoring the social context and user needs in all strategies
Underinvestment in the development of the team's competencies for the chosen strategy
Underestimating the impact of organizational culture on the success of strategy application
7. What should be done?
Ensure data quality and ethics in Data Driven: implement data governance, quality metrics (completeness, accuracy, relevance), regular audits, and appoint data owners.
Limit No Data to the sphere of fast low-risk tasks: for strategic decisions, collect minimally sufficient data, conduct quick experiments
Strengthen Data Informed through synthesis: combine quantitative metrics with qualitative insights (custdev/UX, interviews), introduce explicit weights of arguments in decision‑making, and conduct cross-functional reviews.
Discipline Data Inspired: use a risk‑limited innovation portfolio, stage-gate and pilots, limited budgets for research tasks, clear criteria for transitions between stages and investment horizons.
Embed the social context and user needs: conduct regular research, stakeholder mapping, social/ethical impact assessments, and monitoring the impact of implementations.
Invest in competencies: develop data literacy, train decision frameworks (DACI/RAPID), create communities of practice, close gaps through targeted hiring and development of specialized employees (CDO, Data Steward, etc.).
Create a data culture and reward system: provide leadership sponsorship, rituals (demos, retrospectives), transparent metrics and data catalogs, as well as KPIs and incentives related to the use of high‑quality data in solutions.
8. The most important thing is that there is no "silver bullet" – there is no single approach for all situations.:
Data-driven is an excellent tool for solving everyday/operational tasks or in conditions of stability, i.e. for making 80% of all decisions. It is an operational management and planning/control tool
Data-informed is essential when creating new products, working with people, and planning for the near future. It is a tactical management and planning tool.
Data-inspired helps you create strategies and find new solutions, and work for the long-term future. It is a strategic management and planning tool.
A person is still a key element of any approach. He can only spend his experience and expertise not on daily routine, but on intellectual work. And according to research, 95% of companies unsuccessfully implement data-based solutions precisely because of the wrong choice of strategy for a specific situation. Success depends not on the data itself, but on the culture of its use and the understanding of the limitations of each approach by managers.
An example of which tasks each strategy can be used for is shown in the table below.
Data-driven | Data-informed | Data-inspired |
A/B tests Evaluating the performance and stability of new features | Development roadmap Prioritize the development of new features | Strategy Search for new opportunities |
Useful materials
useful links
