8. “Every day, we create 2.5 quintillion bytes of data
- so much that 90% of data in the world today has
been created in the last two years alone.” - IBM
9. “The number of transistors in a dense integrated
circuit doubles approximately every two years.” -
Moore’s Law
13. Why Now?
Cheap storage and computing
Algorithms for big data - big data for algorithms
Progress in machine learning (AI! Deep learning!)
Evolving consumer behaviour and user expectations
Excellent mature tools for data manipulation and analysis
Increasing awareness of the power of data
17. Data-Driven Business is Eating the World
Data-driven
products and services
Data-informed
product development
Data-informed
business
Data
Macro
Futures
18. Data-Driven Business is Eating the World
Data-driven
products and services
Data-informed
product development
Data-informed
business
Data
Macro
Futures
19. Data-Driven
Products
Products using data and algorithms to
solve problems and create value
Data plus machine learning algorithms
NOT about traditional databases or units
of data
Digital natives → Data natives
20.
21.
22. Product management = crafting
the right product with UX +
Data + Machine Learning +
Product & Business Strategy
23. Data-Driven Business is Eating the World
Data-driven
products and services
Data-informed
product development
Data-informed
business
Data
Macro
Futures
24. Data-Informed Product Development
Product Strategy
Ideation
Prototyping
Evaluation
Macro Insights
Development
Testing
Business Analytics
Product Analytics
A/B Testing
29. Data-Informed Business
Daily (hourly?) health of your business on dashboards
… democratically available
… with ability to quickly get meaningful and actionable
Insights
… with ability to explore
… 80% of questions answered without analyst involvement
32. Data-Driven Business is Eating the World
Data-driven
products and services
Data-informed
product development
Data-informed
business
Data
Macro
Futures
36. Data-Driven Business is Eating the World
Data-driven
products and services
Data-informed
product development
Data-informed
business
Data
Macro
Futures
37. Data ≠ Results
The Year is 2017 and search still sucks on most websites.
Forbes: “Why Investments In Big Data And Analytics Are Not Yet Paying Off”
Getting results from data requires
● New roles
● New collaborative mindsets
● New organizational structures
● Not getting stuck in one of the four boxes
40. AI is
A. That thing that will kill us
B. A subfield of computer
science
C. Machine learning algorithms
D. Deep neural networks
E. Automation, robotics
F. All of the above
41. AI is
A. That thing that will kill us
B. A subfield of computer
science
C. Machine learning algorithms
D. Deep neural networks
E. Automation, robotics
F. Artificial cognition
G. All of the above
Terminator
49. AI is
A. That thing that will kill us
B. A subfield of computer
science
C. Machine learning algorithms
D. Deep neural networks
E. Automation, robotics
F. All of the above
50. AI is
A. That thing that will kill us
B. A subfield of computer
science
C. Machine learning algorithms
D. Deep neural networks
E. Automation, robotics
F. All of the above
51. Machine learning - algorithms that allow
computers to learn without being explicitly
programmed.
52. Training Data + Labels Model
Training
Coefficients!
New Data Model
Training
Coefficients!
Predicted
Labels
54. AI is
A. That thing that will kill us
B. A subfield of computer
science
C. Machine learning algorithms
D. Deep neural networks
E. Automation, robotics
F. All of the above
56. AI is
A. That thing that will kill us
B. A subfield of computer
science
C. Machine learning algorithms
D. Deep neural networks
E. Automation, robotics
F. All of the above
58. AI is
A. That thing that will kill us
B. A subfield of computer
science
C. Machine learning algorithms
D. Deep neural networks
E. Automation, robotics
F. All of the above
60. What are the
implications of
this
product/user
shift?
● See above
● Product strategy!
● Requires higher
precision/lower errors
● Closer integration of all
data functions
● Great data first
● Not your old-style IT
department
● Not optional (AI startups?)
62. Risks
Data security
Privacy issues, tech surveillance, social media monitoring
Ethical implications of complex machine learning models
Complexity of software systems and interactions (new this year)
Over-focus on metrics leading to a local minimum
State and big company players
66. In modern companies, we’re looking to solve
relatively open ended problems in innovative
ways with the newest technology often by
producing fairly complex systems, whose shape
is unknown at the start.
Producing unknown complex outputs requires
complex inputs and complex processing
systems.
68. A few notes on modern organizations
● Engineering departments are like diabetes...
● From hierarchies to networks
● Companies as knowledge propagation mechanisms
● Multi-disciplinary approach is often a must
● Diversity on multiple dimensions
● Continuous deployment for systems - continuous
development for people
● Executives and board who have walked the walk earlier
71. Yes, do study some programming (but
know that good software engineering
is art on top of science)
72. Data: Everyone needs it!
“I keep saying the sexy job in the next ten years will be statisticians. People think
I’m joking, but who would’ve guessed that computer engineers would’ve been the
sexy job of the 1990s? The ability to take data—to be able to understand it, to
process it, to extract value from it, to visualize it, to communicate it—that’s going to
be a hugely important skill in the next decades, not only at the professional level
but even at the educational level for elementary school kids, for high school kids,
for college kids. “
Hal Varian, Google’s Chief Economist
January 2009
73. Jobs of the Future
Backend Engineer
Statistician
Data Engineer
Product Manager
Product Owner
(Tech) Project Manager
Frontend Engineer
Mobile Engineer
Data Scientist
Machine Learning Engineer
Product Analyst
UX Researcher
Business Analyst
Embedded Engineer
74. Jobs of the Future + Domain Knowledge
Backend Engineer
Statistician
Data Engineer
Product Manager
Product Owner
(Tech) Project Manager
Frontend Engineer
Mobile Engineer
Data Scientist
Machine Learning Engineer
Product Analyst
UX Researcher
Business Analyst
Embedded Engineer
75. Design, implement,
maintain data platform
Ask questions
Extract data
Clean and join big data
Analyse, and visualize data
Build data-based feature prototypes
Statistics, ML, AI
Communicate results
Ask questions
Extract data
Analyse and visualize data
Communicate results
Data Engineer
Data Infrastructure
Data Scientist Product Manager
Business Analyst
Machine Learning Engineer
Build Skills - Not Silos
76. Best data people I hire these days
come from domain fields
(aeronautics, physics, biology, pure
math, archeology, etc)
77. People are not resources
Hire for:
● Common sense
● Curiosity and interest
● Communication skills
● Ethics
● Basic technical skills
Develop:
● Technical skills
● Business skills
● Presentation and story-telling
skills
● Management skills
Modern life, there’s an app for that. The leading character in my story is technology and data. But equally important are people.
How can we move from creating just operational and reporting and into creating new products and business models.
Amount of data (the quote), cheap storage and computing (cloud, powerful machines in DCs), infrastructure and tools, algorithms. Also mention increasing awareness of the power of data and companies using data successfully.
Evolving consumer behavior and expectations on products to be customized : users are not only mere consumers of the data that companies produce any more, they are the producers and influencers through their posts, likes, shares, recommendations, reviews, ratings etc
Personalization/customization - it is also the effect of the amount of data being produced, can’t consume all or search through everything, we need to customize based on user preference (preset/selected) and/or usage data.
Big Data, what exactly is it?
Bigdata is like a healthy diet, everyone talks about it, nobody really knows how to do it, everybody thinks everyone else is doing it, so everyone claims they are doing it. (Rephrasing a quote from Dan Ariely)
There are four primary aspects of how data-driven businesses are working to creating the effortless experiences and tools we use daily. The four pillars of these are:
One. Data-driven products and services:
Next generation of products and services: by looking closely and collection data about key content, users, we can create better products and services our users actually want and need. Every part of our product can be more efficient: for ex, personalised search suggestions are no longer a novelty - they are now expected of all applications. We expect suggested completion on everything (think of all the search field in all your apps), but this completion is based on behaviour of all users before and also this particular users. For purchases made online, we can have sophisticated follow up scenarios. We can write text automatically knowing other texts, machine translation is in some ways a solved problem. Reputations can be managed when we buy things online (Fraud). User generated content is a key driver of the content economy too.
Data-driven product development:
How do we decide which features to implement or not? How should we prioritize our work? How can we enable the fast-iteration product development cycle that Jacqueline was talking about in the first section? We need to create key metrics for our business - and then some - and then use them consistently and automatically to inform ourselves about where we want to go.
[New, faster ways of developing products: as discussed before, how do we figure out what our products should be like? We talk to our users, but we may also have so many, we do fast iterations and experiment. In order to do this, we need to be able to create experiments with a subset of our users, and track to see how happy they are. We create success metrics (proxy for how happy users are), and then we have automated ways of figuring out how our experiments compare to our controls. For example, a new recommendation algorithms - we want to see whether people look at more articles or objects, or are more likely to buy something on a site. ]
Data-Driven Business:
Business and technical efficiency: Quarterly reports, monthly or even weekly cycles of data collection for key business metrics are becoming a key of the past. In order to optimize business, we want key metrics (usually around users, their happyness, and revenue) be available live or near-live. Being able to do simply sophisticated analyses, allows us to detect vulnerabilities and issues - as well as opportunities. It can also dramatically improve decision making in an organization.We can predict which customers you might lose _now_ and offer them something to prevent it. Technical efficiency: health, resource utilization, optimization of services (say, based on day-night patterns). Customer loyalty, churn, profitability, user segmentation
Data vs Competition:
Competitive advantage: most companies view data as their competitive advantage. User-generated content. Analyzing their users like, building better personalized products, thus getting more users and better data can create a kind of a moat, a competitive advantage for your business. It is not enough to create beautiful products, creating effortless products requires optimizing user processes so they seem effortless - to do so we need to both use fine-tuned data analysis of user behaviour as well as that of our content.
There are four primary aspects of how data-driven businesses are working to creating the effortless experiences and tools we use daily. The four pillars of these are:
One. Data-driven products and services:
Next generation of products and services: by looking closely and collection data about key content, users, we can create better products and services our users actually want and need. Every part of our product can be more efficient: for ex, personalised search suggestions are no longer a novelty - they are now expected of all applications. We expect suggested completion on everything (think of all the search field in all your apps), but this completion is based on behaviour of all users before and also this particular users. For purchases made online, we can have sophisticated follow up scenarios. We can write text automatically knowing other texts, machine translation is in some ways a solved problem. Reputations can be managed when we buy things online (Fraud). User generated content is a key driver of the content economy too.
Data-driven product development:
How do we decide which features to implement or not? How should we prioritize our work? How can we enable the fast-iteration product development cycle that Jacqueline was talking about in the first section? We need to create key metrics for our business - and then some - and then use them consistently and automatically to inform ourselves about where we want to go.
[New, faster ways of developing products: as discussed before, how do we figure out what our products should be like? We talk to our users, but we may also have so many, we do fast iterations and experiment. In order to do this, we need to be able to create experiments with a subset of our users, and track to see how happy they are. We create success metrics (proxy for how happy users are), and then we have automated ways of figuring out how our experiments compare to our controls. For example, a new recommendation algorithms - we want to see whether people look at more articles or objects, or are more likely to buy something on a site. ]
Data-Driven Business:
Business and technical efficiency: Quarterly reports, monthly or even weekly cycles of data collection for key business metrics are becoming a key of the past. In order to optimize business, we want key metrics (usually around users, their happyness, and revenue) be available live or near-live. Being able to do simply sophisticated analyses, allows us to detect vulnerabilities and issues - as well as opportunities. It can also dramatically improve decision making in an organization.We can predict which customers you might lose _now_ and offer them something to prevent it. Technical efficiency: health, resource utilization, optimization of services (say, based on day-night patterns). Customer loyalty, churn, profitability, user segmentation
Data vs Competition:
Competitive advantage: most companies view data as their competitive advantage. User-generated content. Analyzing their users like, building better personalized products, thus getting more users and better data can create a kind of a moat, a competitive advantage for your business. It is not enough to create beautiful products, creating effortless products requires optimizing user processes so they seem effortless - to do so we need to both use fine-tuned data analysis of user behaviour as well as that of our content.
Confusingly enough all of this is called Ai-first
There are four primary aspects of how data-driven businesses are working to creating the effortless experiences and tools we use daily. The four pillars of these are:
One. Data-driven products and services:
Next generation of products and services: by looking closely and collection data about key content, users, we can create better products and services our users actually want and need. Every part of our product can be more efficient: for ex, personalised search suggestions are no longer a novelty - they are now expected of all applications. We expect suggested completion on everything (think of all the search field in all your apps), but this completion is based on behaviour of all users before and also this particular users. For purchases made online, we can have sophisticated follow up scenarios. We can write text automatically knowing other texts, machine translation is in some ways a solved problem. Reputations can be managed when we buy things online (Fraud). User generated content is a key driver of the content economy too.
Data-driven product development:
How do we decide which features to implement or not? How should we prioritize our work? How can we enable the fast-iteration product development cycle that Jacqueline was talking about in the first section? We need to create key metrics for our business - and then some - and then use them consistently and automatically to inform ourselves about where we want to go.
[New, faster ways of developing products: as discussed before, how do we figure out what our products should be like? We talk to our users, but we may also have so many, we do fast iterations and experiment. In order to do this, we need to be able to create experiments with a subset of our users, and track to see how happy they are. We create success metrics (proxy for how happy users are), and then we have automated ways of figuring out how our experiments compare to our controls. For example, a new recommendation algorithms - we want to see whether people look at more articles or objects, or are more likely to buy something on a site. ]
Data-Driven Business:
Business and technical efficiency: Quarterly reports, monthly or even weekly cycles of data collection for key business metrics are becoming a key of the past. In order to optimize business, we want key metrics (usually around users, their happyness, and revenue) be available live or near-live. Being able to do simply sophisticated analyses, allows us to detect vulnerabilities and issues - as well as opportunities. It can also dramatically improve decision making in an organization.We can predict which customers you might lose _now_ and offer them something to prevent it. Technical efficiency: health, resource utilization, optimization of services (say, based on day-night patterns). Customer loyalty, churn, profitability, user segmentation
Data vs Competition:
Competitive advantage: most companies view data as their competitive advantage. User-generated content. Analyzing their users like, building better personalized products, thus getting more users and better data can create a kind of a moat, a competitive advantage for your business. It is not enough to create beautiful products, creating effortless products requires optimizing user processes so they seem effortless - to do so we need to both use fine-tuned data analysis of user behaviour as well as that of our content.
Innovation is a process
There are four primary aspects of how data-driven businesses are working to creating the effortless experiences and tools we use daily. The four pillars of these are:
One. Data-driven products and services:
Next generation of products and services: by looking closely and collection data about key content, users, we can create better products and services our users actually want and need. Every part of our product can be more efficient: for ex, personalised search suggestions are no longer a novelty - they are now expected of all applications. We expect suggested completion on everything (think of all the search field in all your apps), but this completion is based on behaviour of all users before and also this particular users. For purchases made online, we can have sophisticated follow up scenarios. We can write text automatically knowing other texts, machine translation is in some ways a solved problem. Reputations can be managed when we buy things online (Fraud). User generated content is a key driver of the content economy too.
Data-driven product development:
How do we decide which features to implement or not? How should we prioritize our work? How can we enable the fast-iteration product development cycle that Jacqueline was talking about in the first section? We need to create key metrics for our business - and then some - and then use them consistently and automatically to inform ourselves about where we want to go.
[New, faster ways of developing products: as discussed before, how do we figure out what our products should be like? We talk to our users, but we may also have so many, we do fast iterations and experiment. In order to do this, we need to be able to create experiments with a subset of our users, and track to see how happy they are. We create success metrics (proxy for how happy users are), and then we have automated ways of figuring out how our experiments compare to our controls. For example, a new recommendation algorithms - we want to see whether people look at more articles or objects, or are more likely to buy something on a site. ]
Data-Driven Business:
Business and technical efficiency: Quarterly reports, monthly or even weekly cycles of data collection for key business metrics are becoming a key of the past. In order to optimize business, we want key metrics (usually around users, their happyness, and revenue) be available live or near-live. Being able to do simply sophisticated analyses, allows us to detect vulnerabilities and issues - as well as opportunities. It can also dramatically improve decision making in an organization.We can predict which customers you might lose _now_ and offer them something to prevent it. Technical efficiency: health, resource utilization, optimization of services (say, based on day-night patterns). Customer loyalty, churn, profitability, user segmentation
Data vs Competition:
Competitive advantage: most companies view data as their competitive advantage. User-generated content. Analyzing their users like, building better personalized products, thus getting more users and better data can create a kind of a moat, a competitive advantage for your business. It is not enough to create beautiful products, creating effortless products requires optimizing user processes so they seem effortless - to do so we need to both use fine-tuned data analysis of user behaviour as well as that of our content.
Data-informated vs data-driven
Train your intuition
Take the guesswork out of it. Collect data and talk to users. Why do both? Because if you’re living big, you won’t be able to interview everyone; because users do one thing and say another. Still talk to users because your data may not surface all the problems, because your metrics may not be capturing general discontent.
Bring it back to collaboration and contributions from a variety of specializations
There are four primary aspects of how data-driven businesses are working to creating the effortless experiences and tools we use daily. The four pillars of these are:
One. Data-driven products and services:
Next generation of products and services: by looking closely and collection data about key content, users, we can create better products and services our users actually want and need. Every part of our product can be more efficient: for ex, personalised search suggestions are no longer a novelty - they are now expected of all applications. We expect suggested completion on everything (think of all the search field in all your apps), but this completion is based on behaviour of all users before and also this particular users. For purchases made online, we can have sophisticated follow up scenarios. We can write text automatically knowing other texts, machine translation is in some ways a solved problem. Reputations can be managed when we buy things online (Fraud). User generated content is a key driver of the content economy too.
Data-driven product development:
How do we decide which features to implement or not? How should we prioritize our work? How can we enable the fast-iteration product development cycle that Jacqueline was talking about in the first section? We need to create key metrics for our business - and then some - and then use them consistently and automatically to inform ourselves about where we want to go.
[New, faster ways of developing products: as discussed before, how do we figure out what our products should be like? We talk to our users, but we may also have so many, we do fast iterations and experiment. In order to do this, we need to be able to create experiments with a subset of our users, and track to see how happy they are. We create success metrics (proxy for how happy users are), and then we have automated ways of figuring out how our experiments compare to our controls. For example, a new recommendation algorithms - we want to see whether people look at more articles or objects, or are more likely to buy something on a site. ]
Data-Driven Business:
Business and technical efficiency: Quarterly reports, monthly or even weekly cycles of data collection for key business metrics are becoming a key of the past. In order to optimize business, we want key metrics (usually around users, their happyness, and revenue) be available live or near-live. Being able to do simply sophisticated analyses, allows us to detect vulnerabilities and issues - as well as opportunities. It can also dramatically improve decision making in an organization.We can predict which customers you might lose _now_ and offer them something to prevent it. Technical efficiency: health, resource utilization, optimization of services (say, based on day-night patterns). Customer loyalty, churn, profitability, user segmentation
Data vs Competition:
Competitive advantage: most companies view data as their competitive advantage. User-generated content. Analyzing their users like, building better personalized products, thus getting more users and better data can create a kind of a moat, a competitive advantage for your business. It is not enough to create beautiful products, creating effortless products requires optimizing user processes so they seem effortless - to do so we need to both use fine-tuned data analysis of user behaviour as well as that of our content.
There are four primary aspects of how data-driven businesses are working to creating the effortless experiences and tools we use daily. The four pillars of these are:
One. Data-driven products and services:
Next generation of products and services: by looking closely and collection data about key content, users, we can create better products and services our users actually want and need. Every part of our product can be more efficient: for ex, personalised search suggestions are no longer a novelty - they are now expected of all applications. We expect suggested completion on everything (think of all the search field in all your apps), but this completion is based on behaviour of all users before and also this particular users. For purchases made online, we can have sophisticated follow up scenarios. We can write text automatically knowing other texts, machine translation is in some ways a solved problem. Reputations can be managed when we buy things online (Fraud). User generated content is a key driver of the content economy too.
Data-driven product development:
How do we decide which features to implement or not? How should we prioritize our work? How can we enable the fast-iteration product development cycle that Jacqueline was talking about in the first section? We need to create key metrics for our business - and then some - and then use them consistently and automatically to inform ourselves about where we want to go.
[New, faster ways of developing products: as discussed before, how do we figure out what our products should be like? We talk to our users, but we may also have so many, we do fast iterations and experiment. In order to do this, we need to be able to create experiments with a subset of our users, and track to see how happy they are. We create success metrics (proxy for how happy users are), and then we have automated ways of figuring out how our experiments compare to our controls. For example, a new recommendation algorithms - we want to see whether people look at more articles or objects, or are more likely to buy something on a site. ]
Data-Driven Business:
Business and technical efficiency: Quarterly reports, monthly or even weekly cycles of data collection for key business metrics are becoming a key of the past. In order to optimize business, we want key metrics (usually around users, their happyness, and revenue) be available live or near-live. Being able to do simply sophisticated analyses, allows us to detect vulnerabilities and issues - as well as opportunities. It can also dramatically improve decision making in an organization.We can predict which customers you might lose _now_ and offer them something to prevent it. Technical efficiency: health, resource utilization, optimization of services (say, based on day-night patterns). Customer loyalty, churn, profitability, user segmentation
Data vs Competition:
Competitive advantage: most companies view data as their competitive advantage. User-generated content. Analyzing their users like, building better personalized products, thus getting more users and better data can create a kind of a moat, a competitive advantage for your business. It is not enough to create beautiful products, creating effortless products requires optimizing user processes so they seem effortless - to do so we need to both use fine-tuned data analysis of user behaviour as well as that of our content.
The Turing test, developed by Alan Turing in 1950, is a test of a machine's ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human. Turing proposed that a human evaluator would judge natural language conversations between a human and a machine designed to generate human-like responses. The evaluator would be aware that one of the two partners in conversation is a machine, and all participants would be separated from one another. The conversation would be limited to a text-only channel such as a computer keyboard and screen so the result would not depend on the machine's ability to render words as speech.[2] If the evaluator cannot reliably tell the machine from the human, the machine is said to have passed the test. The test does not check the ability to give correct answers to questions, only how closely answers resemble those a human would give.
Commoditization of everything else. Not only.
Commoditization of everything else. Not only.
Machine learning engineer - that’s where true innovation. Compress collaboration