seizing the machine-learning opportunity how to define and implement your …ja... ·...
TRANSCRIPT
© Dain Studios 2016© Dain Studios 2016
Seizing the Machine-Learning Opportunity –How to define and implement your AI strategyTulevaisuuden tuotekehitys
Helsinki, May 24, 2018
Ulla Kruhse-LehtonenCEO, Co-founderTel. +358 45 639 [email protected]
© Dain Studios 2016
2
CEO, Founding Partner, DAIN Studios
Vice President, Consumer Analytics and Insights, Sanoma
Director, Consumer Analytics, Nokia
Management Consultant, Accenture, XLENT, Nexus
Economist, Labor Institute for Economic Research
PhD, Economics, Helsinki School of Economics
Information Leader of Year 2013, Finland
One of the 32 coolest and most influential women in Nordic tech by Business Insider Nordic, 2018
Ulla Kruhse-Lehtonen
© Dain Studios 2018© Dain Studios 2018
20+ Clients14 Industries5 Countries
3 StudiosHelsinkiBerlin
Munich
Data + AI+ INsight
From AI Strategy to Execution
2 years old
Team of 1810 Data Scientists
/ Engineers8 PhD
Own AI productsTravel AI
Smart Recruit
DAIN
© Dain Studios 2018
4
Data Scientist.The Sexiest Job
at Sanoma.
Source: Sanoma Strategy 2013
© Dain Studios 2016
Growth provides challenges and opportunities for all market players
© Dain Studios 2016© Dain Studios 2016
© Dain Studios 2016
Going beyond the hype...
© Dain Studios 2016© Dain Studios 2016
Machine-learning applications across industries
© Dain Studios 2016© Dain Studios 2016
Leverage data for existing business optimization as well as for new business
Internal DataExternal Data
New Business
Current Business
Data Partnerships
Business OptimizationBusiness Optimization
Data as a Business
Collaborate with external partners.Exchange data to enable new offerings or business models which would not be possible alone
Provide 3rd parties with access to your data assets, insights, and/or analytical capabilities to enable them to grow and improve their business (e.g. Data Market Platform)
Utilize external data sources to enhance your own data asset to enable further optimization of business processes
Combine internal data to further optimize existing business and processes to enable new offerings
© Dain Studios 2016
Data, Analytics, and AI play a significant role in the development of intelligent products and services
Source: Eric Rice, 2011
© Dain Studios 2016
Data and AI change the way how you develop productsExample self-driving cars
Tesla way of developmentMore than 200.000 cars deliver data
BMW way of developmentClose to <50 cars deliver data
EXAMPLES
© Dain Studios 2016© Dain Studios 2016
Unleash operational inefficienciesEXAMPLES
© Dain Studios 2016© Dain Studios 2016
Business models are changing
From:
To:
Source: Siemens Mindsphere
EXAMPLES
© Dain Studios 2016© Dain Studios 2016
Foundation of Smart Cities, Factories, EcosystemsEXAMPLES
© Dain Studios 2016© Dain Studios 2016
It all starts with a vision
What are the business goals want to achieve?
Where do I want to go with my business?
What are my prioritized use cases to get there?
It starts with a vision
© Dain Studios 2016
Define the ambition level for data
Data used for current business, product development, and new
business areas
AmbitiousModerate
Use data for the optimization of your current business
Data seen as an enabler Data seen as a strategic asset
Mainly internal data used Use internal and external data for differentiation
Focus on core business Own market seen widely
No/limited commercialization of internal data
APIs enable data as a business and data partnerships
Example
© Dain Studios 2016
Identify, define, and prioritize the AI opportunitiesEXAMPLE
Customer Journey / KPIs / Use Case Mapping AI & Data Opportunities
To
uch
po
ints
Ob
jecti
ves
KP
Is
Efficient in sales and marketing, continuously collecting valuable customer insights for product design & our partners
Optimize
conversion
Capture user
profiles
Best ownership
experience
Upgrade, resell
Appeal Guide Reassure Serve Support Enable Ensure longevity
Discover Consider Prefer Get started Use Extend use Upgrade
Always on, always serving our customers, and continuously engaging the market with the Nokia brand and products.
Best shopping
experience.
No hassle
onboarding
Efficient
marketing
Brand advocacy
# Reach (organic, paid)
#uniques (own sites)
# Engagement
# Likes
# newsletter registrations
Retention, upsell
Most lucrative
brand
# activation
# NPS
# conversion*
# active devices
# usage index* *
# NPS
# conversion*
# accounts (Care)
# number of care cases
# sentiments
Best customer
care
Understand and
engage users
Consistent omni-
channel experience
Desired
products
Dimensions: Overall, Country, touch point, Device model
# conversion
# basket value
# accounts (eCom)
# NPS store
Data Potential + User Volume Total Data Value
Data Potential is driven by the following:• Relevance of data for energy-
service provisioning• High business impact potential• Utilization in many use cases• The quality of the data is high
User Volume is driven by the following:• Number of accessible users
and/or devices (out of total customer base)
• Number of user/device records
The Total Data Value describes the data value opportunity assuming that value creation is realized via
analytics and automation, and the use cases are successfully
integrated into relevant business processes
Data Integration Effort*
Da
ta P
ote
nti
al
low high
low
high
10
6
11
17
16
15
12
18
19
Quick wins
© Dain Studios 2016© Dain Studios 2016
Assess your company’s analytics maturity level and set a target state
Source: Davenport & Harris, 2007
EXAMPLE
© Dain Studios 2016© Dain Studios 2016
Embedding analytics into business processes
Service Roll-Out
Spearheads
Analytics/AI Modeling
It is easier to roll out purely technical data products (e.g. recommendation engines) than products that involve people having to change their way of working (e.g. marketing automation).
© Dain Studios 2016© Dain Studios 2016
Leadership, governance, and a right incentive scheme drive the change toward a data-driven organization
Leadership
Governance Incentive
Data is a strategic asset for future
business
• Set the vision and drive direction –ensure continuation
• Thought leadership
• Drive and steer implementation of vision and strategy across the company
• Resolve conflict of interest or trade-offs
• Provide motivation for whole organization to head towards common direction
• Measure progress along defined KPIs• Incentivize data-driven innovation
© Dain Studios 2016© Dain Studios 2016
Data success is not only about data, analytics, and technology: How to drive an analytics culture?
Break down silosLead from the top AND
the middleSharpen the business
strategyExecute effectively
• Silos are culture killers• Ensure data capabilities are
used in the most beneficial areas for the company
• Most impactful analytics requires cross functional collaboration
• Requires full leadership attention
• Build passion for data-driven decision making
• Upskill workforce in analytics
• Derive AI strategy from business strategy
• Define common metrics and incentives for whole organization
• No opt-out allowed for strategic targets
• Utilize defined KPIs for systematic implementation
• Take action - learn and iterate
• BUT understand that analytics is a journey
© Dain Studios 2016
22
Company Personas
© Dain Studios 2016
We have identified six company personas(A highly unscientific presentation)
Black Box Optimist Details, Details Pessimist
No Rush Covering our Backs Smart
© Dain Studios 2016© Dain Studios 2016
The Black Box Optimist
Company rationale:
“We are behind and need to do something quickly…”
”Machine learning is so complicated…”
“Let’s get a tool where we can dump our data in and get insights out – it will cost money, but then it’s done…”
Challenges:
• Overestimate the possibilities of technology
• Underestimate the impact on organization, required competences, process changes
© Dain Studios 2016© Dain Studios 2016
Details, Details
Company rationale:
“First, we need to have the business strategy, digital strategy, technology strategy, and marketing strategy in place…”
“We need detailed business case calculations for the next 5 years in place before we can start with execution…”
“All roles and the organizational structure need to be defined and approved…”
Challenges:
• Drowning in complexity
• Stagnate and lose time by answering questions you only will be able to address when you get going
© Dain Studios 2016© Dain Studios 2016
The Pessimist
Company rationale:
“We don’t have anything (no people, no data,...)…”
“These kinds of projects always fail (“the project that shall not be named”)…”
“Let’s start with something very small that doesn’t disturb our core business…”
Challenges:
• Hard to make a business difference as use cases siloed, manual, and small
• Company-level transformation not happening
© Dain Studios 2016© Dain Studios 2016
No Rush
Company rationale:
“At some point, we will probably need to ramp up our data capabilities, but our business is going well so let’s not rush anywhere…”
“Anyway, how would someone from the outside be able to tell us what to do? We are the experts of our business...”
Challenges:
• Past success is no guarantee of future results; total unpreparedness is dangerous
• Individuals fear that data and analytics will challenge their role / position
© Dain Studios 2016© Dain Studios 2016
Covering our Backs
Company rationale:
“Let’s hire McKinsey/BCG/IBM/Accenture/… because they know what to do…”
“If it fails anyway, we don’t get the blame as we used the big, expensive consultants…”
Challenges:
• Trying to outsource the thinking (and execution) instead of getting own hands dirty and leverage own domain knowledge
• Not learning as an organization
© Dain Studios 2016© Dain Studios 2016
Smart
Company rationale:
“We want to be leaders in digital transformation…”
“Our products and services need to be intelligent and adoptive…”
“Our sales and marketing processes need to be based on solid customer understanding…”
“We are not afraid of change; we embrace the journey…”
Challenges:
• Take the whole organization onto the journey
• Satisfy the stock market
© Dain Studios 2016© Dain Studios 2016
Key takeaways
1
2
3
Don’t outsource thinking – you need to define your Business Opportunities according to business priorities.
AI is a journey – you need to build a analytics culture across the whole company led by the leadership to make a difference.
People first – you don’t need to know exactly where to start but get a couple of good people and place them wisely to start specifying things
© Dain Studios 2016
31
Recruiting Data Scientists (and other data people)
© Dain Studios 2016© Dain Studios 2016
AI and Big Data mean new roles for a company
Analytics Strategist Data ScientistConsumer Data Privacy and Protection Officer
Data Steward/ Custodian
Solution Architect Data Architect Big Data Engineer Database Developer
Business / Data Science
Legal
Technology
© Dain Studios 2016© Dain Studios 2016
© Dain Studios 2016
DO
• Invest in people: recruit key persons and improve the analytics skills of business people
• Build passion for analytics at every level of the organization
• Find common ways of working together across organizational silos ensuring end-to-end delivery of results
• Define common metrics that guide strategic and operational decisions
DON’T
• Expect immediate ROI. Analytics transformation is a long-term effort
• Do not believe that technology resolves the business questions of your company
• Outsource data and analytics to one company function
• Allow individual departments to opt out of strategically important implementation targets
Do’s and Don’ts in AI Execution - Summary
© Dain Studios 2016© Dain Studios 2016
Ulla Kruhse-LehtonenCEO DAIN Studios Finland, Co-founder
[email protected]: +358 45 639 3125
Helsinki – Berlin - Munich