dynamics day 2015: systems of intelligence in action
TRANSCRIPT
Dynamics Day 2015
Systems of Intelligence in Action
Experimentation!=
Learning From Mistakes
1MALLEABLE: It must be cost effective to change the way that the system or process behaves. Ideally for just a subset of usage or users.
WHAT WE NEED
2
3
OBSERVABLE: The system or process must be sufficiently well instrumented such that we have the information available to test the hypothesis.ENGAGED: The system or process must be being used at sufficient scale and in a sufficiently ‘real’ environment for any results to be significant.
An experiment is a procedure carried out to verify, refute, or establish the validity of a hypothesis… Experiments vary greatly in goal and scale, but always rely on repeatable procedure and logical analysis of the results.
SYSTEMS OF EXPERIMENTATION
MALLEABLE
OBSERVABLEENGAGED
Experiments
MALLEABLE
OBSERVABLE
ENGAGED
“arguments about whether or not a feature idea is worth doing or
not generally get resolved by just spending a week implementing it and then testing it on a sample of users, e.g., 1% of Nevada users.”
There are over 500 different K-Cups
There Are Cartridge Loaded Software Systems
Microsoft Azure Demand Center
Internal Systems & Product/Platforms
Dynamics CRMMarketo
Web Email Call Center
Data Warehouse
PowerBI
SystemsOf
Record
SystemsOf
Experimentation
SystemsOf
Engagement
MALLEABLE
OBSERVABLE
ENGAGED
Storage is Cheap…
…Data Is Not
Smart Appliances (Think Printers + Ink)
Existing ‘Dumb’ Product
SensorTagIoT Suite
Smart Screen Mobile Apps(s)
Raw Data Store
???
Personalization Platform
PowerBI
SystemsOf
Experimentation
SystemsOf
Engagement
MALLEABLE
OBSERVABLE
ENGAGED
“…Our tool works best when you’ve got at least 1 million or so
app installs to work with…”
Towards Systems
of Intelligence
Flipping The Model
Inputs OutputsIT System
ExpertiseInterpretationExperimentation
Flipping The Model
Inputs OutputsIT System
ExplanationExperimentation
Automation
Models
Machine Learning
Example:Building SystemsProvider (~HVAC)
Much of the Tooling is Free
Is Becoming More Widely
Available
Is Becoming Easier to Use
And Is Entering The Public Consciousness
And can be crowd-sourced
In Boots & All
Use Cases We Are Seeing
Big Data Use Cases, BARC Research 2015
STOREOPTIMIZATIO
N
TRACKING, TARGETING
SEGMENT OF 1SMART
PRODUCTSSMARTASSETS
PRODUCT IMPROVEMEN
T
CUSTOMERACQUISITION &
RETENTION
USER ANALYTICS
&OPTIMIZATiON
SERVICE/B2B RETAIL SUPPLY CHAIN
FRAUD&
COMPLIANCE
• Embrace Experimentation• Malleable, Agile, Cartridges• Observable; Capture More Data• Engaged; Ship Early & Learn
• Systems of Intelligence• Doing Data Science <- Easier• Knowing What We Can Do <-
Hard• Analytics Inside Your
Organization• Survey for Opportunities• Most CRM Systems• Customer Facing Apps/Sites• IoT