Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
December 1st, 2017
BIG DATA & ANALYTICS
TURN INFORMATION INTO ACTIONABLE INSIGHTS
Marcus Hartmann
Bisnode Group Director AnalyticsManaging Director SWAN Insights
Less than 0.5% of all data we create is ever analysed and used.
Source: MIT Technology Review
BIG DATA, NOISE & UNLOCKED INSIGHTS
0.5%
BIG DATA DRIVES CHANGE AND…
Unstructured data from social networks, websites and forums.
Text, images, videos, links, unrelated and incomplete.
Requires unstructured database and new science.
…CREATES MANY OPPORTUNITIES
Big Data Use Cases 2015 – Getting real on data monetization
- 12 -
Figure 4: Which problems/challenges would you like to address with big data technologies?
(n=431)
IT (45 percent) and management (43 percent)
top the list of thought leaders and drivers be-
hind big data in enterprises (Figure 5). The re-
maining business departments are still very
passive as a whole and are driving the topic
considerably less often. We, too, can confirm
this based on experience from our own projects
with clients. In many cases, either IT uses and
suggests big data technologies as a solution to
address old or new challenges from the bottom
up or management has recognized the strategic
benefits of digitalizing and using data and is
driving the topic from the top down. But enter-
prises can only ensure the long-term success of
big data when other departments accept and
utilize it. There is obviously still work to do in
this regard.
Upon closer examination, the data shows that
management is a decisive factor for the pro-
gress of a big data initiative (
Figure 6). In 61 percent of companies where
management is the thought leader or driver be-
hind this topic, big data initiatives are already
integrated in business processes. If big data ini-
tiatives are in pilot stage or conceivable in the
future, management is the thought leader in
only 46 percent and 34 percent of cases re-
spectively. What is also striking is that respond-
ents named operational departments such as
sales more frequently when big data is already
firmly anchored in their business processes.
57%
55%
51%
50%
46%
31%
27%
23%
19%
6%
1%
Analysis of large volumes of data
Better or new data analysis possibilities
Building predictive models
Analysis of information from differentdata sources (polystructured)
Faster delivery of data for analysis
Acceleration of decision-making
Monitoring/Analysis of streamingdata/Complex Event Processing
Better cost-benefit ratio for analyticalenvironments
Automation of decision-making
There are no pressing challenges inour company
Others
Embeddedreal-timecognitivetechnology
APIintegration;
Prescriptiveanalyticalmodels;Automateddatapreparationprocesses;
Changeofbusinessprocesses;AdvancedAnalyticsasaService
…
One-OffexperimentalBigDataAnalyticsEco-System
Highperformancedatamanagement;Predictiveanddescriptivealgorithm
Semi-automatedpredictiveanalytics
Scorecardsanddescriptiveanalytics
2020
2000
CustomerProjects ProcessandSystemIntegrated
StructuredData UnstructuredData
PROJECTSCOPE
NEEDS VISION
CONSUMERS AND CORPORATES LEAVE DIGITAL FOOTPRINT EVERYWHERE.
Consumers and Company are online and leave information about themselves scattered everywhere on the Internet.
How can we find those isolated footprints and reconstruct a consistent profile?
This is where Contextual Intelligence helps.
A HUGE POTENTIAL LIES IN ONLINE DATA
CONTEXT IS WHAT TURNS INFORMATION INTO INSIGHTS.
Information is scattered everywhere, and unrelated. The key is to bring in the outside world after separating the wheat from the chaff.
… and to connect dots between data and contextualize information in order to deliver insights that truly matter.
BRING CONTEXTUAL INTELLIGENCE TO DATA
Worldwide news Social data Open data Closed data
Millions ofRSS feeds
Profiles, posts,connections
Nearly everythingcan be found
Purchased ornegotiated data
CONTEXT
BENEFIT FROM THE POWER BIG DATA
ENRICH
Context mining Artificial Intelligence Network ScienceCreate new information
by connecting dotsLeverage data to let
machines take decisionsUnreveal hidden insights
through mathematics
QUALIFY TARGET DETECT REFINE PREDICT REVEAL CLUSTER EXPLORE
HOW CONTEXTUALINTELLIGENCE CAN HELP?
CONSUMER DATA ENRICHEMENTWITH ONLINE INFORMATION
Matching and data enrichment based on web social profiles.
1. Professional profiles
2. Social profiles
3. Affinities
4. Clustering enrichment
5. Enrichment from company websites
PREDICT CONSUMER MOVES FROM CALL DATA RECORDS
Customer challenge:
Wanted to know if customer will move to another location and change provider.
Bisnode solution:
• Analyzing numbers calling and numbers called creates a network;
• Network science predicts future customer moves from one city to another, based on network calls.
RANKS LEADS FROM CRM AND IDENTIFIES THE MOST VALUABLE CUSTOMERS
Customer challenge:
Thousands of leads in CRM – too much work to call them one by one.
Bisnode solution:
1. Leads ranked and sorted by the propensity to buy the car;
2. Lead's relationship with competitive car models;
3. Detailed characterization and enrichment of existing CRM contacts.
PREDICT RESIDENTIAL CUSTOMERS AT RISKOF CHURNING
Customer challenge:
High residential customers churn rate –wanted to know which ones will be at risk in the future.
Bisnode solution:
Customer ranking sorted by the propensity to leave the operator;
Customer data enriched by web information and characterized in classes;
Advices on the best way to run preventive campaigns to reduce organic churn.
THANK YOU
Marcus HartmannBisnode Group Director AnalyticsManaging Director SWAN Insights