At the end of March 2014, people across
various industries gathered in Boston for the AnalyticsWeek Big Data &
Analytics Unconference (http://boston.analyticsweek.com/). The overarching purpose of the
weeklong event, as shared by the organizers, was to spread information and
capabilities across the Boston Big Data and Analytics community.
A few of my Centric Boston colleagues and I
made it a point to attend - some of it at least. Each day of the conference had a different
theme and we soaked in the scenes on day one – Big Data, day three – Insurance
& Financial Services, and day four – Marketing. Although we attended discussion on different
topics, several reoccurring themes emerged. After comparing notes, here’s our recap of the
top four themes:
1)
Realizing the promise. (While companies are recognizing the
importance of Big Data, questions still remain about how best to take action
and what kind of impact we can hope to see.)
Expectations for Big Data across industries remain huge. Progress is absolutely being made, including great work from start-ups like DataRobot (http://www.datarobot.com/) and Nutonian (http://www.nutonian.com/) along with advancements from established players like Wayfair (http://www.wayfair.com/) and Oracle (http://www.oracle.com/index.html). But we also heard that to get high impact outcomes and mass proliferation, there will need to be transformational thinking and broader delivery. Two distinctive analogies about Big Data adoption were used by keynote speakers:
Expectations for Big Data across industries remain huge. Progress is absolutely being made, including great work from start-ups like DataRobot (http://www.datarobot.com/) and Nutonian (http://www.nutonian.com/) along with advancements from established players like Wayfair (http://www.wayfair.com/) and Oracle (http://www.oracle.com/index.html). But we also heard that to get high impact outcomes and mass proliferation, there will need to be transformational thinking and broader delivery. Two distinctive analogies about Big Data adoption were used by keynote speakers:
1.
From Paul Sonderegger (Oracle)
http://merage.uci.edu/ResearchAndCenters/CDT/Resources/Documents/[197]Paul%20Sonderegger_bio.pdf: Where
we are now, with Big Data is analogous to where the world stood just as
electricity was invented. The invention of electricity was
revolutionary, but not every house on the block had it, initially, and many,
many uses for it had not immediately been conceived. So, we’ve done some things with Big Data, but
reality is that while most companies have expectations that they will be using
Big Data over the next few years, most also don’t have a clear vision of
exactly how new capabilities will transform their business.
2.
From
Chris Lynch (Atlas Venture) http://www.atlasventure.com/team/chris-lynch: The world-changing invention of the Internet
came first, but to realize sweeping impacts and adoption, the subsequent
inventions of the World Wide Web and Browser were required.
The focus on building momentum is undeniable,
with enterprises spending about $8M on average on Big Data in 2014, according
to Louis Columbus citing IDG - http://www.forbes.com/sites/louiscolumbus/2014/01/12/2014-the-year-big-data-adoption-goes-mainstream-in-the-enterprise/. Development
efforts will continue to mature and evolve into repeatable value-based
activities. Leaders will emerge!
2) Thinking in terms
of a data stream of consciousness. (Successful
companies will utilize their own data, as well as data from third-party providers,
to deliver context throughout the endless data collection journey.)
On both day one (Big Data) and day three (Insurance & Financial Services), points were made on what constituted a firm’s Big Data: Successful companies harness their own data, combine it with external data, and use it all to adapt to a world that’s going to change more and more quickly, leaving slower companies in its wake. When a question was posed on the Financial Services and Insurance panel about utilizing proprietary data vs. buying data from outside the firm, the simple answer was, Both are required. Financial Services and Insurance companies aggregate data directly from their customer (demographic, preference-based, transaction history, navigation of firms’ digital media) but they are also big buyers of 3rd party data and will continue to be. Some examples include census data and weather data (especially for P&C Insurers).
On both day one (Big Data) and day three (Insurance & Financial Services), points were made on what constituted a firm’s Big Data: Successful companies harness their own data, combine it with external data, and use it all to adapt to a world that’s going to change more and more quickly, leaving slower companies in its wake. When a question was posed on the Financial Services and Insurance panel about utilizing proprietary data vs. buying data from outside the firm, the simple answer was, Both are required. Financial Services and Insurance companies aggregate data directly from their customer (demographic, preference-based, transaction history, navigation of firms’ digital media) but they are also big buyers of 3rd party data and will continue to be. Some examples include census data and weather data (especially for P&C Insurers).
All of this data, both structured and
unstructured, will help deliver value to the firm and its customers assuming
context and relevance can be established.
Think predictive analysis, including highlighting actions that can drive
a customer to better personal outcomes.
Speaking of customer value…
3) In service to our fellow
man*. (When it comes to defining objectives for the use of Big Data,
customers should come first. Consumer
demand for transparency in actions and personalized experiences are only going
to increase.)
To achieve critical adoption and value, Chris Lynch clearly put forth the notion of simplicity. “We need data science that can be consumed by more than data scientists” (more on these scientists later). He also highlighted that there are very real privacy concerns and natural resistance to change, so use of customers’ personal data has to deliver value and be transparent for those customers, from the start. With that, there is a positive cycle that can be established - give me some data and I’ll provide you with answers to challenges you face in your daily lives (e.g. furniture purchasing decision, investment choices, insurance coverage, etc.). This was especially clear on day four (Marketing) - value comes from analyzing customer behavior and buying patterns across all channels (omnichannel) to ensure a personalized shopping experience for every customer. This draws on customer experience principles and can help in customer retention (treat people right!). Of note: Measurement on mobile is a lot more complex than measuring traditional website activity.
To achieve critical adoption and value, Chris Lynch clearly put forth the notion of simplicity. “We need data science that can be consumed by more than data scientists” (more on these scientists later). He also highlighted that there are very real privacy concerns and natural resistance to change, so use of customers’ personal data has to deliver value and be transparent for those customers, from the start. With that, there is a positive cycle that can be established - give me some data and I’ll provide you with answers to challenges you face in your daily lives (e.g. furniture purchasing decision, investment choices, insurance coverage, etc.). This was especially clear on day four (Marketing) - value comes from analyzing customer behavior and buying patterns across all channels (omnichannel) to ensure a personalized shopping experience for every customer. This draws on customer experience principles and can help in customer retention (treat people right!). Of note: Measurement on mobile is a lot more complex than measuring traditional website activity.
If firms want to maximize the impacts they can
achieve from Big Data and Analytics, they are best served to think beyond their
firms’ needs and desires and instead think in terms of how to create value for their
customers (businesses or consumers) with Big Data.
* This phrase was
borrowed from an essay by Albert Einstein – An
Ideal of Service to Our Fellow Man - http://www.npr.org/templates/story/story.php?storyId=4670423
4) For many, the data
scientist is still a mythical creature. (A consensus among experts seems to
exist, that the ideal resource has a unique skillset & is difficult to
find.)
The Data Scientist. Who is this? How do you find them? What can you expect from them? This was a central topic during this unconference, and has been a topic at other events we’ve attended. Defined as, a resource who can combine technology ability (Hadoop, NoSQL, etc.) with statistical aptitude (think quant or actuary), sprinkle in some business expertise or acumen, and round it out with the ability to communicate effectively. No wonder why there is a supposed shortage of these experts (The United States alone faces a shortage of 140,000 to 190,000 people with analytical expertise – per McKinsey http://www.mckinsey.com/features/big_data). Nonetheless, the speakers worked to try and answer how you identify and test these resources prior to hiring them. Some sound advice included paying particular attention to what people do as hobbies – do they have Big Data blogs? Do they opine on Big Data challenges via networking groups? Also, specific light problems can be delivered, via the interview process, to test some aspects of problem solving.
The Data Scientist. Who is this? How do you find them? What can you expect from them? This was a central topic during this unconference, and has been a topic at other events we’ve attended. Defined as, a resource who can combine technology ability (Hadoop, NoSQL, etc.) with statistical aptitude (think quant or actuary), sprinkle in some business expertise or acumen, and round it out with the ability to communicate effectively. No wonder why there is a supposed shortage of these experts (The United States alone faces a shortage of 140,000 to 190,000 people with analytical expertise – per McKinsey http://www.mckinsey.com/features/big_data). Nonetheless, the speakers worked to try and answer how you identify and test these resources prior to hiring them. Some sound advice included paying particular attention to what people do as hobbies – do they have Big Data blogs? Do they opine on Big Data challenges via networking groups? Also, specific light problems can be delivered, via the interview process, to test some aspects of problem solving.
Another way to think
about this dilemma comes from Jeff Kanel, Centric’s National Business
Intelligence Practice Lead. Jeff says that it makes sense to analyze the skillsets
needed for specific objectives and consider building a team of complimentary resources
vs. trying to find all the skills in a single resource - http://centricconsulting.com/data-scientist-sightings-will-mostly-be-proven-a-hoax/. In fact, Centric is currently expanding its
national analytics practice to fulfill current needs in this space.
So where does that
leave us? The takeaways from this conference and from our own experiences point
to the fact that getting a handle on Big Data is important – expectations are
emerging both from corporate leaders and from consumers. How companies specifically use the information
Big Data provides, such as building teams, interpreting the data and applying
that knowledge to enhance the customer experience, still remains to be seen. Companies that invest the time to figure out
such questions now will be leaders of tomorrow.
Centric Boston
colleagues Jeff
Aalto, Errol
Yudelman, and Brian
Sedor collaborated on this piece. http://centricconsulting.com/boston/
No comments:
Post a Comment