1. October 2, 2013
DataWeek: Oh no, I’m running a data-driven cult!
hugeinc.com
info@hugeinc.com
45 Main St. #220 Brooklyn, NY 11201
+1 718 625 4843
2. Oh no, I’m running a data-driven cult.
Leala Abbott
Senior Content Strategist
Huge Inc.
Leala Abbott is a Senior Content
Strategist at Huge, helping to
shape, define and interpret
metadata and taxonomical needs
into content structures everyday.
She holds a Masters Degree in
Information Science from Rutgers
University.
3. “Story and Data need each other.
Data without Story is Mythology.
Story without Data is Propaganda.”
- Huge / Content Strategy Motto
5. What is propaganda?
“information of a biased or misleading
nature, used to
promote or publicize a
particular point of view”
Media Propaganda by: Trosious
6. Data as propaganda.
The big data method.
The scientific method.
hunches
refine
hypothesis
hypothesis
Gaming the system
results
test
results
test
7. What story are you telling?
Susceptive to fallacy.
Many clients use the end of the
quarter to crunch all the data
and re-enforce a narrative
with senior management.
Open to truth.
• Put data in the passenger
seat (not the drivers seat
or the trunk).
• Use data to guide, not just
narrate.
• Reporting needs to be
ongoing, fluid, and light.
• Leave the artfulness of the
quarterly report behind.
8. The question you have to ask yourself is
if data tells you something that goes
against the current gospel, do you share
these truths?
10. The path.
1
Have a goal.
2
Give instruction.
4
Ask technology.
Lack of the proper filtering mechanisms,
both from and human and tech
standpoint to abstract the signal from the
nose. The frequent metaphor people use
"is drowning in data”. Big data wont
magically solve your problems, you’ve
got to have a clear objective.
3
Balance cost / accuracy.
Everyone talks about the role of 3rd party
data targeting, personalization and
optimization but not all data is created
equal. We need to do a better job
balancing the cost of the data with actual
accuracy it provides.
They aren’t drowning in data, they are
waving for help. People need guidance and
proper insight into interpreting the data.
Some of the concepts used in attribution
are so technical and arcane that the
marketing lead simply doesn't know the
right questions to ask.
Data is being created at breakneck pace.
So correlating, interpreting and analyzing
it needs to happen just as quickly. This is
where the technology comes in as
humans can’t do it fast enough, or older
systems just aren’t structured to do it
efficiently. You need flexibility and
innovation.
11. Start converting.
Establishing new goals.
Your approach to data
should grow and change
as you do.
Find inefficiencies.
Your architecture (and
your org) should support
the efficient filtering of
data.
Create efficiencies.
Bring structure to your
data or provide an
interpreter.
Business change
Establish new goals, new teams, and more refined
and efficient data, to deliver the real truths to
your organization.
13. Turning data into information.
Planning.
Structure.
T
argeting.
Put together a strategic
plan, which can be used as
a guideline for your
approach to working with
data.
Audit to find LHF.
Whenever feasible give the
data deeper, semantic
meaning through
structures prior to it being
collected.
Don’t try and take it all in at
once, target specific data
sets the meet primary
goals, and objectives first
to start modeling.
Prioritization and
measurement.
Consolidation and aggregation.
Integration.
Gather and house the data
in a centralized repository.
Don’t have customers
emails across 4 databases.
Enable the data to be
integrated into business
systems and decision
making processes. End the
ownership fighting.
Prioritize within your
targeted data sets and
determine LOE. Start
small. Is it measurable is it
valuable?
Data tells a story, because of this it can also tell propaganda.
Gaming the system: statistics can be faked as noise can become signal through bogus selective-picking, and sampling.
Data shouldn’t be the last chapter in your story. Otherwise, you risk utilizing data to just support what has already happened (the good and the bad)Many clients use the end of the quarter to crunch all the data and re-enforce a narrative with senior management. Because of the artfulness of the reporting, this type of process takes up a significant amount time and energy from departmental senior management. This is just a flat out waste of resources - as an old NBC colleague of mine said: "Imagine how well rested and productive we'd all be if they just gave us the time off rather than spending 3 weeks a quarter developing the quarterly report". Which by the way, wasn't an exaggeration.The philosophical shift from that to using data as a form of guidance is a monumental one and shouldn't be underplayed.
Time to repent.
More data, shouldn’t equal more problems.
Turning data into information.
LHF low haning fruit, system with like strcutres or with some structure“Smart Data” means information that actually makes sense. It is the difference between seeing a long list of numbers referring to weekly sales vs. identifying the peaks and troughs in sales volume over time.Smart data is data from which signals and patterns have been extracted by intelligent algorithms. Collecting large amounts of statistics and numbers bring little benefit if there is no layer of added intelligence.