10.01.2015 Views

Accenture Technology Vision 2013

Accenture Technology Vision 2013

Accenture Technology Vision 2013

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Trend 3. Data Velocity<br />

<strong>Accenture</strong> <strong>Technology</strong> <strong>Vision</strong> <strong>2013</strong><br />

Your 100-day plan<br />

In 100 days, create a data velocity strategy to match the<br />

speed of your insights to the window of opportunity<br />

available to act on them.<br />

• Survey business units to determine where they have<br />

critical decision-making bottlenecks.<br />

• Determine which decisions are data-processing<br />

dependent.<br />

• Prioritize opportunities based on the cost-benefit<br />

tradeoffs for accelerating the decision. Speed<br />

costs money.<br />

• Create tactical deployments based on the prioritized<br />

opportunities with current tools and methods.<br />

• Data champions should update the data catalog to<br />

incorporate data-processing criteria.<br />

crucial for IT groups to still rely on<br />

non-real-time data where possible,<br />

blending fast and slow to solve<br />

problems cost-effectively. Of course,<br />

skilled developers already do this;<br />

for some, it harks back to their<br />

training in engineering. Enterprise<br />

systems are no different. You might,<br />

for example, precompute much of<br />

your customer analytics. The batch<br />

analysis from your weekly churn<br />

report may tell you that you’re<br />

in danger of losing a longtime<br />

customer, but discovering that<br />

that customer, already identified<br />

as at-risk, is currently browsing<br />

lower-cost service options on your<br />

Web site could give you the hook to<br />

find a way to keep from losing your<br />

customer altogether.<br />

As the challenges of accelerating<br />

data become more sophisticated—<br />

involving soaring volumes of<br />

unstructured data, for example—the<br />

trick will be to apply “hybrid insight”<br />

as often as possible. This calls not<br />

only for changes in architecture<br />

but for changes in skills as well. It<br />

requires that software-engineering<br />

leaders seek out and reward<br />

developers who demonstrate a<br />

definite “speed mindset.”<br />

Those who show an aptitude for<br />

blending real-time insight with<br />

batch insight—and knowing when<br />

to use each—will be extremely<br />

valuable. For example, take the case<br />

of a healthcare-insurance provider<br />

that has to meet a tight deadline<br />

on a government mandate to<br />

predict subscriber attrition. The IT<br />

group can run the work in batches<br />

to meet the compliance deadline,<br />

but over the long term, the group<br />

can use the insights from that<br />

work to understand and predict<br />

“attrition propensity” in every other<br />

interaction with insured patients and<br />

40

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!