Professional Documents
Culture Documents
HBR Alteryx Report 6
HBR Alteryx Report 6
sponsored by
SPONSOR PERSPECTIVE
Rick Schultz
Senior Vice President, Marketing
Alteryx, Inc.
VALUE DRIVEN
Its easy to see why data analytics has become so important. LoB managers are increasingly being
asked to deliver more value from their organizations data and associated systems. Analytics software helps them use data for business decisions that can lead to a better customer experience,
more product and service innovation, optimized business processes and, ultimately, competitive
differentiation.
Using data-analytics tools can also help organizations gain three important benefits: lower labor and
IT costs, higher revenue, and increased employee productivity, according to technology research
firm IDC. Thats an extra benefit organizations will get over their less analytically oriented peers,
says Dan Vesset, IDCs VP of business analytics and big data research.
Yet two serious challenges have so far hampered LoB decision makers from realizing data analytics full potential. First, many LoB data analysts are still burdened with old-school legacy tools
and processes. They begin by manually poring over Excel spreadsheets, Access databases, and
printed reports. Then they do rudimentary data manipulation and reporting using solutions that
werent built to access multiple data sources. Finally, they share their reports via email.
Second, many of todays tools arent able to provide true self-service. Because these tools were not
designed to be used by LoB analysts, users can find themselves relying on a virtual army of data
specialists and data scientists, often working under the auspices of a centralized analytics group,
to prepare, blend, and analyze dataor even to do the actual reporting of what they believe are
the key insights. Worse, the survey finds, few decision makers are truly satisfied with the tools
they currently have access to. figure 1 Of course, all this means wasted time and expense for both
LoB managers and their organizations alike.
Whats needed are data-analytics tools that are far easier to use. That way, they can be deployed
by both nontechnical managers and LoB data analysts. (The latter are more familiar with the
nuances in the data and the day-to-day business problems at hand.) These tools also need to be
highly repeatable, so that initial setups can be used by others, and extremely scalable, to process
the enormous volumes of data from both inside and outside the organization that todays managers and data analysts must analyze.
FIGURE 1
HIGHLY SATISFIED
39
29
Data preparation/blending
42
25
Data integration
42
42
Advanced analytics
57
Data visualization
37
84
Spreadsheets
34
USING: Percentage of respondents who say they use these specific tools; HIGHLY SATISFIED: Percentage who say theyre
very or extremely satisfied with the tools ability to provide the information they need for decision-making.
SOURCE HARVARD BUSINESS REVIEW ANALYTIC SERVICES SURVEY, MARCH 2016
HELD BACK
One big challenge for LoB decision makers and analysts is that, as a result of working with spreadsheets and other manual tools, they often glean information by gut reaction. Thats because
older tools, unlike todays more advanced tools, cannot dive deeply into data and allow users to
easily spot trends across disparate systems. Instead, these older tools offer data results that are
relatively rudimentary.
To be sure, self-service data-visualization tools are now available, and many are somewhat helpful for eyeballing localized data compiled as charts and graphs in dashboards. But they were not
meant to delve into mountains of raw data across multiple storage areas. And they certainly cant
handle much-needed data cleansing, preparation, and predictive analysis.
As a result, data analysts in business can find themselves spending up to 90 percent of their time
on ensuring the quality of their data, says Ray Wang, CEO of Silicon Valleybased researchers
Constellation Research. Thats because these managers must ensure that their data is relevant,
cleansed, and prepared before they can spin it into data-visualization dashboards. And in most
cases, they must do this manually. This leaves only 10 percent of the time for discovering business patterns and insights, Wang adds.
Another limitation of todays analysis tools is that theyre unable to effectively work with data
from multiple sources. For todays decision makers and data specialists, thats a big issue. They
need to be able to access data from multiple sources, including internal storage locales such as
data warehouses, CRM and ERP systems, and myriad Excel and Access files. But these managers
also need to analyze data from the growing number of external data sources, such as social media
and cloud-based platforms.
Yet another challenge highlighted by the survey: more than 60 percent say they depend on data
analytics specialists and data scientists for their data-analysis work. In addition, many respondents say theyre dissatisfied with both the speed of these specialists responses and the quality
of their work. figure 2
Unfortunately, this is an old, tired story, says Wang. He notes that data analysts have for years found
getting access to enterprise data sources with help from IT to be a time-consuming process. Often,
data analysts find themselves waiting in line for attention, he explains. Adding new requests for
additional data sourcessuch as external social-media dataonly makes matters worse. When
that happens, Wang says, IT needs to apply new rounds of data cleansing, transformation, and
integration to these sources.
FIGURE 2
38%
62%
NO
YES
19
Ability to compile information from inside/outside the organization
19
Speed
31
Quality of work
SOURCE HARVARD BUSINESS REVIEW ANALYTIC SERVICES SURVEY, MARCH 2016
IDC analyst Vesset says this issue is likely to only get worse in the future. The current shortage of
skilled and qualified data staff will not only persist, he says, but also extend to include data architects and data-management experts. While colleges and universities are trying to help by offering
more data classes, Vesset adds, it will take some time before the new pool of talent becomes fully
integrated into the labor force.
All this leads LoB managers and data analysts to be disappointed with their current data analytics process, and to seek tools that are better suited to their needs, rather than tools designed for
data scientists. As the survey shows, a majority of respondents put a high value on ease of use,
accessibility, speed, and other user-friendly characteristics. But unfortunately, no more than a
third say theyre getting these features from current tools. figure 3
The need for customization is another limitation of many current data analytics tools. LoB managers often need to customize existing reports to analyze a different geography or time frame, or
make other tweaks. In most cases, this requires the original data analyst who developed the report
to take time away from his or her regular work and customize the report. While this may not be
FIGURE 3
59
EXCELLENT
23
Scalability
69
25
Repeatability
69
26
Speed
75
Accessibility
32
79
Ease of use
35
IMPORTANT: Percentage of respondents who say these characteristics are either very or extremely important for their
decision-making; EXCELLENT: Percentage who say the data analytics tools they use are either very good or excellent
at these characteristics.
SOURCE HARVARD BUSINESS REVIEW ANALYTIC SERVICES SURVEY, MARCH 2016
a big deal for the occasional customization, it can become problematic for any data analyst who
has to support a large number of LoB managers, or who works with managers needing frequent
customization. Thats why its somewhat disturbing that fully half the survey respondents say
they customize analytics reports for others somewhat frequently, while fewer than 10 percent
say they never do it. figure 4
A senior analytics researcher at a major automaker, says hes well aware of this kind of extracurricular customization work. He spends many overtime hours on just such tasks. Whats more, he
believes his time could be spent far more productively. After a data project is completed, for example, the researcher wishes that a retroactive roundup could take place: Nobodys going, Lets
look backward in time to evaluate what we did right and wrong, and whats best going forward.
FIGURE 4
51%
SOMEWHAT FREQUENTLY
23%
VERY FREQUENTLY
18%
NOT FREQUENTLY
8%
RARELY/NEVER
SOURCE HARVARD BUSINESS REVIEW ANALYTIC SERVICES SURVEY, MARCH 2016
SELF-SERVICE EMPOWERMENT
Its clear, then, that legacy data analysis tools are becoming prohibitively time consuming and
costly. Whats needed are solutions that are automated, easy to use, repeatable, and scalable
enough to process the enormous volumes of data companies now draw from multiple sources,
including data streams from both inside and outside the organization.
Self-service data analytics tools are the answer. This technology can empower LoB data analysts
to take charge of the entire analytical process instead of relying on data specialists from IT to
create the critical tools they need or to do the analysis work for them. Data managers using these
tools can now deploy self-service data solutions that easily prepare, cleanse, blend, and analyze
all relevant data. With these self-service tools, you dont have to deal with IT at all, says Wayne
Eckerson, president of research and consulting firm Eckerson Group. While IT still needs to provide governance, at least for now, data managers dont need to wait in line for IT to clear a data
mart for them.
As a bonus, managers with these tools can do all this work on their own, without any need to code.
And they can glean data from virtually any available source. This also empowers them to share
their analytic results across the business with an easily accessible graphical user interface (GUI)
environment. They can make more comprehensive and insightful business decisions at speed,
ultimately keeping their companies highly competitive.
ON CLOUD NINE
How much do users like these self-service data analytics tools? A lot. Preethi Elango, an enterprise
data strategist at Southwest Airlines, is so high on the technology, she says theres a self-service
movement within her company.
As Elango explains, self-service data analytics tools give Southwests data managers more ownership and faster access to their data than ever before. Our self-service tools are user friendly and
intuitive, she says. And they empower users to integrate across multiple data sources to discover
business information and innovate with their data.
Elango adds that Southwest Airlines enterprise data is siloed, disparate, complex, and in varying
degrees of messy and clean in data-storage areas across several different platforms. Our data
is growing every year, she adds, and our business users have been struggling with the inability
to quickly consume, integrate, cleanse, and analyze all this data. She adds that her team wasnt
thrilled with the data analysis tool it was using, mainly due to its limitations in data integration
and manipulation.
Thats because those legacy self-service data analysis tools simply cannot perform the cleansing,
preparation, and integration required to generate the useful business results that analysts need
for making the best business decisions.
Before Southwest Airlines deployed its self-service solution, the overall process was extremely
time consuming. It also made outcome-sharing quite difficult. But now, Elango says, having such
powerful self-service data analytics tools has created a faster path from data to insight.
REVOLUTIONARY
The automakers senior analytics researcher likes self-service data analytics too. In fact, he calls
the technology the new software revolution for data. Thanks to a new self-service data analytics tool, his companys LoB managers can now answer business questions much more quickly
and easily than ever before, and with little of their older systems complexity. Whats more, LoB
managers get their data results all in one place via highly drillable and customized dashboards.
As a result, the researchers data team finds less need for the legacy data-visualization platforms
its used for years. He adds, Its phenomenal.
As an example, the researcher cites a recent project that evaluated his companys global travel
expenses and itineraries. The goal: discover insights to cut travel costs while also making travel
more efficient. The projects scope was massive, consisting of more than 13 million expense records
collected over the past eight years. To dig in, the team began by methodically inputting the booking data, expense data, and organizational data, such as which divisions traveled the most. Then
they deployed the legacy tools at hand. In all, it took about three months before the team could
create the first reports for LoB managers. While they were able to add new insights on price optimizations and more effective travel outcomes, the whole project was a grind, the researcher
says, adding, It took its toll on people.
Along the way, the automaker installed a new self-service data analytics tool, and the researcher
used it to rewrite the program for independent verification. It took him less than one day.
Now, when new travel information is input, its automatically updated by the new system. I have
a highly efficient system thats repeatable for use in the data division, the researcher says. I can
call up all this data by pushing a button and have the up-to-the-minute status and a new set of
results. And its all in dashboards that are accessible for our LoB managers and business leaders.
be data driven, he believes, making it essential for organizations to build data-driven cultures that
promote collaboration between IT and the business. This will empower everyone with shared environments and connected tools, he says, so they can all work together and learn from each other.
Not to mention that through 2020, IDC predicts spending on self-service and related data tools
will grow 2.5 times faster than will traditional IT-controlled tools for similar functionality.
Self-service data analytics has supplanted IT-driven applications, according to researcher Eckerson.
He adds that the shift could be as big as the social revolution that occurred when affordable cameras shifted from black-and-white to color. In data delivery, he says, were moving from a
monochrome world to a Kodachrome one.
Leading-edge data managers know this well, and theyre swearing by their self-service data
analytics tools. Many have already left the black-and-white world far behind. When I got my
self-service solution, the world opened with an unlimited possibility for solving problems, the
automakers researcher says. Now, if someone took my self-service data analytics tool away from
me, Id go buy it myself. Seriously. I have to have it.
JOB FUNCTIONS
SENIORITY
REGIONS
hbr.org/hbr-analytic-services