Strategic Insights and Clickworthy Content Development

Month: September 2017

Want to Succeed at Data-Driven Transformation? Start Slow

Data-driven transformation efforts often fail because companies are moving too fast or too slow. If they’re racing into it, they’re probably not getting the details right. If they’re trying to get everything into a perfect state before doing anything, they’re wasting precious time. Either way, ROI suffers.

If you want to strike the right balance, you have to think carefully about what you want to do and why you want to do it. You also have to move at an appropriate speed which means balancing short-term and long-term goals.

Start small

Boston Consulting Group recently published an in-depth article about data-driven transformation. In it, the authors raise a lot of good points, one of which is taking small steps rather than giant leaps.

Taking small steps enables businesses to succeed more often, albeit on a smaller scale. If the project is a success, the money saved or earned can be used to fund subsequent steps. Small successes also help keep teams motivated. However, executing too many small projects simultaneously can be counterproductive.

“The first mistake I see is doing 200 proof of concepts. The second thing I see is people start to do a pilot, even at scale, but [they think] first we need a developer to reinvent the system and then we’ll get the value at the end,” said Antoine Gourévitch, a Senior Partner and Managing Director at BCG. “I’d rather find a quick and dirty way to connect the IT systems and get [some immediate value] rather than doing a full transformation. You can have a transformation over three to five years, but at least I need to do the connection between my pilot at scale and the dedicated systems for the data platform that’s needed to be of value as we go.”

The third challenge is prioritizing pilots or projects. Value is the criteria there. Without a prioritized roadmap, “cool” projects may take precedence over projects that deliver business value.

Three steps to better ROI

BCG offers a lot of good advice in its article, not the least of which is breaking short-term and long-term goals into a three-step process that enables quick wins while paving the way to data-driven transformation. The three steps are:

  • Use quick wins fund the digital journey and learn
  • Design the company-wide transformation
  • Organize for sustained performance
“This WAS a different cybersecurity conference experience, and I really enjoyed all of the interaction and honest discussions.” (2017 Attendee) LEARN MORE

Within those three steps, BCG specified actions companies should take. Upon close review, it’s clear that some of the recommended actions, such as “fail fast,” apply to more than one step. If you read BCG’s article and ponder the graphics, it will get you thinking about how to scale success in your organization.

BCG also presents a five-level transformation model that includes vision, use cases, analytics, data governance and data infrastructure. Gourévitch said data governance tends to be the most problematic because data isn’t viewed as a corporate asset, so data owners may hesitate to share it.

Bottom line

Companies often move too fast or too slow when becoming data-driven organizations. When they move too fast, they can overlook important details that cause initiatives to fail. When they move too slow, they risk losing competitive advantages.

Somewhere between 200 pilots and one massive transformation effort is a balance of short-term and long-term goals, investment and ROI. The challenge, as always, is finding the balance.

Why Advanced Analytics Is in Your Future

Basic reporting and analytics are now competitive table stakes across industries. As 2020 approaches, more companies are using sophisticated algorithms to drive higher levels of efficiency, reduce costs and risks, drive additional revenue, improve customer experience and more. If organizations want to become truly agile in today’s dynamic business environment, they have to continually improve their operations and evolve the ways they’re using analytics.

“If you’re not using advanced analytics yet, you’re in trouble,” said Bill Franks, chief analytics officer at the International Institute for Analytics (IIA). “Twenty years ago, if you were doing some type of analytics you had competitive advantage. Now if you’re not doing analytics, you’re falling behind. If companies don’t push to adopt the new stuff, it’s going to become a problem over time.”

What Is Advanced Analytics?

Advanced analytics, like data science, lacks a standard description, although characteristically, it involves prediction. Deep learning, neural networks, cognitive computing, and AI come to mind because the algorithms have capabilities traditional input/output systems just can’t provide.

“What’s commercially possible to do has expanded significantly,” said Chris Mazzei, chief analytics officer at professional services company EY. “Decreasing technology costs and the explosion of data changes what’s possible to do with analytics, and [the possibilities] are growing every year. That, combined with competitive pressures means if you’re not looking for ways to reduce costs, enhance customer experience, create new products and services, if you don’t want to manage risks radically different and better, you’re in trouble.”

Most companies start with basic analytics and then increase the level of sophistication as they begin to realize the limitations of their existing systems. Disruptors are an exception because they use advanced analytics early on in an attempt to outthink and outmaneuver the existing players.

Whether your company is trying to compete more effectively or just stay relevant, advanced analytics is in your future, sooner or later. The question is whether your company will lead or follow. Either way, now is the time to learn all you can about advanced analytics so you understand what benefits it can drive for your company.

Even Small Businesses Should Care

Not so long ago, only large companies could afford the tools and specialists necessary to take advantage of advanced analytics. However, as more capabilities are made available through cloud-based services and as more of the complexity is abstracted, more businesses are able to advantage of advanced analytics without spending millions of dollars and hiring data scientists.

For example, lawn care aggregator site LawnStarterstarted using prescriptive analytics about two years after the founders defined the business concept. The initial goal was to decrease customer churn.

“We have a customer risk model and a provider risk,” said Ryan Farley, co-founder or LawnStarter. “We have thousands of lawn care providers in our system and the number of jobs they have ranges from tens to hundreds. Sometimes they take on too much. Before we had predictive analytics, we had to wait for the problem to become obvious.” Now LawnStarter is able to operate in a proactive way rather than a reactive way.

In all fairness, Farley wasn’t a typical entrepreneur. Previously, he worked for Capital One, which has been using predictive analytics since the 1990s to improve the ROI of its direct mail campaigns. When LawnStarter was founded, the founders wanted to do “cool stuff” rather than follow the traditional method of starting a company, building a product, and writing code. Fortunately, LawnStarter and machine learning platform provider DataRobot were part of the same Techstars accelerator program, so LawnStarter became one of DataRobot’s beta customers.

“We were like, ‘This is so cool! There’s predictive capabilities in our data sets!” said Farley. “We started out doing it for fun, but then we realized there was actually business to be had there. Shortly thereafter, we started investing in the data infrastructure to where we can compile our different data together and make sure everything we’re collecting is consistent and accurate.

Tips for Getting Your Company Started with Analytics

n today’s fast-paced economy, businesses need access to insights faster than before. While periodic reporting still has its place, organizations are looking for deeper and more timely insights that can help them make better decisions, cut costs, improve efficiencies, reduce risks and drive more revenue.

It’s pretty obvious from all the hype around the topic that many powerful things can be done with analytics, but it isn’t always obvious where one should begin. We asked some experts — they or their firms presented in the Data & Analytics Track at Interop ITX this month — where they thought businesses should start, and here’s what they had to say.

Prove value, not concepts

Some organizations spend too much time trying to get everything in a perfect state before using analytics, which wastes valuable time and overcomplicates what could be a simple beginning. The best way to start is to choose a project that has the potential to demonstrate value without requiring a lot of extra work or heavy investment.

“When we build something that returns value to the organization right away, people start buying in because they see the ROI and the other types of value it provides like efficiencies in the workforce, short time to solution or short time to value,” said Kirk Borne, principal data scientist at Booz Allen Hamilton. “It can be something simple.”

For example, a financial services company managed to save $1 billion simply by analyzing web clicks, Borne said.

“They already had web analytics in place. They just needed to pay attention to what the signals were telling them,” said Borne. “It doesn’t have to be a complicated model or involve complicated data to prove value.”

Analytics can start anywhere

Analytics can begin at any level in an organization, whether it’s an executive who wants an answer a strategic question or a line of business manager or staff member who needs to solve a tactical problem.

Five years ago, the Association of Schools and Programs of Public Health(ASPPH) assigned data analytics as a part-time job to its current director of data analytics and another employee. The association had been sending members periodic reports, but as the volume of data grew, it became obvious ASPPH could provide more value to its members with analytics.

“We started out small, providing a few dashboards to our members,” said Christine Plesys, director of data analytics at ASPPH. “They didn’t have to wait four months to receive a report.”

Now ASPPH is teaching its members about the best practices in data analytics and how to use the data ASPPH provides for strategic planning and internal benchmarking purposes. In addition, its data analytics staff has grown to four full-time employees

Get an executive sponsor

First efforts can be difficult to get off the ground if no one at the executive level understands the potential value of the project. To minimize that challenge, it’s wise to have an executive involved who will help the project succeed.

“One of the things some people miss is it’s not just a chief data officer or a data scientist you should be talking to,” said Booz Allen Hamilton’s Borne. “Sometimes it’s the chief financial officer or chief marketing offer because those people hold the purse strings on the kinds of investments that need to be made.”

Expect the unexpected

When people receive reports, they often have more questions that require a new report to be built. Analytics dashboards enable individuals to explore data in a more iterative fashion which needs to be considered when launching a first attempt.

“A lot of people think analytics is a new word for data warehousing or business intelligence and then they try to run their [analytics] project the same way,” said Karen Lopez, senior project manager and architect at data project and data management consultancy InfoAdvisors. “At a base level, you’re building a Q&A system because you don’t know [all of the questions] you’re going to ask.”

It’s also difficult to anticipate what unexpected circumstances might arise, especially when launching an analytics project without the help of an expert. For example, it may be difficult to get the necessary data from IT or from another department because they don’t want to share it.

People new to analytics also tend to overlook data quality. Poor data quality can cause spurious analytical results and perfect data quality is virtually unattainable. It’s wise to understand the tradeoffs between data quality and the time and expense it takes to get it into a state that’s “good enough” for the purposes it will serve. An expert can help you find that balance.

Bottom line

Too often, first attempts are derailed by overcomplicating the problem or attempting to solve a problem that is too complex to be solved well with existing team members and tools. The best first analytics project is one that can demonstrate value quickly and cost-effectively. If you succeed, it will be easier to make a case for follow-on projects. If you fail, you’ll learn a lot without wasting months or years and several million dollars.