Lessons from the front lines


By Chris Brahm and Lori Sherer

Almost every executive team we know is grappling with a big question: How can data and analytics create real value and protect us from disruption? After all, no CEO wants to be “logged in to Netflix” or “to Amazon”.

As business data skyrockets, signs of analytics-driven innovation are appearing across industries, and the application of machine learning, robotics and automation is quickly becoming a reality. Even the most analog companies scramble to invest in analytics. Their hopes are high, their investments substantial; but the results have been inconsistent at best.

In a recent Bain survey of 334 executives, more than two-thirds said their companies are investing heavily in big data. Unsurprisingly, 40% expected a “significantly positive” impact on returns, while 8% predicted “transformational” results (see figure).

Despite these high expectations, 30% of these executives said they did not have a clear strategy for integrating data and analytics into their businesses. In our experience, too many companies focus on investing in the technology and talent associated with advanced analytics, without thinking about the broader changes they will need to make to fully deploy analytics and achieve results. favorable.

Data and analytics advocates love to quote how Netflix skillfully uses analytics to personalize the user experience and secure some of the most popular entertainment content available today, raising its profile in the market. Incumbents are also using advanced analytics to transform their business models. Think about progressive insurance and its pluggable car devices which allow customers to earn discounts for safe driving, or the colorful MagicBands of Walt Disney World, which have transcended their mission of following customers to serve as Symbols suitable for Instagram a special holiday just waiting to happen.

These success stories leave late users wondering what big steps they can take to catch up. Many will rush to invest in the latest providers of analytics software and infrastructure and hire data scientists, but the ultimate winners will align those investments with their strategic and organizational needs in ways that lead to action and results. For a business trying to figure out how to be successful, here is what we recommend:

Put business science ahead of data science. A company’s advanced analytics goals should mirror the larger goals of the business, enabling it to amplify its most profitable products, services, and processes. Coca-Cola, for example, uses sophisticated social listening tools to spot influencers who could help the company promote its signature brand to key customer groups. Healthcare providers have deployed predictive analytics to guide preventive care, leading to better patient outcomes and lower costs.

Design scans with the “last mile” of adoption in mind. The best analytics solutions emerge when data scientists and business stakeholders work together, define the requirements for success early on, and keep end users at the heart of decisions. As the team makes critical design choices, such as the right analysis method, members will need to consider how end users will act on those results. Unstructured data and machine learning are at the forefront today, but it’s not always the most intuitive options for frontline workers who manage nuanced customer situations under tight deadlines. Sometimes analyzes just hinder organizational dynamics. In many retail organizations, for example, merchants regularly replace the sophisticated assortment algorithms of local stores. This may be the right answer in some cases, but if an organization is to invest in analytical innovation, it must anticipate and plan for potential obstacles to adopting the outcome of those investments.

Look far beyond traditional analytics. Companies have long relied on structured business data from their core systems and business sources to inform their decision making. Today, more companies are experimenting with unstructured data from social media, web scraping, customer interaction transcripts, log files, sensor data, images and other content accessible to the user. public. Insurers, for example, are tapping into new sources of data to help them assess risk, just as companies in the sharing economy are using social media and web data to identify customers who could harm their brands. Advanced machine learning techniques also help companies automate even sophisticated processes that only knowledge workers previously handled. Simply put, the bar of what constitutes competitive analysis capability in most businesses is rising.

Find the shortest path from insight to action. More data isn’t necessarily better. In fact, it is often the opposite. Rather than collecting more data, most businesses would gain more from maximizing the performance of their existing data and becoming agile enough to act on information quickly. After all, value comes from the action, not the entry. To illustrate the difference, compare a large airline that sends the same generic customer survey to its most valuable customers after every flight with another that uses its operational data to identify customers on a delayed flight so they can send an apology. preventive measures and a coupon. The old airline collects the same performance data; it’s just not using it. The latter mobilizes existing data to build loyalty.

Test, learn and iterate. The smartest businesses don’t wait to have the perfect analytics solution in place. They go out and pilot their new approaches on real customers and processes, even if their tools are barely viable, and then they are constantly perfecting them. This agile approach represents a huge change for companies accustomed to slower, waterfall-based processes that strive for “perfect” analytical results, a goal that is nearly impossible on the first try. Algorithms are never perfect; they are designed to adapt as new information emerges. Consider how far ecommerce recommendation engines have come. A few years ago, these engines served as unnecessary toy recommendations to a customer who had recently purchased a doll for a friend’s child’s birthday party. Companies have spotted these missed opportunities, updated their algorithms to eliminate one-time purchases, and can now recommend products that more accurately reflect customer interests, increasing the odds of a sale.

Manage the transition of advanced analytics with a clear goal. Advanced analytics will only lead a business if it doesn’t have a solid operating model to link strategy and execution across departments and functions. Senior managers need to define who does the work and how, and what capabilities the business needs, both in terms of talent and technology. Leaders will need to consider the extent to which they centralize their analytical capabilities, governance and operations. They will weigh the potential use of frontline data against the investments in technology and data infrastructure that these efforts require. And they will revisit these questions as the needs of the business evolve. Even analytics leaders like Facebook and Google have changed their operating models over time, adjusting the level of leadership and activity centralized versus distributed as their needs change.

The amount of data available is expected to nearly double by 2020, when it will likely reach 44 zetabytes, according to IDC Digital Universe. The scale, scope, and complexity of data and the companies that use it have bypassed the ability of humans to process it without intelligent automation. Advanced analytics just isn’t an option, it’s a must for any large organization.

An ecommerce CEO summed it up: “We’re a small business, but we have 30,000 SKUs. Just to price them smartly, and not by the same rudimentary “cost-plus” factor, it would literally take up 100% of our staff’s capacity. Yet with advanced analytics, I can price a SKU based on how many competitors offer that SKU, its price, its cost, and we can even factor in past behaviors and the value of an individual customer who could look at that particular SKU, and it’s almost fully automated. These breakthroughs in analytics have a huge economic leverage effect on my business. “

Businesses must respond to the imperative of advanced analytics. This means closing any results gaps and reducing the “measurable impact time frame” of their analytical investments. With a pragmatic approach, companies can harness the full potential of advanced analytics.

Related: With Advanced Analytics, it’s people (not data) that stand in the way of change

Chris Brahm is a partner who leads Bain & Company’s global advanced analytics practice. Lori Sherer is also a partner in the Advanced Analytics practice. Both are based in San Francisco.


Comments are closed.