"Big data” frequently means different things in different discussions. Many times, people talk about characteristics of the data, such as the volume, velocity, va¬riety, variability or characteristics such as the complexity or unstructured nature of their data. Some corporate leaders also worry also about the “big investment” of big data, mixing related subjects such as IT infra¬structure, cloud computing, machine learning, new visualization techniques and new technologies like Hadoop. At Alorica, the data science team is focused on discovering and unlocking the “big potential” within its big mountain of customer engagement data. Right about now, you’re probably asking yourself, “Ok, but what is Alorica?”
Alorica is the largest provider of customer engagement solutions to the U.S. market, and third largest in the world. The company employs 100,000 team members worldwide, providing customer care, technical support and a wide variety of other services to Fortune 500 clients. By analyzing 1.7 million customer interactions daily, Alorica is transforming itself in ways that delight clients and their customers.
“What’s in It for Me,” you say?
It’s the deep insights and stories buried within the data that excite this team, such as the ability to build Silicon Valley-style predictive models for predicting customer behaviors, dramatically improving operations and creating methods of engaging with brands that customers embrace. Under the hood, the team builds and employs machine learning propensity models, similarity indexes and artificial intelligence models to dynamically route calls, guide operations and inform decision-making. This is the big investment Alorica has made towards disrupting an industry and accelerating growth—all in the daily pursuit of realizing its brand vision to make lives better, one interaction at a time.
Where to start?
So, how could your company embark upon a similar data science journey? The first step is understanding how analytical models various “data products” are similar in their ability to self-adjust and improve with more data. For example, Internet search responses are a personalized list not just to the search term, but also to the history of searches for that individual. Another example? Movie or product recommendations from Internet companies can be surprisingly accurate because they incorporate the viewing data and ratings from large numbers of other “similar” consumers in an effort to more accurately predict consumer preferences.
Once you understand those distinctions, you can begin building your own data product by asking yourself if you can numerically “rank” different things, delighting customers is highly contextual.
Finding analysts and data scientist with the right skills, placed in the right roles, and cultivating the best organizational processes, is incredibly important in a data-driven organization
For example, offering a soda on a hot day to a customer standing in a line outside your business might be effective in a retail setting, but wouldn’t work as a customer experience solution for a manufacturing company. Therefore, it’s important to include contextual information into this numerical ranking. Whether it be movies, products, customers, or employees, ranking them is the first step towards determining the “similarity” between two different things and letting the numbers guide the data product.
Finding the Right People Matters
Finding analysts and data scientist with the right skills, placed in the right roles, and cultivating the best organizational processes, is incredibly important in a data-driven organization. Described by Forbes and Harvard Business Review as the “Sexiest job of the 21st Century,” data scientists are harder to find than unicorns. What do I mean by that? It seems impossible to find individuals with all the requisite skills to make the model work— advanced statistics, machine learning, operations research, mathematics, programming, business and interpersonal communications skills. That’s why, at Alorica, data scientists work in 2-3 person interdisciplinary groups, ensuring the needed skills are covered. An effective data science team is also frequently working across the organization, pulling data from many difference sources and drawing upon the skills from colleagues in business intelligence, data engineering, data architecture and even marketing.
The Sweet Spot on the Bat: Turning Singles into Home Runs
Inside most companies, there are always more ideas about how to conduct business than their resources, so leaders must decide to prioritize. This is where advanced analytics and data science can have the most impact inside a company.
As a first step, every proposed project should make a “forecast” of what success looks like. To be specific, every forecast needs to be a time-series and must be measurable, so within the first days, weeks and months, the company has an initial dashboard against which to measure actual results .
The discipline to quantifiably forecast what success looks like also helps leadership teams prioritize which “home run” projects bring the most bang for the buck. It also helps identify those high effort, low return projects that need to be cancelled. Frequently, companies spend too much time on routine base hits, or low effort, low return projects, often termed identified as “low hanging fruit.” Those efforts that are high reward, but also high effort, could be redesigned to achieve most of the benefits for a fraction of the effort, thereby turning into a home run.
Lather. Rinse. Repeat.
The real learning actually occurs after organizations complete their prioritizations, finalize their project forecasts and measure their actual results. Analysts and data scientists show their real value in trying to explain why actual results are high, or low, versus the forecasts. Understanding why earlier forecasts were wrong is where additional data-driven insights occur and frequently leads to new projects and project forecasts. Then, it’s lather, rinse and repeat, a cycle that leads to continuous improvement over time. As a process, this is the internal engine of a big data-driven organization, using data driven insights to prioritize engineering efforts, with business intelligence measuring actual performance versus forecasts.
Making Lives Better with Big Data
Alorica is betting big on advanced analytics and data science—not only to build proprietary data products, but to become a data-driven organization from top to bottom. With 100,000 customer engagement professionals around the globe, interacting with customers of the biggest brands in the US more than 600 million times a year, we have the opportunity to create insanely great customer experiences that will only improve with more data. As we continue to build an interdisciplinary data science team, and unlocking the potential of this big mountain of data, we’re continuously finding new ways to makes lives better, one interaction at a time—and so can you.