Dawn Hudson - Fotolia
There are a multitude of reasons why an organization might outsource the analysis of data they have already collected. Companies frequently partner with third-party providers to drive the speed and sophistication of their analytics insights and to connect these insights to action.
Amaresh Tripathy, global business leader of analytics at Genpact, said his company has seen a significant uptick in demand for analytics outsourcing in the wake of COVID-19 challenges.
"Increasingly, we see such relationships become strategic, where partners provide insights and take part in enabling the action as a result of those insights with digital tools and change management activities," Tripathy said.
This often works as a center of expertise model, where the partner brings together a cross-functional team that combines business and technical skills with industry accelerators. Service providers bring digital tools and partnerships with other technology vendors and greater access to third-party data to drive speed and sophistication.
Enterprises are also struggling with the growing volume and variety of data types.
"The outsourcing of analytics not only helps in reducing the cost of analytics but also increases the value and speed of analytics," said Sameer Dixit, general manager of data, analytics and AI/ML at Persistent Systems.
What is analytics outsourcing?
As companies rely more on analytics, they frequently do not have the resources to perform these analyses. Benjamin Taub, CEO of Dataspace, said outsourced analytics typically addresses two main shortcomings: a lack of expertise in the approaches and technologies, and a lack of hands to do the grunt work of bringing data together and analyzing it.
Taub often sees outsourcing teams brought in as experts to consult on and run projects, provide supplemental staffing, or as specialty firms that provide deep knowledge of a specific business or analytics technique.
Historically, analytics outsourcing was driven by the banking, financial services and insurance industry to implement risk and fraud analytics, said Alex Bekker, head of the data analytics department at ScienceSoft. But now he is seeing analytics outsourcing is growing across more industries, especially healthcare and retail, with the biggest areas of interest predictive and prescriptive data analysis.
"By outsourcing these advanced analytics types, companies receive ML- and AI-driven recommendations and forecasts for the next optimal step in their business processes," said Bekker.
Here are 10 reasons enterprises outsource analytics.
Analyzing new data types. There has been significant growth in the use of outsourcing firms with cross-domain knowledge for weaving new data types into analytics workflows, said Rahul Prasad, head of the data and analytics practice at Infostretch. For example, finance is outsourcing the adoption of unstructured data such as news feeds, and market research data in analytics. Insurance companies are outsourcing domain expertise to improve video analysis. Retail is outsourcing analytics to improve hyper personalization and optimization of supply chain and inventory.
Improving forecast accuracy. "The main reasons that enterprises partner with service providers is to drive faster, smarter access to their data and greater business value," said Tripathy. Some enterprises are forming outsourcing relationships that tie compensation to business outcomes like improved accuracy. For example, Genpact has a few strategic partnerships where it has committed to outcomes such as improve forecasting accuracy by x% or reduce inventory levels by y% for a CPG company, which clients use to transfer or share the risk.
Developing algorithms. When Ben Schein, vice president of data curiosity at Domo, a BI platform, was running a data science team at Target, they primarily handed off a lot of the manual or less advanced data preparation tasks to a team in Bangalore. Now he is seeing a shift to outsourcing more advanced skills like algorithm development as these basic tasks are automated.
These kinds of relationships should ensure that the outsourcing partner is not creating a black box. Even if the creation or maintenance is outsourced, it needs to be a living model.
"If I want to create an algorithm for likelihood of repeat purchase, I do not want a static algorithm. I want to be able to configure the algorithm and even own the code where possible," Schein said.
Shoring up lack of internal expertise. Enterprises may also outsource analytics tasks to address a lack of internal expertise and people to do the implementation and testing work. "Just like traditional IT outsourcing, you can get a lot more done in a given period of time with analytics outsourcing," said Taub.
The trick is to really understand what you need. An outsourcing partner can help teams navigate the subtle distinctions between data scientists and Python programmers that happen to know some of the data science libraries.
Accessing domain knowledge. Outsourcing gives enterprises access to data analytics expertise in the moment when it is needed and lets them quickly find specialists with up-to-date skill sets in the specific areas that are required, said Charles Miglietti, CEO and co-founder of Toucan Toco, a BI platform. This expertise can take the form of know-how to evaluate, select and maintain the analytics tech stack. It can also encompass the business aspects. For example, in the pharmaceutical and consumer goods industries, experts can help sort out data collection issues that are laborious and require agreements covering large distribution networks.
"Pick a partner with domain expertise in your space and you can also anticipate getting valuable new insights into industry best practices, and powerful new sector-specific ways to use data and benchmarking metrics to guide your business strategies," Miglietti said.
Adopting breakout technology. Enterprises often turn to analytics outsourcing when implementing a breakout technology that is far outside their existing skill levels. "These types of projects are more difficult to execute, but also offer new sources of revenue and a unique differentiation to customers," said David Tareen, director of AI and analytics at SAS Institute. For example, his team worked with one utility that wanted to use drones and computer vision to monitor underground heat pipes to locate leaks and schedule repairs. This required heat-detecting cameras mounted on drone devices streaming video data to a deep learning model that was trained to identify the location of small leaks.
Improving data storytelling. Analytics outsourcing can improve data storytelling, a narrative approach to data interpretation and analysis that makes business communication persuasive and memorable.
"As a rule, people request analytics outsourcing when they have massive amounts of data, but they can't read it and can't leverage it to the full potential," said Ivan Kot, director of customer acquisition at Itransition Group. "At the same time, they create all kinds of narratives, like brand visions or monthly plans, and in most cases these narratives aren't connected with data."
For instance, operational managers may see their goal in just improving their metrics, like customer acquisition or customer lifetime value. Heads of teams may want to improve operational managers' KPIs. CEOs may want to launch a new branch of business, while the founders want to create hundreds of job opportunities in their native countries.
"All the narratives are related, but almost nobody has a bigger picture and can't understand how their efforts contribute to the common mission," said Kot. A BI consultant can accumulate all existing narratives and connect the dots across different views of the data into a coherent story that guides everyone.
Accelerating time to market. Outsourcing firms can also help companies integrate analytics across new acquisitions, said Mike O'Malley, senior vice president at SenecaGlobal. His company worked with Ability Networks, a healthcare tech company, to rapidly innovate and scale its clinical and administrative cloud workflow products for connectivity and analytics for hospitals. This helped Ability rapidly integrate technology from several key acquisitions into its core platform, which provided valuable time to market advantages.
Curating reporting. Outsourcing firms can also help to prioritize important analytics to improve specific kinds of decision-making processes, Dixit said. His firm has seen a lot of adoption of analytics outsourcing in business functions related to sales, marketing, support and security that help prioritize outliers relevant to specific business outcomes. His team helped one customer analyze its support ticket data to make the self-service process more customer friendly. This has led to a significant reduction in support ticket volume, saving millions of dollars.
Balancing skill sets. Although many enterprises have invested in consolidating their data, the art and science of uncovering meaningful insights requires a balance of new skills including domain understanding, statistics, technology and storytelling. Candidates with these highly specialized and high-demand skills are difficult to hire, provide attractive career path and retain for many organizations.
"Outsourcing analytics passes the burden of bringing these wide variety of intersecting skills to analytics companies which focus on training, grooming and mentoring these high-demand skills," said Sandhya Balakrishnan, region head, data analytics and engineering at Brillio.
While outsourcing is not an ideal situation in all cases, the use of it in data analytics makes it easier to distribute the workload, provide different perspectives on models and reduce bottlenecks while often reducing costs. As such, outsourcing of data analytics functions is becoming more heavily established at all levels.