The pros and cons of outsourcing data analytics

Companies outsource all kinds of IT and business functions to service providers, including some that are quite strategic. Increasingly, that includes data analytics, one of the more competitively advantageous areas of the technology spectrum.

With data analytics outsourcing, organizations hire service providers to perform analytics on the data they provide to the outsourcing company. Industry research shows demand for the service is on the rise.

For example, a 2017 report by market research and consulting firm Hexa Research said the worldwide data analytics outsourcing market is set to expand at a compound annual growth rate (CAGR) exceeding 30 percent between 2016 and 2024, reaching annual revenues of more than $6 billion by the end of the forecast period.

Growing awareness about the advantages of data analytics is a key market driver, according to the Hexa report. Businesses are gradually realizing the importance of analytics in maximizing revenues and in identifying consumer choices, it said, and not every organization is equipped with the required knowledge and resources for effective data analysis.

In addition, the scarcity of professionals for data analytics is hampering the development of competitive data analysis. This has further fueled the demand for analytics services. The Hexa report divides analytics into three main types — predictive, prescriptive, and descriptive — and says descriptive analytics holds a major share in the overall market. The firm forecasts significant growth for prescriptive analytics, because of its widespread adoption among organizations.

Within the category of outsourcing in general, services can be differentiated by offshoring and onshoring.

“Cloud delivery enables easier access to data from regions of the world where labor is less expensive, and this may reduce costs for the ongoing management of algorithms,” says Katy Ring, research director of cloud & IT services at research firm 451 Research. “But actually machine learning technology is more likely to significantly bring costs down in this area. Engineering of the data management systems themselves probably makes more sense to source offshore, rather than the analytics.”

With machines taking the lead on data processing, “the concept of offshore will progressively get diluted over time,” says Beatriz Sanz Saiz, global analytics partner leader of advisory services at consulting firm EY. “It is less about offshoring or [onshoring] and more about how humans and machines work together to get the best outcome.”

Outsourcing data analytics might not make sense for all types of organizations, or for all kinds of analytics. Clearly there are potential benefits, but there are risks as well. Here are some of the pros and cons of outsourcing data analytics, based on insights from industry experts.

Pro: Access to skills in short supply

It’s well known that people with certain IT skills are in short supply, and this includes professionals who specialize in areas such as cloud computing, advanced analytics, big data, data lakes and data science. Outsourcing companies can help close the gap by providing this kind of expertise.

“As data volumes are expanding, trying to keep pace within the traditional data center is proving impossible,” Ring says. “This is driving the requirement to manage the data estate in the cloud by providing access to big data on Amazon Web Services, Microsoft Azure and Google Cloud Platform.”

For this, organizations need cloud management platforms so that they can provision big data lakes and manage data loads and transfers from individual consoles, Ring says. “However, it is challenging to operationalize this type of approach with an IT team that does not have the [appropriate] skills in-house. Outsourcing can provide organizations with access to such skills.”

Con: The risk of choosing the wrong provider

Deciding on which service provider to partner with can be a challenge with any type of outsourcing engagement, and data analytics is no exception.

“Vendor selection could be daunting with many touted ‘best in breed’ technologies,” says Alison Close, research manager of finance and accounting, BPaaS, and analytics services at International Data Corp. “While cost is obviously a major factor in the vendor selection process, cultural fit and alignment of teams plays an equally important role.”

Companies expect much more strategic, high-touch partnerships today, where resources are ingrained in everyday operations, communication channels are effective, and delivering on business outcomes is vital, Close says. “These are outcomes that extend beyond just cost reductions,” she says.

Pro: Industry expertise

While some data analytics functions are universal, others can be specific to certain sectors such as healthcare and financial services. Finding an outsourcing partner that has deep industry expertise can be a huge competitive plus.

“Providers who may hone domain expertise in retail, for example, will have specific analytics services offerings like customer lifetime value analysis, store sales analysis, profitability analysis or market basket analysis — all techniques very specific to this industry,” Close says.

“They can also provide benchmarking data/metrics to show a point of comparison to industry standards or other players in the industry,” Close says. “This could be used as competitive advantage.”

Con: Cost vs. value trade-offs

Once a predictive model is created and converted into a product by an external service provider, it needs to be operationalized for as long as required, Ring says. That means tweaking and redeploying rules in the algorithm so that the insights it delivers remain meaningful.

“Data changes constantly, so the model must not be left to degrade,” Ring says. “These constant updates come at a price, however, and that price will be more than business lines are used to paying internal IT services for BI [business intelligence] reports.”

In fact, the whole issue of the cost of outsourcing services could be a challenge, especially for larger organizations with more complex operational models.

Getting executive buy-in across the enterprise and funding “could be a challenge, especially if you’re trying to centralize data sources that exist in silos and different lines of business are involved in that decision making or funding process,” Close says.

Pro: Easy scalability and a quick path to analytics maturity

Beyond the ability to acquire data analytics skills, outsourcing services can help organizations quickly build up an analytics infrastructure that might not be easy or even possible to do in-house.

Data analytics “has become a natural part of doing business, and is nowadays more than ‘just’ data warehousing and business intelligence,” says Jorgen Heizenberg, research director of data and analytics at Gartner. “This requires a level of scalability and complexity that is not always found in-house. One of the most common reasons organizations look for external [analytics] support is because they lack the [internal resources] to meet this growing demand.”

Another consideration is the potential to keep costs down while acquiring these analytics capabilities, Heizenberg says, because these providers often leverage analytical assets such as frameworks and accelerators.

Taking advantage of the technology expertise of the provider, whether it’s implementing a data warehouse or bringing in robotic process automation or cloud-based tools to improve operational efficiencies can be a big plus, Close says. “Outsourcing data analytics to a third-party provider could also help introduce more innovative solutions that an enterprise may not currently be considering,” she says.

Con: Losing control of data storage and of analytic models

Any outsourcing arrangement generally means giving something up, such as control and in some cases even employees. With analytics outsourcing, one of the biggest resources to be sacrificed is analytics models.

“For insight as a service, the customer has typically given the data to the service provider in order for the service provider to give the customer back the answer,” Ring says. “In this model, the customer never owns the logic or the algorithm. Consequently, when the customer exits, all that is owned is the data and the recommendations, not the models, approaches, framework or configuration.”

Companies that are outsourcing data analytics might also encounter concerns about where their data is actually being stored and whether the storage location is the best option for them, Close says. “Is it stored internally onsite, in a dedicated environment just for your organization at the provider’s data center?” she says. “Or in a hosted ‘shared’ public cloud environment at the provider’s data center?”

Pro: Ensuring ongoing data protection compliance

As data volumes grow, the process of managing and analyzing the data can put organizations at a greater risk of non-compliance with a host of regulations.

Differences between governance and security policies across data source systems create challenges for companies when it comes to auditing data in data lakes, Ring says.

“Particularly as requirements around personally identifiable information and the General Data Protection Regulation (GDPR) go into effect, the need for more easily audited data can be a catalyst for seeking an external outsourcing partner,” Ring says.

Con: The need to own the data management strategy

To support the democratization of data within organizations, a chief data officer (CDO) is needed to champion a company-wide strategy for the capture, management and sharing of data, Ring says.

The self-service analytics and governance layers need to be architected the right way to enable a range of use cases over time, Ring says, and this is why a CDO role is so important.

“The CDO is ultimately responsible for business and IT alignment around data management,” Ring says. “If the [organization that’s outsourcing] does not have this internal role in place, there will be limited success for an outsourced approach.”

Pro: Greater potential to leverage the value of data

It’s been said that data is the new currency for businesses, and there’s certainly a lot of potential to leverage analytics for business gain.

“With machines taking the lead on data processing, the value chain of data and analytics will fundamentally change,” Saiz says. “As in every digital business, disintermediation will happen so value will stay at the very two ends of the business, either at the data end or at the decision support/business insights end.”

Within this context, a possible benefit of outsourcing is the opportunity to leverage a data marketplace and build alternative business models, “with an independent third party running the custody of data from multiple organizations and promoting the concept of [an] anonimized and secure data exchange platform,” Saiz says.

Con: Potential for conflicts

Everyone goes into outsourcing agreements hoping the relationship will go smoothly for all parties. But issues can crop up that jeopardize the harmony of the arrangement. That’s especially true if companies have not been thorough in creating contracts.

When outsourcing data analytics “organizations often forget to include key contracting terms such as termination, data governance, IP [intellectual property] ownership, liability, metrics and SLAs, pricing model, and additional capacity and renewal costs,” Heizenberg says. That can lead to potential conflicts during or at the end of the engagement.