Managing Director & Senior Partner
Related Expertise: デジタルHR, デジタル/テクノロジー/データ, 組織設計
Don’t look now, but shared-service organizations (SSOs) are about to get an upgrade. For the last three decades, SSOs have enabled companies to generate efficiencies of scale, reduce labor and processing costs, and increase the quality and accuracy of services. By combining transaction functions such as HR, IT, finance, procurement, and legal into a consolidated hub for the entire company, SSOs have delivered significant business value. In fact, a new BCG survey of more than 1,100 senior executives from ten industries in 40 countries found that companies with effective SSOs are far more likely to show strong revenue growth and high margins.
But new technologies such as robotic process automation (RPA) and artificial intelligence (AI) have the potential to handle functions at far lower costs (reductions of 20% to 80%) and greater accuracy, without any direct oversight from a person. Will these technologies be the death knell for SSOs? Our experience working with clients across a variety of industries suggests otherwise. Rather than making SSOs obsolete, these technologies will turbocharge them and allow them to generate even greater gains.
To create the next-generation SSO, companies need to focus—urgently—on three priorities. First, they should get the basics right, by creating holistic, end-to-end (E2E) processes that go beyond merely executing transactions and are built with a service approach and a customer-oriented mindset. Second, companies should deploy the new technologies to accelerate the evolution of their SSO. Third, they should develop the required underlying capabilities.
While the benefits of an SSO may seem obvious, many of the companies we work with lack even a basic SSO foundation. In a 2015 BCG survey of senior executives across major industries worldwide, fewer than 50% indicated that they are focused on improving SSO performance through the right behaviors and culture, and only 20% reported that they are consistently driving the SSO vision and mandate. Only 8% have a clearly defined service catalog, and only 10% have clearly aligned the role of service owner for specific processes. The results on actively creating a good customer experience are similarly dismal. (See “Taking Shared-Service Organizations to the Next Level: Five Capabilities to Drive Performance,” BCG article, March 2015.) Today, many companies still have multiple processes for a specific task, or they have partially standardized only a few steps—an ad hoc approach that limits their ability to improve and digitize these processes.
To make these crucial business improvements, companies need to get the basics right, by developing holistic, E2E processes across functions to increase efficiency, effectiveness, and overall business value. These processes need to be measurable and repeatable, and—critically—they need to be designed with a service mindset that puts the customer (either inside or outside the company) at the forefront. When a customer makes a request, the ideal E2E process looks at every step that needs to happen to meet that request and streamlines each step to create the best experience for that customer. This may involve some tweaks to current processes, or it may involve rethinking them from scratch—especially given that many processes are built around the function delivering the service, rather than the customer who requires that service.
In addition to holistic, E2E processes and a service mindset that puts the customer first, several other actions are required to get the basics right:
Together, these actions form a systematic means for SSOs to function more effectively and evolve beyond their traditional role as mere transaction processors to become full-service solution providers that can partner directly with business units. (See “Stages of SSO Maturity.”)
In creating an SSO, companies typically follow a standard path, gradually increasing the scope of processes handled by the SSO and expanding the breadth of the support that the SSO can offer business units. This path spans three main stages of maturity:
In addition, we are now seeing a fourth level of maturity emerging: fully digital shared services. At this level, the SSO uses digital processes for all internal customers, mass customization, and mass application of human-to-machine interfaces and AI.
SSOs have traditionally used several automation technologies to digitize processes, including optical-character recognition, voice recognition, and business process management. The next wave of automation technologies, such as RPA and AI, can help accelerate the evolution of SSOs.
Notably, companies that are just beginning this journey can use these technologies to get the basics right from the start, creating a virtuous loop. For example, fully automated software platforms can handle all aspects of finance functions, such as management reporting. Companies that implement these platforms can quickly establish holistic E2E processes with a global process owner, feedback mechanisms from users, and other critical elements discussed above.
For some processes, companies may choose to use RPA, a software that mimics the activities of an employee on a desktop computer or in the background, accessing company-specific applications. RPA robots (or bots) can navigate among screens, open files, select data fields, copy and paste data, and log in and out of different systems—all without any intervention from a person. RPA works well on rules-based processes that don’t require any judgment calls. And it is scalable: once a bot reaches capacity, the company can easily add another bot to handle the same tasks.
However, companies using only RPA as the primary automation tool run the risk of fragmenting a process by automating only the rules-based parts of that process. Instead, many companies are now exploring a branch of AI known as cognitive computing. Unlike RPA software (which can follow only explicit rules), machine-learning software can—as the name suggests—“learn,” making it highly relevant for SSOs. Cognitive computing uses data mining, pattern recognition, and natural language processing to make decisions. The system, which mimics the way the human brain works, can solve problems without assistance from a person. It can synthesize information from a tremendous variety of inputs—analyst reports, tweets, call center transcripts, emails, texts, newspapers, and blogs—and make complex decisions according to that information. In this way, machine-learning software can generate insights with minimal (or no) intervention from people, making it a very effective tool for automating processes end to end.
Both RPA and cognitive computing have clear advantages, starting with costs. Bots have faster processing speeds than any person, and they do not need lunch breaks—or any other rest. They can operate 24 hours a day, 365 days a year, at much lower costs. A typical onshore employee costs about $80,000, including benefits. An offshore employee or contractor costs about $30,000. Yet bots cost just $15,000 or less. And as the technology evolves, that cost will likely decrease.
The benefits go beyond cost, however. RPA and cognitive computing make the SSO more flexible and responsive to employees (for example, during holidays). They can scale up to meet spikes in demand or to handle predictable peak loads, such as closing the books at the end of a month. Bots eliminate human error, offer detailed reporting and process control, and facilitate the data aggregation that SSOs need to better support business units as a full-service solution provider. (For real-world examples, see “Bots in Business.”) Notably, these new technologies do not require major investment at the front end or changes to a company’s IT suite. In fact, they can build on investments that companies have already made in automation technology.
At companies across all industries, SSOs are using RPA and AI to improve the accuracy of transactions, reduce costs, and accelerate the transformation to a full-service solution provider.
For example, a multinational bank struggled with complex, unwieldy general-ledger accounting processes. IT systems were outdated and incompatible with one another, and the company had eight general ledgers, requiring employees to pull semistructured data from multiple applications and screens. Many tasks were repetitive and had to be handled manually for about 7,000 journal entries each month. By implementing RPA—along with standardized templates and email reporting—the company was able to automate about 85% of journal entry steps. That allowed the company to reduce the number of full-time employees by 40% to 45% while improving accuracy.
In another example, a large online beauty and fashion retailer went even further, applying cognitive computing to improve processes. The company had limited capacity to process invoices, yet sales growth was accelerating. In response, the company used cognitive computing to configure the system to learn the flow of invoices through the natural processing sequence (including the part that employees handled directly). As a result, the system was able to scan both paper and electronic invoices—even those that had non-uniform formats—and match them against purchase orders. It sent rejected invoices back to suppliers and flagged remaining problem invoices to agents for action. This resulted in a significant improvement in workforce productivity and a reduction in processing time.1
1. For more information on this retailer’s use of cognitive computing, see Celaton, Case Study: Streamlining the Accounts Payable Process.
To be clear, companies need to avoid some typical pitfalls in implementing these kinds of tools. SSO technologies are not a silver bullet or a solution for flawed data. Appropriate change management principles and communication are key.
Success in the changing environment will require new capabilities. On the basis of our experience working with clients, we believe there are two types of important underlying capabilities: technological and value-added.
The first set of capabilities are those needed to enable advanced technologies such as RPA and cognitive computing. As these tools become more prevalent, the critical skills needed within the SSO will shift away from underlying process expertise and toward digital capabilities. Many job profiles in today’s SSOs will become obsolete as processes are automated. Yet new jobs will appear, for data scientists, AI integrators, and people who can configure and control robots. For example, companies will no longer need someone who is an expert in closing books but instead will require someone who can convert general accounting principles into code. As a result, companies need an ecosystem to recruit, hire, and retain employees with digital skills.
This transformation is not unusual or unique—manufacturing has undergone a similar transition. Far fewer people work on factory production lines today compared with a decade ago, and many repetitive tasks are now handled by robots, which are faster, less expensive, and more accurate. Yet manufacturing plants require employees with new skills, such as programmers and machinists who work on the robots. Software testing is a similar story. Years ago, companies used people to test and debug new programs, typically in an SSO. Today, software can handle much of that work, but there is a greater demand for people to work on other aspects of the development process, such as designing more intuitive interfaces and a better customer experience.
Moreover, the SSO resources freed up through automation can be invested in building a second set of value-added capabilities, such as customer service management, lean service delivery, data analytics, and decision support. Furthermore, implementing RPA and AI will generate far more data (both structured and unstructured) than companies are currently required to manage, and today, no single entity governs this data in most organizations. Multifunctional SSOs serve as a central hub for multiple sources of data, putting them in a unique position to become the purveyors of this data and enabling a wide variety of analytics across the organization. Ultimately, these capabilities will become table stakes for most organizations, though companies can differentiate themselves by building these capabilities faster and deeper than the competition.
Many companies have taken some steps to create an SSO, but most have further to go. Given the increasing value at stake, the opportunity cost of not making progress on the basics or the deployment of technology is higher than ever. The gap between companies that get it right and those that don’t will continue to widen, in terms of revenue growth and margins.
To gauge their progress in the SSO arena, companies should assess their current level of maturity. This analysis will point to specific areas where companies can start to create value, in terms of functions, processes, and subprocesses. It will also highlight areas where new technology can help companies advance—and potentially even leapfrog some steps in the evolution to a full-service solution provider. For these areas, companies need to create a plan to take action, not over the next two years but in the next quarter.