The building blocks, or ideals as we prefer to call them, not only include technology driven solution ideals but, in equ
Driving a Holistic Approach to Intelligent Automation | 1
REDWOOD POSITIONS ITSELF AT THE INTERSECTION OF ERP, COGNITIVE COMPUTING AND RPA
Author:
Tom Reuner, Managing Director, HfS Research May 2016
Executive Summary Business operations leaders and global process owners continue to struggle to find answers to the dual mandate of driving down costs while simultaneously increasing the quality of service delivery. As organizations move toward As-a-Service, “time to value” is By guaranteeing outcomes and by becoming another fundamental challenge, thus putting innovative taking a business function point of approaches to automation center stage. In the past labor arbitrage was leveraged to address process inefficiencies and to find cost saving view, Redwood is adding a opportunities. The emerging notion of Intelligent Automation is providing distinctive voice to the discussion on an opportunity to overcome this dichotomy by de-coupling routine service Intelligent Automation delivery from labor arbitrage. Redwood’s approach to Enterprise Process Robotics is adding a distinctive voice in those discussions by addressing automation from a business function point of view rather than from a task or sub-process level. Instead, Redwood is focused on back office/administrative processes for finance (RoboFinance™), supply chain (RoboSupplyChain™) and human resources (RoboHumanCapital™) that are largely standardized but with some variation in each organization. Creating deep domain expertise in those processes allows Redwood not only to robotize those variations and gaps, but Redwood is offering robotized also to guarantee outcomes to its clients.
process IP, not task or swivel
For example, Redwood has a set of marquee clients for whom it has deployed RoboFinance successfully. However, in the views of its clients Redwood’s differentiation comes from the fact that they are being perceived as understanding how processes work and how business rules have to be applied rather than engaging through technology. This understanding is borne out of a deep domain knowledge of business functions and in particular of the leading ERP systems. In the words of one of Redwood’s clients: “We were talking to financial people. We did not talk to software people.”
chair automation
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 2
Supporting Clients on their Journey into the As-a-Service Economy Almost every discussion we have in the industry these days comes around to “digital transformation” at some point and often the dialogue includes reference to the customer experience of digital disruptors such as Uber and Airbnb. For most enterprises that were not “born in the cloud,” the journey toward “digital” means transforming a traditional delivery capability to support this digital transformation. Crucially, those enterprises have to find ways to deal with as well as write off legacy applications. This is particularly prevalent is sectors such as Financial Services, but also more horizontally around ERP applications. Unlike Uber, which could set up its IT architecture as well as service delivery in a Greenfield approach, most organizations continue to grapple how to overcome legacy environments? HfS describes the journey for organizations into the As-a-Service Economy against this backdrop. The building blocks, or ideals as we prefer to call them, not only include technology driven solution ideals but, in equal measure, change management ideals. Without a change in mindset, technological innovation is unlikely to fulfill its potential. Exhibit 1 outlines the eight ideals of the As-a-Service Economy:
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 3
Exhibit 1: 8 Ideals to Achieving Intelligent Operations in the As-a-Service Economy
Source: HfS Research, 2016 The disruptive shift on this journey that is facing most organizations is the time to value for new products and services but equally the expectancy of a real time delivery of services. Thus, the notion of Intelligent Automation as placeholder for a set of disparate yet overlapping and interdependent process automation approaches is increasingly coming to the fore. As organizations have to address both the quality of their processes as well as the time it takes to deliver services, throwing labor in low-cost locations alone at problems is not a feasible option anymore.
Intelligent Automation Needs to be Assessed as a Continuum By decoupling routine service delivery from labor, we believe Intelligent Automation is one of the most disruptive forces IT and business process operations have ever seen. Intelligent Automation will affect every organization, whether on the supply or buy side. Historically, it was cheaper and more efficient to throw labor at issues that arose from badly structured business processes and applications. By looking at the sheer magnitude of people working in delivery centers in low cost locations, the scale of this disruption becomes tangible. However, we must
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 4
be careful not over-simplify the arguments. The discussion should not just be about cost and taking out FTEs as we have seen the early phase of the market development. Rather, the focal point should be how humans can be augmented by these innovations. Put in more simple terms, the advances in Intelligent Automation have to be part of the journey toward the As-a-Service Economy. Thus, the context for discussions on Intelligent Automation should be service delivery and the starting point should be use cases, but not technologies or even tools. Consequently, we need to approach Intelligent Automation from a holistic and more transformational perspective, as we highlight in Exhibit 2. Transformation means integrating innovations such as RPA or Autonomics into the broader service delivery cababilities. Therefore, to achieve the desired process efficiencies, it is crucial to call out the interdependency with other, often more low-level automation concepts including runbook, scripting and scheduling. As Exhibit 2 suggests, different automation concepts are both overlapping and interdependent. Consequently, broad service orchestration capabilities as well as automation frameworks have to be developed. That is the thought process for why we have termed the model in Exhibit 2 the HfS Continuum of Intelligent Automation, providers are deploying the various depicted automation options in that model depending on use cases and process maturity. Hence, in business operations we have not yet reached an industrial scale of deployments, but are advancing from a sub-process level toward end-to-end processes. Further critical reference points depicted on the model are the characteristics of the process as well as of data/information. For instance, RPA works well for rules based processes with standardized language while Autonomics is progressing toward unstructured information and dynamics language. In broad terms the evolution of Intelligent Automation will progress towards unstructured data as well as toward less well defined processes. Painting with a broad brush, business processes lack the standardization of IT processes that have been standardized through methodologies such as ITSM and ITIL. Fundamentally, organizations should assess these options within the Continuum or see them as part of a sourcing toolbox. In many cases, the journey toward fully automated processes will be undertaken by bundling or integrating different sets of tools and technologies. HfS believes a strong and plausible scenario for an end-state of process optimization is the notion of highly verticalized insights. Industrialization, standardization and automation will support the provision of vertically relevant business insights at an industrial scale. However, there are no short cuts to get to this endstate, but the tools of Intelligent Automation have to be integrated with analytics capabilities as well as cognitive engines.
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 5
Exhibit 2: The HfS Continuum of Intelligent Automation
Source: HfS Research, 2016 Zooming in on RPA, the core value proposition is twofold: First, it is about recording, scheduling and the playing back of process steps that could not be automated with other technologies or where the cost would not support the business cases. Second, the extraction of data across heterogeneous systems at scale. In the past it was deemed easier and cheaper to apply near endless labor in low cost locations to those problems. Within the notion of RPA, algorithms or bots perform tasks just like a business analyst or transactional agents. However, none of these tools are turn-key solutions or even silver bullets. Therefore, the discussion with clients should start with understanding and mapping the end state or put in more simple words articulating what is meant to be achieved. Based on many discussions with industry stakeholders, HfS has enhanced its Continuum of Intelligent Automation. First, we have introduced a layer of service orchestration that cuts across and underpins the broad notion of Intelligent Automation. This reflects both the need for orchestrating the diverse set of automation approaches as well as platforms that offer broad integration capabilities. Second, based in particular on discussions with Redwood, we have introduced ERP on the same level as traditional BPM and Workflow rather than using the more generic terms “business process automation” that we used for historical reasons. The thought process here is the
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 6
need for integration with these approaches in predominantly business process centric scenarios. At the same time deep domain knowledge around either of these categories but in particular, knowledge of ERP offers a way of differentiating the value proposition for tool providers. As we will detail in this report, Redwood plays in different segments of the Continuum in Exhibit 2: it is positioned at the intersection of RPA, Cognitive and ERP while offering service orchestration including cognitive capabilities at the same time.
Redwood Offers Robotized Process IP, Not Just Technology Redwood’s broad automation capabilities are a compelling example of how a solution can bring together overlapping and interdependent approaches within the Continuum of Intelligent Automation. The company aims to be part of the strategic process stakeholder discussions created by the ascent of RPA while offering a markedly different approach than the leading RPA vendors. Redwood was founded in 1993 with a single-minded mission—to help organizations eliminate the costs, risks and wasted time associated with manual tasks. Over time Redwood has developed a broad set of automation capabilities, but its sweet spot and strongest differentiation lies in its deep domain knowledge of ERP based processes and in particular around SAP® and Oracle. Consequently, the emphasis is on handling back-office processes end-to-end and not looking for low hanging fruit to optimize. Hence, the key differentiation focuses mainly on three aspects:
» Instead of taking a user centric view and mimicking what users do, Redwood looks at processes in their entirety. Consequently, human activities get robotized as well as application and system tasks, which are connected directly with the ERP or other business systems in order to maintain process integrity.
» In order to automate entire processes such as Order to Cash or Procure to Pay, traditional RPA vendors have to start from scratch and teach their robots how to handle every step. In contrast Redwood has robots that have the complete knowledge of finance and business processes and how to efficiently execute them. [how this is done technically, library of artefacts and process maps etc.].
» Scalability based on a 3 tier architecture that can either sit on premise, public or private cloud and which interacts with all necessary applications via a single instance. Crucially, this solution does not grow exponentially because Redwood orchestrates the processing directly within the underlying ERP system. Redwood’s vision is to provide robotized end-to-end processes as a self-service to the customer, which can be either internal (back office) processes or external (front office) to customers. Thus, service orientation is addressed by having the following characteristics at the same time:
» » » »
Fast Efficient Fully transparent and governed Agile to changes
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 7
The objective is to eliminate as many manual steps within the end-to-end process to improve the service quality that is delivered to the end customer (internal or external). Looking at Redwood’s approach around its process robotics offering in more detail, it is significantly different yet complementary to the leading RPA vendors. It is different in that Redwood doesn’t approach customer situations from a task or sub-process level and then work its way up in the organization or going after ‘swivel chair’ front office processes like claims management or mortgage processing. Rather, Redwood is focused on back office/administrative processes such as finance (RoboFinance™), supply chain (RoboSupplyChain™) and human resources (RoboHumanCapital™) that are largely standardized but with some variation in each organization. The starting point in those discussions is the discovery of any required parameters so that process owners can configure the automation steps easily and consequently the time to value is significantly reduced. Examples for this are pre-defined rules, for example if an order is less than 100% or the asset value is less than 10% of the net book value then intervention is required. These adjustments can be maintained by the process owners themselves and the action of robots will be based on any new instructions. In a nutshell, it is a top down approach starting with a holistic view of an end-to-end process rather than bottom up as in the case of RPA vendors that start with tasks and sub-processes and work their way up through processes. At the same time Redwood offers a broad set of tools that are cognitively intelligent and aware of:
» » » »
The process steps which have to happen Both the explicit and the implicit rules of the process The meaning of the information which is created and processed Their context within other processes and of robots with which they are interconnected and work collaboratively with in a coherent and orchestrated manner.
An example of this is the robotization of SAP’s financial revenue recognition process where robotization is performed end-to-end and where the robot remembers and recalls activities it has done before. Specifically, it has learning, memory and experience beyond just a single execution cycle. Details of the use cases for Redwood’s RoboFinance offering can be found in Exhibit 3. Exhibit 3: Redwood’s RoboFinance Offering
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 8
Source: Redwood Beyond the specific use cases in Finance & Accounting, the exhibit calls out the Redwood’s Managed Control and Compliance Environment. This has evolved out of the broad Enterprise Process Automation capabilities for SAP® and Oracle. Thus, Redwood can support clients in managing the heterogeneity in Shared Service Centers (SSC) by outlining the dependencies that need to be managed across different parts of a process. At the same time, it is a conduit for clients that work in outsourced environments, to orchestrate the process flows between Global Process Owners (GPO) and BPO providers. All too often, a stakeholder on their own does some automation but what is missing is the orchestration across the two. This might lead to highly optimized results on one side but these efficiencies might get lost in the nebulous world of reverse SLAs and interdependencies. The net result is that there is often not that much improvement in the end-to-end processes. Here is where the Control Environment cuts in as there are processes where the hand off happens from one department to another. Therefore, it is important for those events to be controlled, even though the functions are spread globally, be it the retrained organization, front office, back office, separate SSCs or other functions. A financial, supply chain or human capital management process is linked end-to -end – whether actually operationally integrated or not.
Lessons from Redwood’s Automation Clients Reflecting the nascent state of the market around Intelligent Automation, there is a scarcity of relevant client references. In order for the market to evolve, insights into the lessons learned in the early deployments are critical. Insights from some of Redwood’s marquee clients provide valuable reference points and learnings around
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 9
automation. A point in case is Royal DSM, a Dutch Life Sciences and Materials Sciences company active in health, nutrition and materials.
Royal DSM In 2012, DSM implemented a Financial Shared Services Center to support the company’s growth strategy. DSM opted for a captive shared services center, set up in India, as they felt that having the employees connected to the company made more strategic sense than outsourcing. Part of this process was a standardization exercise as DSM had grown significantly through M&A resulting in an enterprise architecture with 11 different SAP systems. To cope with this heterogeneity DSM put a SSC framework on top of the processes to move toward managing the transactions seamlessly. Executives described the mindset of the migration as ”migrate first, do the optimization later.” The initial time frame for the project was estimated to be 4 to 5 years. Half way through the project, DSM’s management asked whether the project could be accelerated, which led to the inception of a global improvement program. The goal was to move up to the top 25% of the benchmark. However, at that point DSM was ~40% off the benchmark. The fundamental question was; how could you lower the cost while raising the quality of the processes at the same time? Global sourcing is a lever for cost, but quality is not guaranteed. This prompted DSM to think differently and explore the option of driving automation through its processes. As Redwood had already engaged with DSM around the Financial Closing Cockpit in their SSC Framework, it had already demonstrated their deep process knowledge. Hence, DSM decided to expand with Redwood rather than going with other automation tools providers. Inside DSM opinions were divided. Some executives believed robotics could be transformative, while other executives doubted that robotics would work in an SAP environment. Thus, change management was key and a proof of concept was meant to be the conduit. A proof of concept not only suggested that cost went down, but crucially there was an improvement in overall quality as well. However, the CFO raised the valid concern: “Don’t come back to me in two months’ time, saying that you can’t answer any questions anymore as the robots are doing all the work.” Therefore, control points had to be built in. The eventual rollout was done in a phased approach beginning with an initial global business unit. First, DSM simulated with robots, a month end close that had already happened in order to ascertain they had taken the right steps. The result was exactly the same figures but a much faster delivery. The main challenge was to set up the business rules and decide what they were as these steer the whole process. Furthermore, user acceptance testing yielded very positive results and suggested that the change management process was starting to manifest itself. Based on these results, three further business units followed after which a re-evaluation was scheduled. After a continued positive business case, the rollout was expanded across the whole organization which took another three months without any major incidents. Looking at the results the best case expectation was that across the board up to 80% reduction of manual activities could be achieved.
» In a subsequent review the manual activities reduced were recalculated and found to be higher than expected at 89%.
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 10
» The input used to be based on spreadsheets, mails and other sources. Now the robot picks up that information and incorporates it into that journal, sends it back, and assures compliance without anybody touching it.
» The quality went up significantly, as everything, which now was in the journal, was actually booked. Based on the project the work of the people has changed. It moved from largely transactional work to higher value tasks, including re-assigning work and moving people into new positions. In the words of DSM executives, the differentiation of Redwood was “...that we were talking to financial people. We did not talk to software people. We really talked to financial people who knew how the processes work with in-depth knowledge in for instance general entries and reconciliation work. We talked about the business rules […] It is not about the technology. It is about the results, it is about the financials and it is about how you do the processes.” In contrast, many competitors were perceived as talking about software. Differentiation also came through because Redwood met DSM’s key prerequisite: The full integration of automation into SAP. As DSM suggested, in SAP you can see what is automated and what is manual, which they felt was a huge advantage to drive change management as people could see the level of manual intervention. At the same time people didn’t have to learn a new system. The positive results culminated in DSM being able to apply robotics to one legal entity company code per week, which in their words was “unheard of for an SAP implementation.” Reflecting on the project, executives at DSM suggested that automation “will change the way we think about SSC. “Going forward it needs to be decided whether labor arbitrage is still a sensible strategic approach. “Because if you take away labor, that is the best labor arbitrage you can have.” Thus, you need a vision how you transform knowledge work.
SAP® The project at SAP adds more color on the RoboFinance offering. Over the years, SAP has built out a complex yet mature SSC organization to centralize transactions around P2P, O2I and R2R. The objective of which was to standardize the processes, leaving one person responsible for a specific end-to-end process. In 2008, SAP launched a “Finance Better Together” initiative to deliver a single virtual team. In 2013 SAP embarked on a R2R transformation project with a view to centralize, standardize and automate what was a very manual process. By its own admission, SAP has a very complex revenue recognition cycle, where deferral of revenues had to wait until invoices were issued (usually at month-end or quarter-end) and deferrals took up to two days. The Concur acquisition in December 2014 not only added to this complexity but also caused disruption in the transformation. SAP executives commented: “When completing the spreading reporting we had to do Sarbanes Oxley (SOX) checks for errors. However, there were too many transactions and too many steps to cover.” Before transformation, SAP took four weeks to publish and seven working days to close their entities.
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 11
Primary drivers for the transformation project were to gain efficiencies and standardize processes through a centralized approach. Compliance around SOX, International Financial Reporting Standards (IFRS) and tax regulations were also drivers for change. Furthermore, SAP had data quality targets to meet in order to gain compliance. Having re-evaluated their entire close process, SAP decided to implement Redwood’s RoboFinance solution specifically geared to the R2R function, RoboClose™. Based on handbooks that described standard operating procedures in offshore delivery centers, robots took over the majority of manual steps. The main achievements included:
» A reduction in the time it takes to publish and to close entities from four weeks and 7 days respectively to three weeks and four days respectively.
» Redwood’s robots now perform 100% of the revenue recognition tasks, eliminating the 197 man days required per close period.
» Globally standardized processes. » Guarantee of 100% SOX compliance. » Strong positive feedback from employee surveys. SAP executives summarized the achievements as followed: “With RoboClose the spreading report is completely automated. Robots now check invoices and deferred revenue, and remove it from the deferred report automatically. It originally took 90 minutes for each spread report to run, but now it takes just a few minutes for each. SAP can now re-order processes and save time. For example, we allow certain entities to start their spreading reports if they're finished with invoicing, so they no longer have to wait a whole day. We can even run all entities in parallel and track each one.” Similarly, the strategic impact of RoboClose especially on employee morale was described as: “RoboClose allows us to give more control back to the regions so people can finish earlier. It's a good Shared Services Center model. We want people to be able to get home on time and have already squeezed a lot of hours out of the close to keep it more manageable.”
The Future of Intelligent Automation As the reference cases have shown, the deployment of Intelligent Automation should not be confined to notions of removing FTEs and just cost, but should be part of an organization’s journey into the digital age. Or, as we at HfS call it, the As-a-Service Economy. As part of this transformation, the emphasis should be on how experienced staff can be augmented by automation and other innovative approaches. Human augmentation is starting to emerge as the mainstream narrative around Intelligent Automation. However, this augmentation will be part of a fundamental transformation of knowledge work. While many generic back-office jobs as data entry and compliance will be phased out, new higher value positions will be created. Organizations but also education institutions have to prepare for and adapt to this disruptive shift that is starting to happen.
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 12
Cognitive engines and artificial intelligence will play an increasing role in the automation tool box. Similarly, we will see emerging technologies such as Natural Language Processing, Deep Learning and Context Aware to mature and necessitate the need for broad service orchestration capabilities. Artificial Intelligence in its various guises will be a critical conduit of expanding and integrating automation strategies into Big Data scenarios. Thus, the end-game is likely to be around domain-specific insights The more we standardize, industrialize and automate, the more the focus will be on highly vertically infused insights. It is here where value will be created and differentiation will be achieved. All this necessitates broad service orchestration capabilities. For instance, we are starting to see ServiceNow platforms being expanded to and integrated with Intelligent Automation capabilities. Thus overcoming legacy ways of managing IT environments with process innovation. Redwood has demonstrated how its approach of looking at process automation from a business function point of view with end-to-end responsibility can contribute to orchestration within selected domains. As an industry we have to widen these discussions to governance, organizational models and testing in equal measure as what is today a nascent mature continues to grow and mature.
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 13
About the Author Tom Reuner Tom Reuner is Managing Director for IT Outsourcing Research at HfS. Tom is responsible for driving the HfS research agenda for the “As-a-Service Economy" across SaaS applications, cloud eco-systems and IT. Together with his HfS colleagues Tom continues to develop ground breaking research around process automation and cognitive computing in both IT and business processes. A central theme for all of his research is the increasing linkages between technological evolution and evolution in the delivery of business processes. Tom’s deep understanding of the dynamics of this market comes from having held senior positions with Gartner, Ovum and KPMG Consulting in the UK and with IDC in Germany where his responsibilities ranged from research and consulting to business development. He has always been involved in advising clients on the formulation of strategies, guiding them through methodologies and analytical data and working with clients to develop impactful and actionable insights. Tom is frequently quoted in the leading business and national press, appeared on TV and is a regular presenter at conferences. Tom has a PhD in History from the University of Göttingen in Germany. He lives in London with his wife and in his spare time he is trying to improve his culinary skills in order to distract him from the straining experience of being a Spurs supporter. Tom can be reached at
[email protected] and followed on Twitter @tom_reuner.
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com
Driving a Holistic Approach to Intelligent Automation | 14
About HfS Research HfS Research is The Services Research Company™—the leading analyst authority and global community for business operations and IT services. The firm helps enterprises validate their global operating models with world-class research and peer networking. HfS Research coined the term The As-a-Service Economy to illustrate the challenges and opportunities facing enterprises needing to re-architect their operations to thrive in an age of digital disruption, while grappling with an increasingly complex global business environment. HfS created the Eight Ideals of Being As-a-Service as a guiding framework to help service buyers and providers address these challenges and seize the initiative. With specific focus on the digitization of business processes, intelligent automation and outsourcing, HfS has deep industry expertise in healthcare, life sciences, retail, manufacturing, energy, utilities, telecommunications and financial services. HfS uses its groundbreaking Blueprint Methodology™ to evaluate the ability of service and technology providers to innovate and execute the Eight Ideals. HfS facilitates a thriving and dynamic global community of more than 100,000 active subscribers, which adds richness to its research. In addition, HfS holds several Service Leaders Summits every year, bringing together senior service buyers, providers and technology suppliers in an intimate forum to develop collective recommendations for the industry and add depth to the firm’s research publications and analyst offerings. Now in its tenth year of publication, HfS Research’s acclaimed blog Horses for Sources is the most widely read and trusted destination for unfettered collective insight, research and open debate about sourcing industry issues and developments. Horses for Sources and the HfS network of sites receive more than a million web visits a year. HfS was named Analyst Firm of the Year for 2016, alongside Gartner and Forrester, by leading analyst observer InfluencerRelations.
HfS Research | The Services Research Company™ © 2016, HfS Research, Ltd | www.hfsresearch.com | www.horsesforsources.com