Leslie Willcocks, London School of Economics and Political Science.
and
John Hindle Knowledge Capital Partners

“Unlike many organisations that have come to RPA organically (i.e. a little at a time), Mars’ vision for automation was one of being scaled from the start.”
Mars executive

At Mars we talked to John Cottongim and Mario Maffie. At the time, before the Coronovirus crisis, John was heading the Enterprise Automation Hub at Mars Incorporated, delivering Robotic Process Automation and Cognitive Automation solutions across Mars’ geographies and product segments. Mario was the Chief Information Officer of Mars Global Services. Mario led all technology needs globally for Mars Incorporated across the HR, Finance, Procurement and Corporate Functions. The case gives insights into how automation was adopted, the business case, challenges, success factors and the business value delivered. We get a glimpse of the future direction with automation, and the executives and researchers discuss the lessons and insights derived from the case.

Background and Evolution

Mars Incorporated is an American global manufacturer of confectionery, pet food, and other food products and a provider of animal care services, with US$35 billion in annual sales in 2018. Headquartered in McLean, Virginia, United States, Mars operates in five business segments around the world: Mars Wrigley Confectionery, Petcare Food, and Symbioscience, the company's life sciences division. During 2019 Mars had 115,000 employees.

During the first half of 2017, Mars identified a series of core capabilities that would enable the company to best undertake an Enterprise Digital Transformation. Those core capabilities included User Centricity/Design Thinking, Analytics & Machine Learning, and Automation. The introduction of these capabilities into the Mars ecosystem was done in a way that has enabled each capability to leverage and enhance each other, enabling positive synergies and creating what Mars has termed the Digital Engine. Along with these focused capabilities came complementary shifts in both process and mindset regarding how various Mars teams identify, manage, and run their digital initiatives.

One arena where the Digital Engine most clearly stands out is in the identification and design of Mars’ digital solutions—be those solutions in Mars’ business Segments (Mars Wrigley Confectionery, Mars Petcare, Mars Food, and Mars Edge) or in traditional business services (Finance, Commercial, People and Organisation, Digital Technologies). The enhanced focus on User Centricity, combined with agile delivery, enabled the shift away from a traditional waterfall-oriented culture. This combination moved the project teams towards a ‘test-and-learn’ mentality, building iterative solutions. Further, these new toolsets enabled the transformation teams to be ‘sero-distance’ from our customers, delivering better solutions that meet specific business needs. The rapid prototyping and pace of projects also enabled improved freedom of thought via quick decisioning and iterative feedback cycles, all while amplifying our associate’s voices throughout the process.

Particular to Mars’ Automation journey, this portion of the Digital Engine began in earnest during Q3 of 2017 with the formal creation of the Enterprise Automation Hub. The Enterprise Automation Hub at the time was a centralised team designed to engage, design, build and support automation solutions across Mars’ business teams in a consistent way. This consistency enabled various business teams to better internalise the hub’s capabilities and therefore best leverage RPA as part of their ongoing transformational activities. Operationally, the Enterprise Automation Hub was designed to cross-organisational boundaries and operate in a manner that effectively blended Digital Technologies, Operations and Transformation functions.

Unlike many organisations that have come to RPA organically (i.e. a little at a time), Mars’ vision for automation was one of being scaled from the start, as was the case for the other Digital Engine capabilities. The intention of scale therefore drove a variety of design considerations in terms of team structure and partner engagement models. This strategic intent, to deliver an automation capability with near-immediate scale, proved extremely successful in achieving rapid buy-in from various business stakeholders as well as providing rapid tangible benefits to partner business teams.

Although the Enterprise Automation Hub quickly came to an agreement on which set of technologies to leverage, Mars also later engaged with multiple delivery and support vendors in order to deliver high quality automations with minimal backlog or organisational friction. A well-defined technology platform, along with strong technical competency of the automation team, allowed Mars’ centralised hub approach to scale from POC-phase (January - April 2018), to hiring core automation team members (ending June 2018), to having 40 in-production automations at the end of 2018, and with over 50 automations in place as of May 2019. Those 50+ automations delivered more than 150,000 hours of productive time back to associate teams, while enabling improved customer experience, higher quality outputs and greatly reduced overall cycle times of effected processes. The diversity of benefits, beyond merely offsetting hourly effort, reflected the team’s partnership, and utilisation of the User Centricity toolkit. With the help of design thinking sessions, the processes that teams were automating could be viewed with a fresh perspective, informing how automations could play a part in delivering efficiency and experiential benefits to all participants.

During 2019, the team continued to move at pace, projecting more than 100 automations in production across an ever-increasing swathe of the global Mars organisation by the end of the year. Given the scale of the Mars organisation, with over 125,000 associates and many contractor partners, the first 100 automations were seen as just the start of a much broader adoption of automation across the Mars enterprise.

Initial Business Case/Challenges

Though the team envisioned automation as a piece of a larger Digital Transformation, there was a formal business case and a set of objectives particular to the Enterprise Automation Hub during its first year of operation. This formal business case was presented with the intention of delivering significant hours of efficiencies, improved quality and reduced cycle times, with the benefits targeted at both internal and external stakeholders. It was assumed that successful execution of these early opportunities would lead to a substantial demand across the enterprise. As such, the hub ran a series of POCs in a variety of business areas. Those POCs were successfully completed and all entered production shortly thereafter. During 2019, the initial business case continued to be executed upon, with favourable results being achieved over the first full year of delivery. The demand within all business areas grew as expected, with the Enterprise Automation Hub being particularly engaged as thought-partners embedded in a variety of transformation efforts taking place across Mars.

Proof of Concepts

An original set of eight proof-of-concepts were conducted simultaneously across various organisations. This method of multiple POCs was enabled by the premeditated scale of the program. This methodology allowed the hub to showcase the capability of the technology in each of several core business areas while greatly reducing the time-to-scale once the technology was sufficiently proven. The POCs took place over the first half of 2018 and were successful in demonstrating the technology’s applicability in a wide range of business-rule-driven tasks.

In parallel to the POCs, the formal staffing of the Enterprise Automation Hub was being completed with a mixture of external talent with substantial automation experience, as well as experienced Mars associates that would enable swift navigation across the Mars global organisation while keeping the team cantered on Mars’ unique culture and history.

Challenges

Challenges faced—many common across technology adoption—included the following:

• o Incorporating a new ‘agile-oriented’ technology within a legacy operating culture. Automation is a technology that promises speed to execution while it tends to be most highly beneficial within the context of a legacy-operating environment. This leads to an Enterprise Automation Hub that was built around agile development cycles to negotiate the waterfall nature of application access, security assessments and Dev/Test/Prod environment differences.
• o Encouraging the teams to use User Centricity as a tool to reimagine their processes, so that the teams did not simply automate a sub-optimal process, but redesign processes where beneficial, in order to best improve customer experience along with quality and efficiency.
• o Given the newness of automation, most business areas requested that a POC be completed within their systems and business processes. Although this is customary, as a program gets under way, it is necessary for a team to push back against POCs that do not have a clear runway to both production and sustained benefits. The demand for POCs, particularly after the technology has been put into production across various business units, is one that the team must carefully navigate to avoid excessive work without production benefits.
• o Constant communication is needed within the organisation to appropriately frame the true delivery capabilities of automation. While software providers are apt to showcase an ‘anyone can do it’ message, the realities within an enterprise setting, is much more nuanced. Given the flexibility of this technology, it must be built with the partnership and guidance of internal teams such as Audit, Security and various compliance organisations. Therefore, in the event that an automation practice decides to federate development to SMEs, there must remain a strong control from the centre in order to enable such a methodology not to run afoul of local governance standards.
• o Moving from a single to multi-vendor paradigm post-POC when the program is ready to scale was another challenge for the team. This was tackled at Mars through an RFP process to select multiple vendors who provided a breadth of service-types. Mars settled on three particular vendors who fit different spaces within the solution space and included a technology integrator, a boutique automation-focused organisation, and a large well-known consultancy. This range of capabilities is put against the various incoming projects that the hub intakes, enabling the hub to direct the appropriate resources to the challenges at hand.

Success Factors

A core to the success of the Enterprise Automation Hub was well-defined roles, including:

• • Automation Director—Sets strategic direction, technology selection, evangelisation, adoption and awareness.
• • Engagement Lead—Leads incoming demand, initial scoping, stakeholder & vendor management.
• • Development Lead—Leads development and standards, initial designs, code reviews and developer management.
• • Infrastructure Lead—Leads hardware and software management, compliance/security, support and maintenance management.
• • Program Manager—Leads program management, alignment of partner ways of working, Developer Project Manager oversight.

A senior sponsor that is actively engaged and drives the program as a visible and strategic step change in the organisation’s approach to work. The senior sponsor ensured the following:

• • the program upskilled and then enrolled key stakeholders to a shared vision for RPA with clear, agreed and measurable articulation of value and progress.
• • provided oversight of the appointment of champions of RPA in each line of business ahead of delivery to ensure deep and rapid customer partnership.
• • regular communication of progress and learnings to key stakeholders and the organisation to maintain focus and awareness as a change program.

Proactive engagement and planning for Control/Compliance/Governance partnerships before development:

• • The Enterprise Automation Hub engaged with these teams during the very first days after the program was initiated. Having as strong partnership with such teams has enabled a quick build-out of a Corporate-Way of delivering automations. Topics needing definition through these partnerships range from how to create and manage ‘bot’ user accounts, setting access strategies for corporate systems, and enabling the auditability of the automation solutions.

Although many automation software vendors teach the benefits of having the automation capability sit within business operations, Mars’ experience was that being closest to the technology-enablement teams allowed for the most agile ways of working. Benefits of this included:

• • Sitting in Digital Technologies means being closer to potential enablers (to your hardware, systems access, support teams, etc.).
• • Enables awareness of technical challenges the business is facing and fosters engagement with those teams to discuss how automation may support potential solutions.
• • Brings greater sustainability and effectiveness of automations by leveraging business continuity and operational best practices of Digital Technologies in the design of the hub and automations.
• • ‘Digital Technologies’ is global in reach, while business operations tend to be self-contained to the particular unit. Truly scaling an automation capability requires a global perspective.

Business Value Delivered

How well did robotic process automation deliver on the initial business case?

Financial—During the first 15 months the program exceeded all financial and efficiency targets set for the program and was more competitive and faster to capture those returns than for traditional application investments. As the hub matured, simple principles were developed like setting a minimum payback period of 1-year, a minimum threshold of committed efficiency gains, etc. for all automations to be approved. This simplified approach accelerated the viability check before development begins and increased the quality of proposed use cases from our customers.

Operational—The Enterprise Automation Hub delivered automations which helped the business in terms of improved quality, faster delivery, and overall better compliance. Such examples included a bot developed to remove employee access from financially significant systems upon their leave from the company, which could not be automated through standard access systems. This was key from a compliance standpoint for Mars. Additionally, another bot helped to ensure that recipe and ingredient profiles of Mars products were transferred from a legacy system to a more modern system accurately, providing improved quality and faster delivery of other downstream projects. Lastly, an operational finance bot improved Mars’ cash position by gently reminding customers of overdue or late invoices—this bot allowed Mars to see an uptick in more timely remittances.

Strategic—The Enterprise Automation Hub delivered automations which enabled improved strategy and better customer retention. One such example was a bot, which creates Factsheets, detailed product specifications, which are sent to suppliers in order to add a new item to the marketplace. For instance, a seasonal chocolate product that needs to be added to the shelves of a grocer/distributor. The bot allows improved strategy execution so that Mars can get into different markets and onto shelves of suppliers more quickly, supporting revenue expansion and customer satisfaction. The bot also indirectly improves customer retention as suppliers are able to receive the factsheets in a timely fashion to add products on their shelves, benefitting their own top and bottom lines; originally, one factsheet creation could take one to two weeks to assemble manually, but the bot can create these within an hour—improving Mars external selling time.

Lessons Learned

Structuring an Enterprise Digital Transformation strategy, with a component of that strategy being automation, enabled much greater benefits to accrue to the organisation, particularly by leveraging User Centricity methodology. Further support from analytics and cognitive (AI/ML) capabilities again further the value-adding potential of automation.

Ensure hiring deep automation delivery expertise on the team as an early priority, in order to accelerate early solution delivery while partnering with vendors whose ways of working in this rapidly expanding capability are not yet standardised across the industry.

By hiring experienced professionals on the core team that had gone through the initial setup process before, you can avoid the trap of low-value POCs and/or low-quality automations early in the cycle that can erode customer confidence in the hub’s ability to produce business critical automations with tangible and meaningful business value.

Building automations with scale in mind, particularly in a multi-vendor environment, requires a ‘development playbook’ to be created by the hub and shared with all delivery partners (external vendors and internal federated developers alike). This development playbook again requires deep expertise to be created, furthering the need to hire appropriate talent as a starting point. This is not a typical artefact provided by firms working on POCs for your organisation.

Cognitive Automation

Given Analytics (inclusive of AI and ML capabilities), became a core capability as part of the Digital Engine framework. The Enterprise Automation Hub has been tightly integrated with both the capabilities and platforms in this space. Mars is using cognitive automation widely in many analytics use cases and increasingly embedded into applications. Use of AI/ML models connected with RPA has begun in a limited way, primarily focused on the ingestion of third party documents (invoices, POs, deductions management, etc.). Mars continues to explore different solutions that enable not only data extraction from document images (OCR/ICR), but also technologies that enable the tools to ‘learn’ (AI/ML) to recognise and categorise different forms in an automated fashion as well as extract the appropriate data from those forms. Given the global nature of Mars, such technology is used across many different languages. Mars has also piloted AI tools designed to support automated process discovery, including usability issues that drive manual work, as a way of identifying untapped RPA usage.

Automation and The Future of Work

While very hard to predict, it is likely that, by the end of 2021, RPA will be fully deployed across all business functions and geographies against the preponderance of high value use cases at Mars. The role of RPA in the automation layer for operations will likely enter a more traditional ‘continuous improvement and maintain’ phase of the lifecycle. Ongoing application modernisations are likely to retire some automations, and create a need for new automations as well. With significant portions of manual tasks being moved to the automation layer, it may then be possible to develop more advanced and intelligent end-to-end automation use cases.

Our Insights From Mars

In many ways the Mars experience replicates the kind of starting points and trajectories for RPA supported by our research. The link between digital transformation, RPA and cognitive was established from the start, organisationally, technically and operationally. Mars recognised that scaling and integration were the long-term goals, leading to the big business payoffs. Mars realised that RPA needs more centralisation of resources and capabilities than many organisations have liked to think. Mars also recognised the internal in-depth skills and capabilities needed for digitisation and automation—another area too many organisations do not invest into sufficiently.

Mars also moved to a more agile/prototype development model—again a common feature amongst the successful organisations we have researched. More unusually this was allied to their User Centricity tool designed to encourage users to rethink, rather than just replicate existing processes and ways of doing things. Mars also took action to keep user requests focused on business value. These approaches, allied to constant communication with the business user base, seemed to have driven user energy and also accelerated business results.

Mars also used a multiple vendor paradigm, using a systems integrator, automation vendor and a major consultancy to help scale their automation projects. This again has been a regular pattern in our research.

Mars’ critical success factors are consistent with our findings. The Enterprise Automation Hub had well-defined roles. A senior sponsor was actively engaged and drove the program as a visible and strategic step change in the organisation’s approach to work. Early stakeholder buy in and constant communication were needed. Proactive engagement and planning for Control/Compliance/Governance partnerships took place before development.

One interesting angle is Mars’ decision to locate RPA/cognitive automation activity within the IT group (Digital technologies). Experientially, this seems to work for Mars, and practitioners will want to contemplate the reasons given above for this decision, and why it has worked out.

Mars seems well set up to exploit in tandem the RPA, cognitive and digital technologies as they come on stream, and have well-developed mechanisms for ensuring business opportunities and needs that these technologies can serve are identified and delivered on. Mars also indicates the huge amount of managerial work needed, and the challenges faced, to set up an organisation to become strategic with their emerging technologies

Note: This case has been updated from our research for Willcocks, L., Hindle, J. and Lacity, M. (2019), Becoming Strategic With Robotic Process Automation, (SB Publishing, Stratford-upon-Avon, UK).