Chapter 13
P&G’s NGS Transformation
The final two chapters in the book build upon my conviction that when it comes to thriving in the Fourth Industrial Revolution, fortune favors the disciplined and prepared mind. The digital revolution is literally the opportunity of a lifetime. It’s up to us to seize it and to be disciplined enough to succeed. To illustrate how all the surprising disciplines to take off and stay ahead can come together, I show how the various jigsaw pieces of the disciplines were assembled at P&G’s Next Generation Services.
P&G’s Quest for Transformation of Its Global Business Services
The Human Genome Project cost $2.5 billion in 2003 to sequence the first genome. Prices have fallen exponentially to the extent that I get junk mail in my in-box quoting $100 to sequence my genome. When was the last time the cost of IT services dropped way beyond 99 percent in fifteen years? And why shouldn’t the global business services industry, which is essentially a data business, not be experiencing exponential capability increases?
This was the question on my mind when I started working on NGS in 2015. P&G’s GBS had industry best-in-class benchmarks, but there had to be a way for us to find the next S-curve of improvement. To stimulate ideas, we talked with more than a hundred organizations—peer shared services organizations, consultancies, IT providers, VCs, and start-ups. One of those happened to be an Australian-based global software company with revenues of $500 million. I asked them if they had a shared services organization. They did not. I knew they were incredibly efficient in their operations, so I persisted in questioning them on how they ran their HR services like payroll, hiring, and performance management. They said that their HR function operated these. My next question was how large their HR organization was across their ten physical country locations. Their answer blew me away—twenty-five people. More incredibly, they read the surprise on my face and thought that number was too high! They said somewhat defensively that half their HR was dedicated to hiring, since the company was doubling their head count each year. While I picked up my jaw from where it had fallen to the floor, I realized that I had stumbled upon an important insight— the next generation of shared services already existed. It was how digitally native companies ran their internal operations.
A Digital Backbone of Operations
The next few months would throw up example after example of why a digital backbone to the company’s operation in the future was not just desirable but was already starting to exist in the new generation of digitally savvy companies. Benchmark costs of IT and shared services as a percentage of revenue in those companies were less than half of those in large companies. A major wealth and asset management company in New York had doubled its employee productivity four times in the previous ten years. The examples mentioned earlier in the book— including x.ai, the robot admin who was able to do calendaring, or bots writing 90 percent of the short online updates on stock prices or sports results, or ShotSpotter, the AI tool used to triangulate the location of gunshots fired in real time using camera information, and many more—suddenly emerged in support of the concept of a digital backbone of operations.
Digital transformation in most organizations will take three forms—new digital business models (e.g., from retail to online selling), new technology-embedded products (e.g., driverless cars), and digital internal operations (e.g., using AI for wealth management). The future of GBS, including the IT function, was to transform itself to be the digital core of operations in the entire enterprise.
Our research uncovered dozens of disruptive possibilities within the shared services industry to support the goal of GBS becoming the digital core of P&G’s operations. The example of travel expense solutions in digitally native organizations, including Google, Adobe, Netflix, among others, was eye-opening. Most traditional enterprises have rigorous standards on where to book flights and which hotels to stay in. After each trip, all expenses are meticulously documented in expense reports—a chore that most travelers hate.
The process followed by some digitally native organizations was radically different. Before each trip, the traveler logged onto a system and entered the destination and the dates of the trip. Based on a huge database of anticipated costs, the system then provided a trip budget. The traveler was free to book their tickets at any site and allowed to stay anywhere. Their corporate credit card had details of expenses already, and therefore the system didn’t need the employee to create an expense report. Further, if the trip budget was underspent, the policy allowed the traveler to put the savings to use in flexible ways—including staying at a posh hotel during the next trip or even donating the savings to charity. These practices not only saved the enterprise money (up to 30 percent) and eliminated noncore activities in the enterprise (e.g., managing travel agencies, negotiating hotel deals, matching expense receipts to claims) but also improved employee satisfaction for having given them an adult business deal.
Over the next three years of my leadership of NGS, we would run about twenty-five projects of similar 10X potential. We would run these as a portfolio mix where some projects were killed if they did not meet speed and financial criteria and others that overachieved their initial potential. As long as the portfolio as a whole over-delivered its goals, we were fine. The possibilities in creating a digital core of enterprise operations were endless. Here are a few examples:
Can you run the supply chain planning operations of an enterprise, end to end from supply to demand, and in real time? The standard practices in the world today treat every process in manufacturing resource planning (MRP) as siloed exercises involving demand forecasting, demand planning, manufacturing execution planning, transportation planning, and so on. The siloed optimization of these processes in large global manufacturing enterprises can result in thousands of employees filling in the gaps. That’s archaic in today’s world, where modern technology architectures can handle trillions of events per hour and use thousands of AI algorithms to optimize plans across the whole system in real time.
How about a “Siri”-like experience for all enterprise systems, with the ability to cut across siloed systems in the enterprise to provide the solution to most needs? So for instance, instead of going to several individual systems when planning for a new hire to join the company (e.g., security badge, salary, facilities, PC and email, training systems, etc.), could we simply say “Hey Siri, set up a new hire, Jane Smith, to start on March 1, 2020” and then have the process executed across all siloed systems after providing a few more details. This same user experience would be available for business transactions as well.
Could AI be used to dramatically redesign accounts receivables processes? Instead of hundreds of employees in large organizations manually executing processes such as deciding which of the disputes for underpayment from customer companies were valid, could we have algorithms make those decisions?
Could algorithms ingest incoming contract documents from suppliers and red-line (highlight) parts of the proposal that were not compliant with policy or were areas for negotiation?
Could financial forecasts be done more accurately by algorithms than by a combination of the traditional forecasting systems and humans?
Could an AI brain for the purchasing function guide buyers in chosen spend pools on their most complex decisions? For instance, keeping up with supplier and industry changes, identifying pricing trends in real time, and even placing spot-buying orders? Or driving more competition among suppliers by ingesting more external data on new supplier possibilities? And in matching ongoing invoices and payments against complicated pricing tables in contracts to avoid overpayment?
Could we disrupt the entire call center experience using algorithms that could translate from voice to text and search vast internal knowledge bases for answers to complex issues, and in the process provide more service and more choices to users?
Could we eliminate 90 percent or more of all IT outages in the enterprise (i.e., from power supply to network, server, database, data quality, or user experience outages) by gathering signals related to the operation of these in a massive data lake and then using algorithms to predict and self-heal most of the issues? And could we do this simultaneously across the world and across all suppliers?
Could complex global ocean shipping and air transportation across countries and suppliers be made more transparent and simple, especially on the current status of location of goods and actual cost of transportation, and possibly even eliminate the need for suppliers to send invoices?
Could video technology be used to do actual “observed” consumer behavior instead of “claimed” behavior by using video algorithms to trawl through massive volumes of footage and provide data based on actual actions of customers? Observing actual behavior is more reliable than asking people how they might behave in a situation. Doing this at massive scale today is a huge challenge.
Could the massive challenge of synchronizing “master data” in the enterprise (e.g., standard codes and values such as the correct weight or the right SKU code for a product anywhere in the world) be finally fixed?
Hundreds of opportunities along these lines exist in the enterprise and are viable ideas. Whether a given organization can execute any, some, or all of them is a different question. It’s the difference between identifying a viable end state and executing digital transformation successfully. The big question is how to transition from a stable, successful current state to a highly desirable but uncertain future state.
The Challenge of Transition
The question of successfully transitioning from current to future state brings us back full circle to the question of why digital transformations fail. The lessons from the NGS journey have led to the creation of the five-stage model of digital transformation. In particular, I would like to highlight the example of three disciplines that played a critical role at the start of the NGS program.
Strategy Sufficiency
The NGS goals were ambitious—we would disrupt P&G’s GBS, but in the process change the entire shared services industry. That raised the issue of strategy sufficiency. How could a small group of P&G folks disrupt an entire industry? The best strategy in our case was to create an ecosystem effect.
First, NGS had to be more than just a P&G-only group. No matter how strong our group of internal resources, a broader ecosystem of resources would always deliver more transformative capabilities than any one company could. We agreed to define NGS as an open ecosystem that would have three groups:
A dozen or so P&G handpicked resources who would design the 10X ideas and implement them in the base organization.
A half-dozen or so current P&G IT partners like EY, Genpact, Infosys, L&T Infotech, HCL, HPE/DXC, Tata Consultancy Services, and WNS. They would scale up the 10X ideas into products.
A large ecosystem of start-ups brought to NGS via ten of the top VCs in the world. They would bring in the latest disruptive capabilities, which could be complemented and “enterprise hardened” by the IT partners and implemented by the P&G NGS leaders.
Second, the ecosystem would have to be based on a win-win relationship for all participants. The value proposition with the P&G IT partners was as follows—they would bring in the resources and the product development funding needed to create the 10X software product. In return, they would get the intellectual property and the rights to sell the products externally to other companies (that didn’t compete directly with P&G). If the product was truly a 10X digital transformative to P&G’s GBS—a best-in-class shared services organization—then it had to be very commercially appealing to others. The value proposition for the start-ups was to get a foothold at an attractive client in return for co-innovating with us. And the win for P&G was to get 10X disruptions at low to no cost.
Third, even this ecosystem of the three groups would not be large enough to disrupt at large scale. We would therefore need a bigger community and a crowd to support it. Most of the ideas as well as execution capacity for NGS would come from these. The “community” included passionate people within P&G and its immediate partners who were attracted by the high quality of the work and wanted to be a part of it. The “crowd” included unlimited resources available from crowdsourcing the work—via universities, start-up communities, and specialist groups like Kaggle for analytics problems.
Iterative Execution
The NGS operating model was defined very early on to be an iterative, high-risk, high-return execution.
First, NGS would focus only on 10X disruptive initiatives. The core organization would take on day-to-day continuous improvement work.
Second, inspired by Alphabet/Google X, the NGS operation was set up to be a portfolio of projects. As owner of the organization, I saw my role to be like a venture capitalist for the projects. For every ten experiments (projects) I took on, I would kill five, expect to have another four deliver only 2X outcomes, and have one turn out to be a 10X disruption.
Third, to create a rapid “clock speed” of operation, we agreed to a general guideline for duration of each stage of work. This was based on the exponential series of 1-2-4-8-16: one month for landscape assessment, two months for design, four months for hypothesis testing, eight months to complete development and all in-market testing, and sixteen months to complete all deployments.
Finally, the operating model itself for the ideation and deployment were standardized to use design thinking and lean startup.
Change Management Model
The chosen model was to create an “edge” organization that would be made up of handpicked, highly credible operational leaders chartered to drive change in the core organization but would operate behind a “cultural firewall” on the initial stages of each project to reward high-risk, high-return experiments.
The choice of who to assign to NGS was made with change management in mind. The dozen leaders were handpicked by the GBS leadership, based primarily on their high credibility within the core organization as opposed to their technical skills or innovative abilities. These leaders worked full-time with NGS but focused on experiments that were high value (usually a $50 million potential or more) and were critical innovation priorities within their individual organizations’ annual strategic plan. To enable the focus on speed (vs. policy compliance) during early stages of each project, the team was allowed to operate behind a “cultural firewall” that promoted smart risk taking at earlier stages of the projects while protecting them from the natural corporate immune system.
By the end of my role after three years in NGS, the actual results followed the 10-5-4-1 model extremely well. Of the twenty-five odd experiments (projects) that were undertaken, four were definitely 10X and eight were 2X to 5X in nature. For every instance of success, there were two or more that failed. That’s where the culture of rewarding “learn by doing” kicked in. There was no failure—only learning, as long as the portfolio effect of all these experiments (projects) continued to exceed overall financial goals. The decision to place the transformation-leading organization at the headquarters in Cincinnati instead of in Silicon Valley was a brilliant move. At the end of the day, digital transformation is less about the technical capabilities and more about systemically changing people’s thinking. That may have been among the biggest side effects of having established NGS—the inspiration to the rest of the organization that transformation was not just desirable, but that each person could actually contribute to it and lead it within their own roles.
Chapter Summary
Any motivated organization can approach digital transformation using the example of P&G’s Next Generation Services.
Digital transformation can take three forms: entirely new business models (e.g., from physical retail to online), technology-driven new products (e.g., driverless cars), and digital operations. The specific case study from NGS was about creating new digital operations.
The challenge for most leaders is how to transition from current state to a desired future state. That’s where the five-stage model for digital transformation can help.