The utopic machine will have zero fulfillment time and an infinite variety of product options. Fast fulfillment is not easy and requires huge capital investments and deep modeling knowledge. Existing assets frequently are a disadvantage. Proceed with seven investigations to initiate the fast fulfillment project: Speed, Brownian Uncertainty, Physio-Digital Innovation, Streaming Data, Transfer Function Math, Automation Segments, and Decision Models.
Fast, faster, fastest. Smart, smarter, smartest. Choices, more choices, most choices. These are the goals of a design-build fulfillment team. In the preceding chapters, we investigated and explained strategies pursued by the innovators and disruptors to build some of the fastest, smartest, and most choice fulfillment machines. Now it’s your turn, and this chapter describes activities to maximize the probability of success.
Zero-Infinity: The Utopic Machine—Fulfillment occurs instantly or in zero time. All possible configurations of the deliverables are possible, or an infinite product/service selection.
For most businesses, a zero-infinity machine is neither possible nor is it necessary. The first step in machine building, actually step zero, is to describe the zero-infinity target for your business. Do not expect this to be a precise or specific target, rather plan for a moving target that takes shape as you learn more about disruptive trends, technologies, and the possible pathways to achieve customer pivots. Assemble a zero-infinity team to collaboratively investigate and calibrate the zero-infinity target for the business.
Business and the management teams are generally organized to view all new initiatives or innovations as a project. This is a proven strategy that is easily implemented through well-known project planning tools. The team is focused on a specific endpoint, has a specified execution timeline, and a development budget. But this strategy does not play out well in a fast fulfillment innovation project. The endpoint is fuzzy in that the team may only have a 10 percent idea of what the endpoint and its associated processes are. This makes timelines, budgets, and even skill requirements just a set of guesses. Frequently, we will read stories of how successful founders innovated continuously and changed the scope and endpoint when a better idea or pathway came along. In contrast, established businesses operate in a more structured setting. To increase the probability of innovation success, it is, therefore, necessary to first increase endpoint visibility. This can be achieved through a series of step-zero investigations. These are designed to collect data and knowledge that are then used to fertilize the design-build innovation process.
The Seven Investigations
Earlier we learned that functional and process innovations, as opposed to inventive innovations, are the primary value creation drivers in a fulfillment machine. A necessary condition, therefore, for building a fulfillment machine are deep insights and knowledge into what functionalities excite your customer base, and what process inefficiencies or even lack of capabilities are limiting market growth and opportunities. Here are two mistakes that can derail our machine building efforts: Mistake #1—we institutively assume that we know the answers, the reality though is we only have a historical or experience view, not an outside or disruptor view. Mistake #2—we tend to focus on the product/service as opposed to the process for fulfilling the customer need.
Collectively, the earlier chapters introduced and presented methods and strategies for building a fast fulfillment machine. These methods and strategies are the focus theme for seven investigations (Figure 10.1) businesses must conduct as they initiate their fast fulfillment project. These investigations define and shape the unknown knowledge space within which the design-build team will conduct their zero-infinity innovation explorations.
Figure 10.1 Fast fulfillment—Seven investigations
The first two investigations, SPEED and BROWNIAN UNCERTAINTY, provide an initial specification of the target endpoints and associated performance goals for the fulfillment machine. Meeting these targets will be a necessary condition for the project to succeed. The subsequent investigations identify the knowledge needed, available technology, and the underlying transfer functions.
The investigations are organized into a set of four to six questions and are intended to probe how the current business operations can be transformed or expanded through disruptive innovations. Each investigation includes a prospective view sourced from a proven disruptive innovator or innovation. This is not the time to study retrospective views, history can be a handicap in innovative machine building.
Investigation 1: Speed
The primary seduction of the physical-internet-of-things is speed. It allows customers to do things immediately, from anywhere, and at any time. The six investigative questions (Table 10.1) focus on identifying the speed, which will disrupt the current business environment and ensure fulfillment success. The investigative results are critical in setting the target endpoints. A conservative approach will only provide only incremental, the investigation must be aggressive in their approach to identify the correct innovative speed targets. The speed targets must be backed up with data and projections that convince the doubters and excite the innovators.
Table 10.1 Speed investigation
1. SPEED: Investigative Questions
A. W hat is the necessary speed for success? Quantified in four themes: (i) Time Efficiency, (ii) Cost Efficiency, (iii) Quality Enhancement, and (iv) Provider Expansion.
B. W hat is the speed limit at which customers will start pivoting to the fast fulfillment product/service?
C. W hat is the optimal position for the product/service in the frequency/immediacy map?
D. What are the key friction points limiting our current processes?
E. What are the major speed bumps limiting our current processes?
F. W hat are the incremental or enterprise solutions? What is the confessional rational for why this solution does not meet necessary speed?
Prospective View: The operational speed of a fulfillment machine is the culmination of many innovative ideas. If the target speed is high, then the innovators will be driven to design-build the needed ideas. Sir Jonathan Ive was chief design officer at Apple and worked closely alongside Steve Jobs in developing the iPhone. Here is what he has to say about the idea to reality journey.
How Does a New Product Come About At Apple?24—What I love about the creative process, and this may sound naive, but it is this idea that one day there is no idea, and no solution, but then the next day there is an idea. The nature of having ideas and creativity is incredibly inspiring. There is an idea that is solitary, fragile, and tentative and does not have form. What we have found here is that it then becomes a conversation, although remains very fragile. When you see the most dramatic shift is when you transition from an abstract idea to a slightly more material conversation. But when you made a 3D model, however crude, you bring form to a nebulous idea, and everything changes—the entire process shifts. It galvanizes and brings focus from a broad group of people. It is a remarkable process.
Investigation 2: Brownian Uncertainty
Uncertainty and a multiplicity of options are inherent in modern fulfillment systems. The machine must recognize and respond rapidly to all requests that enter the system. The three investigative questions (Table 10.2) identify the range of options and the associated uncertainty that characterize the new normal customer base. The innovative ideas, design constructs, and structural elements of the fulfillment machine will be motivated by the results of this investigation.
Table 10.2 Brownian uncertainty investigation
2. BROWNIAN UNCERTAINTY: Investigative Questions
A. How are the Brownian catalysts changing or creating new customer needs? Described in terms of the four catalysts: (i) An unbounded selection, (ii) Crowdsourced information, (iii) End of the normal distribution, and (iv) Random is the new normal.
B. W hat is the Brownian multiplier effect on the existing/proposed business process or system? Described in terms of the four multipliers: (i) What, (ii) When, (iii) Where, and (iv) Like.
C. H ow are streamlined flows limiting responsiveness to Brownian flows? Describe the speed bumps and friction in terms of: (i) Higher costs, (ii) Decision complexity, (iii) Process complexity, and (iv) Delivery complexity necessary.
Prospective View: Shopify has been a powerful innovator in online retailing and over one million businesses in more than 175 countries have used it as a disruptive tool to take their business online. The founder Tobi Lutke has designed a fulfillment machine that adapts to the unique needs of each of its million retailers. In terms of fulfillment efficiency, Shopify is possibly the closest competitor to Amazon. In a recent interview, Tobi provided insights on seeking process improvements and make trendline predictions in an Internet-driven uncertain market.
Build an Online Business—No Matter What Business You Are In: Three kinds of processes85—There is a kind of process that makes things that were previously impossible to do, possible. That is good. Then there is a kind of process that makes something that was previously possible significantly simpler, which is also good. And then there is everything else. I bet you 99.9 percent of all process that exists in corporate America is the third category, which is just telling people to behave slightly different from what common sense tells them to do. Programmatically predict the unpredictable86: How am I going to predict the future? What I am going to do is I’m going to work very hard on understanding what everyone else, everywhere in the world, has already figured out. And then take those trendlines and try to spin them into the future and figure out what possibility space I have for the future. And then I am trying to figure out how the idea of entrepreneurship fits into that, and how does retail fit into that. What expectations change once some of those things that might currently be very nascent but are sharply increasing in steep trendlines; once they grow into something that everyone does? I think if people were honest about the future, they would admit that that is exactly what they are doing because, again, the future, itself, is completely chaotic and is unpredictable. It is just that you can extrapolate it sometimes.
Investigation 3: Physio-Digital Innovation
Fast fulfillment requires an intelligent and controllable cyber-physical infrastructure that adds intelligence to the machine. The four investigative questions (Table 10.3) identify the touchpoints and entry ramps that link the machine to the Internet of things. This is a critical investigation that will seed the innovation pathway and map key data and decision waypoints. Current methods and solutions promote incremental progress and equilibrium, the investigation must circumvent the current status-quo and motivate innovators to put on their disruptor hats.
Table 10.3 Physio-digital innovation investigation
3. PHYSIO-DIGITAL INNOVATION: Investigative Questions
A. W hat are possible physical pathways a disruptor/disintermediator can follow to efficiently respond to the effects of the Brownian multipliers?
B. H ow could subscription plans to optional products/services facilitate an effective leveraging of the Brownian catalysts?
C. Is there a ready compensator idea that incorporates one or more physio-digital innovations based on identified physical and digital pathways?
D. Is there a Nash equilibrium driving innovation complacency among the existing industry leaders?
Prospective View: Softbank and its founder Masayoshi Son is among the most significant funders of innovative disruptors. This is their plan for the utopic future, review, and discusses it in the context of your business as part of investigation 3:
Softbank Vision—The World 30 Years from Now87—You can own storage that can store a virtually infinite amount of information, knowledge, and wisdom. Digital will become the norm for information vehicles. Not only every electrical appliance but also shoes, glasses, everything will have embedded chips. They will be connected to each other through a limitless cloud and a super-highspeed network. This will provide a whole other level of emotional experiences through “seeing,” “learning,” “meeting,” and “playing.” Not only education but also medicine and work style will change fundamentally. Under such circumstances, the SoftBank Group is committed to accumulating all knowledge and wisdom of humans and artificial intelligence in the cloud, making it the largest asset of humankind. We want to revolutionize people’s lifestyles by working with like-minded companies with cutting-edge technologies and superior business models.
Investigation 4: Streaming Data
The quants will rule, and the fulfillment machine will be data-driven. But data are like bread in a bakery, it becomes stale and hard by the end of the day. It is best when warm and fresh, and that is what you need fresh streaming data. The five investigative questions (Table 10.4) explore how data streams will power the disruptive processes of fast fulfillment. All data have hidden value or utility that needs to be discovered. The disruptors are magicians who know how to conjure this value, and you need to learn this trick too.
Table 10.4 Streaming data investigation
4. STREAMING DATA: Investigative Questions
A. For which business processes has latent or limited data availability restricted decision capability? What was the performance impact of these restrictions?
B. Specific to the identified business processes, what are the needed nX data points and their associated active data frequency?
C. W hat are possible additive business objects (B-Objects) and controllable activities that would accelerate fulfillment times of the targeted business processes?
D. W hat crowdsourced data streams could be integrated as speed catalysts in the fulfillment machine?
E. What technological constraints may be experienced in collecting the data streams?
Prospective View: Venky Harinarayan and Anand Rajaraman are two well known, but not publicly famous, data capitalists in Silicon Valley. Why do I call them data capitalists? They used data as the currency to build the machine and ultimate success. Working alongside Jeff Bezos, they innovated many of the data-driven models that are integral to the Amazon fulfillment machine. Summarized in the following are some of their data capitalization strategies. Review and investigate if they can be applied to leverage streaming data in your business.
Venky and Anand Data Capitalists: Model lite and data rich88— More data almost always beats better algorithms, so when you have developed your parameterized model don’t throw the data away. Make decisions in a collaborative approach with both the model and data, this will provide unique solutions for every instance in a Brownian world. Data dashboard—Provide process innovators with grab-and-go data and keep collecting more data including complementary data. Data provide instant proof, the idea is already validated in the first presentation. The naysayers and doubters with soft roadblocks cannot stop the data-backed ideas. Materialize Decisions89—Partition the problem into a hyperspace. Then use heuristics to materialize or precompute the decisions for the most common subproblems. This will allow the system to make the instant decisions required to operate the VLSF models.
Investigation 5: Transfer Function Math
The well-known design principle, KISS: Keep it Simple Stupid, later also became a management mantra. But KISS has limited applicability in today’s data and Internet-driven world. MATCH: Model and Transfer Complexity to Algorithms is the mantra of fast fulfillment. The four investigative questions (Table 10.5) are designed to explore input variables, output parameters, and management’s ability to control the relationship through the right decisions at the right time.
Prospective View: Uber disrupted one of the oldest service industries, taxis, or the ride-hailing business. Later, Uber expanded into the food delivery business. These have historically been very low technology businesses with almost no innovation (except for Pizza delivery). Uber used the full power of technology including advanced math modeling to radically disrupt the industry. Described in the following is how the seemingly mundane task of food delivery is controlled by a complicated transfer function model.
Table 10.5 Transfer function math investigation
5. TRANSFER FUNCTION MATH: Investigative Questions
A. What input process variables will enhance our ability to model the transfer function, which describe the underlying physics of current and future business processes?
B. What controllable decision variables will enhance our ability to manage and improve the output performance of transfer functions driving current and future business processes?
C. How could probability theory be integrated into decision control methods so as to leverage the process uncertainty into a competitive advantage?
D. Is there an immediate example of a Lite-AI solution that could significantly the performance of a current and future business processes?
Michelangelo—Uber’s Machine Learning Platform90—Predicting meal estimated time of delivery (ETD) is not simple. When an UberEATS customer places an order, it is sent to the restaurant for processing. The restaurant then needs to acknowledge the order and prepare the meal, which will take time depending on the complexity of the order and how busy the restaurant is. When the meal is close to being ready, an Uber delivery partner is dispatched to pick up the meal. The delivery partner needs to get to the restaurant, find parking, and walk inside to get the food, then walk back to the car, drive to the customer’s location (which depends on route, traffic, and other factors), find parking, and walk to the customer’s door to complete the delivery. The goal is to predict the total duration of this complex multistage process, as well as recalculate these time-to-delivery predictions at every step of the process. On the Michelangelo platform, the UberEATS data scientists use gradient-boosted decision-tree regression models to predict this end-to-end delivery time. Features for the model inputs include information from the request (e.g., time of day and delivery location), historical features (e.g., average meal prep time for the last seven days), and near-real-time calculated features (e.g., average meal prep time for the last one hour).
Investigation 6: Automation Blocks
Human capital costs are a constant focus of Chief Financial Officer (CFOs) simply because they represent for most businesses the largest discretionary costs. For a zero-growth company, a common management strategy is cost reduction through headcount reduction, followed by a short-term cheer on Wall Street. Here the quest for automation is directionally opposite to a headcount reduction, or to put it bluntly, a retreating strategy. Automation is the way to design-build those high-growth process. Furthermore, humans are bounded by their decision-making speed and the ability to process a set of fast-changing variables. The four investigative questions (Table 10.6) are designed to first look at what previous ideas were limited by headcount constraints. This ensures that future ideas are not a casualty of the same constraints. Innovators view automation as an opportunity and the investigators will find those opportunities as they answer the questions.
Table 10.6 Automation blocks investigation
6. AUTOMATION BLOCKS: Investigative Questions
A. What activities in planned/future business processes were not pursued due to the projected need for high levels of manual labor or manual decision control?
B. What situation attributes are likely to generate automation challenges in the face of Brownian uncertainty associated with future business processes? Described in terms of the four attributes: (i) Randomness, (ii) Factor Complexity, (iii) False Signals, and (iv) Congestion.
C. What technology enablers could facilitate alpha and sigma automation solutions, which in turn result in faster fulfillment speeds?
D. Are there any disruptive moonshot opportunities or canopy blossom ideas that can only be operationalized with an accompanying automation project?
Prospective View: Marc Andreessen is a legendary venture capitalist in Silicon Valley and has been involved with several very successful innovators, including Facebook, Lyft, and Slack. Marc created the highly influential Mosaic Internet browser and cofounded Netscape. In a 2011 Wall Street Journal essay, he coined the phrase software is eating the world. Summarized as follows is a 2019 revisit of the phrase and future disruptive trends. It says if it can be automated, then automate it, else someone is going to do it first.
Software is Eating the World91—So there is a 70-year journey to basically get everybody on a computer, and everybody on the internet. OK, how does this unfold from here, across industries? I describe it in three claims. First claim is that any product or service in any field that can become a software product, will become a software product. And so, if you are used to doing something on the phone or paper, that will go to software. If you have had a physical product, answering machines, or tape players, boom boxes, like, all the things Radio Shack used to sell. They are all apps on the phone. If it can become bytes, it becomes bytes. Bytes are zero marginal cost, so they are easy to replicate at scale, and become much more cost-effective. The next claim is that every company in the world that is in any of these markets in which this process is happening, must become a software company. Any company that deals with customers, especially consumers, is going to have to radically up its game, in terms of its ability to build the kinds of user interfaces and experiences that people expect these days. And then the most audacious claim is, as a consequence of one and two, in the long run, in every market, the best software company will win. And that does not necessarily mean that a new company that starts as a software company entering an existing market will win, but it also does not necessarily mean that an incumbent that adapts to being a software company will win.
Investigation 7: Decision Models
Speed is a necessary condition for fast fulfillment. Multiply that with complex transfer models and streaming data, and you have a decision-making system that is outside the realm of the human envelope. The only solution is a computerized network of intelligent decision models driven by smart algorithms. The four investigative questions (Table 10.7) list specific challenges the business faces in developing an automated decision system. Next, the questions identify the likely speed acceleration the system will provide. A multitude of solution providers are willing to help with tools and application knowledge and the investigation must shortlist those most likely to help with building the innovative machine.
Table 10.7 Decision models investigation
7. DECISION MODELS: Investigative Questions
A. What are the likely complexity challenges in building decisions models for fast fulfillments? Discussions should be in the context of (i) frequency, (ii) variance, and (iii) fulfill locations.
B. What resources queues will slow down the fulfillment speed and cumulatively increase flow time slack? How can automated process control decision models shrink these queues without additional resources?
C. What are the needed data tracking and cloud computing capability needed to build and implement fast decision control models? Described in terms of software tools, hardware needs and service vendors.
D. What manpower skills do we need immediately if we are going to build a fast fulfillment machine?
Prospective View: Tesla and its founder Elon Musk are revolutionary, and at times arrogant, disruptors in the biggest industry of them all, automobiles. Full Self-Driving (FSD) vehicles are possibly the most complex decision problem in engineering. The autopilot system experiences a hyper Brownian environment and the consequences of a decision error are severe. Tesla plans to achieve FSD soon and plans to distribute FSD as a software upgrade to all recently built Teslas. Their game plan shows how disruptors simplify complex problems through decision models.
Tesla Autopilot Game Plan:92—We develop and deploy autonomy at scale. We believe that an approach based on advanced AI for vision and planning, supported by efficient use of inference hardware is the only way to achieve a general solution to full self-driving. Autonomy algorithms drive the car by creating a high-fidelity representation of the world and planning trajectories in that space. To train the neural networks to predict such representations, algorithmically create accurate and large-scale ground truth data by combining information from the car’s sensors across space and time. Use state-of-the-art techniques to build a robust planning and decision-making system that operates in complicated real-world situations under uncertainty.
Before starting the investigations, a few questions must be answered. The first is what? The investigation must be directed by a threat or opportunity-driven business purpose, or an aspiration, typically associated with an internal or external product/service. It could be small or big, abstract or specific, but it must provide a broad not focused direction for the investigation. Here is one big and one small purpose statement for two of the largest companies in the world, at the start of their machine building journey.
Amazon was launched in 1995, with the mission “to be Earth’s most customer-centric company, where customers can find and discover anything they might want to buy online, and endeavors to offer its customers the lowest possible prices.”
Mark Zuckerburg created Facebook in 2004 with the design goal of a campus wide social media website to connect Harvard students with one another.
The second question is who should be on the zero-infinity team? Greg Christie, who led the software development team for the first iPhone, recalls that it was a shockingly small team.93 Likewise, many start-ups developed their first ideas with less than five members. Small teams can achieve higher levels of communication and are far more likely to have the same level of investigative focus. Listed here are the ideal specification for the zero-infinity team:
• Team size should be between three and five members. The same team must do all seven investigations, but associates should be added as needed to complement a specific investigation.
• Members should have complementary not overlapping skills. Three skill sets must be covered by the team, domain knowledge, technology insight, and data processing.
• Members should not have a confirmatory bias. That is, they are likely to seek and interpret information in a way that supports their existing beliefs.
The third question is how the team should proceed. There is no right way and the recommendation is to follow an approach that works with team dynamics. Here, though, are a few process suggestions:
• The seven investigations should be done sequentially. The sequence was designed for information rollover.
• There should be frequent meetings, ideally every day. Many disruptive innovators were collocated. If this is not possible, all possible online tools should be used for effective interactions.
• The process should be relatively short. The target should be to complete each investigation in one week and the entire activity wrapped up in two months.
• Investigations should be democratized, and all member views and investigative results must be reviewed and analyzed.
• The answers should be brief, one to two pages per question. Data should be generously used to illustrated trends and projections.
• Avoid the boundaries of key performance indicators (KPIs) they can be false signals that steer the investigation away from the disruptive innovations.
The final output is an investigation report that identifies and confirms there are disruptive opportunities in the business. The report also confirms that these disruptions are doable, possibly requiring tools and methods that we have never used before. The report should have a summary page upfront, which presents the findings in a sort of VC (Venture Capitalist) pitch mode. Briefly stated, (i) Market Opportunities, (ii) Technology Enablers, and (iii) Potential Value. The next chapter provides additional details on how the investigation and accompanying report integrates with the overall design-build project.
• In the Utopic Machine, fulfillment occurs instantly or in zero time. All possible configurations of the deliverables are possible. The first step in machine building is to describe the zero-infinity target for the business.
• Seven investigations that initiate the fast fulfillment project are introduced. These investigations define and shape the unknown knowledge space within which the design-build team will design-build their zero-infinity innovation.
• Specifications and action plans for the investigation team are presented.