Analytic-driven resilience in a crisis – part 2

Observations by Dinand Tinholt, Goutham Belliappa, Lynn Lang, and Kathleen Junkiness

Part 1 of this series focused on enterprises that have supply but have lost short-term demand (Group A). These companies find their demand slow to return to previous levels and experience immediate impact of the anomalous situations for the next quarters. Our previous article focused on 10 imperatives we see in the market to move across flattening the impact of the crisis through to business recovery and a new normal.

Part 2 focuses on enterprises that face a surge of demand, compounded by massive and global disruptions to supply. These “Group B” companies have an urgent need to more effectively leverage analytics to learn from the crisis period, and restart with a mind to managing ongoing volatility and spikes in demand given supply-side disruptions, and to better serve the current needs of consumers in unusual circumstances. As we did with Group A, for Group B companies we examine their top imperatives from four categories: Employees, Performance, Demand and supply resiliency, Enterprise communications and brand management

Employees 

Imperative 1: Employee analytics to protect, preserve, and manage frontline employees 

For headquarters staff, the challenges presented by working remotely, defining contingency plans, and maintaining morale – are difficult but manageable, given ample technological solutions. However, in any service deemed “essential,” enterprises and governments grappled with unforeseen demand and quickly redeployed all available resources to service customers on the front line. New store hours require new shift schedules in stores. Unusual peaks in demand both in store and online; call for deployment of head office workers and priorities to support stores and warehouses. Redeployment is rapidly moving across sectors as certain skills and backgrounds (ie. medical) are identified to support (with consent) more vital triage in hospitals and medical facilities. Likewise, individuals with food safety backgrounds now have crucial, transportable skills that are redeployed to grocery and pharmacy. In late March, Aldi and McDonald’s made a staff-sharing deal in Germany that heralded employee leasing at a new scale: a format that almost simultaneously was redeployed across geographies and across sectors around the world.

Figure 1: Food stocker in a grocery store – Source McKinsey 

Hero Pay Programs, part of triage mechanisms to retain staff, heralded media campaigns in every corner of the world to appreciate the services of front-line workers. What do these programs mean as we look forward to restarting and taking stock of the past weeks? We see enterprises continuing to grapple with:

Unprecedented demand for training: For front-line workers, training in prevention measures on the use of material (gloves, face masks, hygienic gel, wash hands constantly, etc.) is essential, but so are ways to track completion, adherence, and competency to drive consumer confidence. While organizations adjust to implement new standard operating procedures, tech startups are also innovating around the challenge. Stopcovid.co is a training services startup that sends text messages with the latest updates to front-line workers in real time. For a more extreme example, Voxel51 is using live, pre-existing cameras to track how preventative measures use artificial intelligence to give a window into social activity in popular public spaces. 

In education, front-line workers are teachers and remote learning is the new bonanza for edtech startups. Meeting unprecedented demand has led online learning solutions to offer free access. Meanwhile, their valuations are rising. As an example, India’s Byju’s online tutoring platform jumped to $8B January, while SnapAsk, which allows students to take pictures of their homework to get help from tutors, successfully completed a funding round in early March.

New algorithms for Workforce scheduling and shift management: Given the contagiousness of respiratory diseases such as the flu and coronavirus, we see the emergence of cohort scheduling to reduce cross-contamination viruses across groups (cohorts) of personnel, not to mention that algorithms may need to be updated to account for higher absence ratios that normally applied for family leaves.

Imperative 2: AI-based people analytics to understand current skills, cross-skill ability, and skill-based adverse event scenario modeling and vulnerability 

Crises make an already complex talent landscape even more complicated across dimensions of skills, shifts, timing, labor laws, geography, availability and so on. “Making the right talent decisions requires insight, and that requires data — lots of it —scientifically used and applied with good judgment.” Failing to address the dynamic nature of the talent landscape is a risk to both the HR function and the viability of the organization. No doubt, intuition still works to start the idea moving, but for the execution of models such as the “resource leasing” model (e.g., engaging fast food restaurant workers with food safety skills), data and analytics are critical to manage the complexity and scale. However, without people analytics, leaders will struggle to find softer skills.

Simple tweaks to augment existing workforce management systems can add tremendous value. Very few companies invested in workforce analytics in good times, but we refer to practices and processes that have already proven valuable in good times that are even more valuable under adverse circumstances. Key parameters include:

1)    Define what matters: Agreeing to outcomes in advance is critical. 

2)    Fill data gaps: Internal sources can provide some relevant data, some data may need to be acquired through new or unconventional means and other data may need to be derived or extrapolated.

3)    Challenge conventional wisdom: Data can often result in “confronting senior individuals with evidence that in some cases contradicted deeply held and often conflicting instincts about what drives success.” Be ready for it. 

4)    Management is a key level: Define what “good” looks like in the data and encourage other leaders to behave in the same way. 

5)    Shifts differ: Understand what kind of shifts are most effective and tailor your work schedule accordingly. 

6)    Skills matter but skills can be taught: Understand skill adjacency from the data and begin cross training immediately as part of the job. 

 Crises present many threats while providing a massive opportunity to leapfrog and provide value. Simple analytical augments to make your workforce management system more active and more transparent can provide immediate value in a crisis. 

Imperative 3: Analytics to enable workforce boundaryless working optimization

Working from home will seamlessly morph into a new normal of “boundaryless working.” Where in the past only 3.6% of the workforce that could work remotely (56% of the US workforce according to some studies and only 29% according to the Bureau of Labor statistics) would actually work from home, we now realize through the current measures that this potential remote workforce is much (much, much) larger. New behavior takes time to become a habit, anywhere between 18 and 254 days, averaging out at 66 days. The habits we are forming today are thus very likely to stick. 

Further enabling boundaryless working will be a key priority for the future – to remain nimble and resilient and to adapt to increased demands from employees. Some companies are struggling to enable a high percentage of remote working. Remote access to critical systems and (data) security are often quoted as key concerns. An additional constraint is network performance, for example the loads that VPN connections can handle. Both during and after the current crisis, analytics (e.g. performance monitoring of networks and applications) can help identify the bottlenecks that have limited remote working to overcome these in the immediate future.

For companies that are demand rich and supply challenged (e.g. telecommunications, grocery, pharmacy, health services) the challenge is to manage uncertainty and volatility. To enable boundaryless working for the workforce in these types of companies, employees need real-time access to the company’s core data to be able to rapidly adapt to changing circumstances – data access and analytics supported by human oversight to adapt to anomalies that can’t easily be predicted by data. 

With constantly changing circumstances, a key question is also whether a company’s workforce is optimally utilized in the right places. Analytics can help monitor workforce capacity and effectiveness to shift resources to areas with peak needs. 

When companies have high demand but supply is challenged, the imperative is to do more with fewer resources in less time. A data-driven, cross-functional approach is needed to understand all the drivers of production or service delivery to redesign them. An analytics-driven “Design to Value” process helps support companies not only meet high demand in less time but also to improve their overall margins. 

In business schools, almost students are taught Michael Porter’s classical model of the Five Forces. This model holds even more true today but can be seen in a different light. The imperative for companies that are supply constrained is to embrace the five different forces and see them as opportunities to change the way products and services are delivered. Analytics can help identify substitute products that a company can produce to meet high demand in markets outside their own (for example a car manufacturer producing ventilators). Equally, it can use for example knowledge graphs to map supply chains – both their own and competitors’ – to identify weaknesses and look for alternatives. 

Performance 

Imperative 4: Graph analytics to understand interactions between employees and customers for physical contact traceability 

While many customers have spent millions on Customer 360 solutions, most have a very limited understanding of the customer’s journey and interaction through their perimeter despite and probably because rich interaction data spread across many disparate packaged siloed systems that often compete on the uniqueness of their silo that reduces interoperability. 

Simple graph-based consolidation of customer and employee data, along with potential contact scenarios, can help a company understand how a customer, prospect or contact moves through their perimeter. This can also help in adverse events such as recalls and contact tracing for contagions. 

Graph-based solutions can help link customer touch points across different silos of the business without breaching regulatory boundaries. Having a good, rich source of customer and employee interaction data can help companies understand journeys and touchpoints that can be optimized in good times for better customer experience and in bad times used to increase automation and reduce physical interaction.

Imperative 5: AI “in-run” performance analysis to provide AI-based assistance for job performance – especially for new hires 

While the measurement of performance is important in good times, it becomes critically important in times of crisis when companies must ramp up tens of thousands of short-, or long-term employees in a hurry. The good news is there is ample availability of comprehensive performance analytic frameworks available for organizations to use. The bad news is that most organizations still operate with gut and intuition and in silos and have a low level of trust in data and management frameworks. 

Figure 2: Wide-Angle Performance Management Framework 

A wide-angle performance management framework combines basic descriptive analytics as well as sophisticated AI and ML to allow continuous monitoring of a wide array of job functions. The setup allows different levels of management as well as the employee to understand how they stand against priority-engineered metrics so they are aware of overall performance in real time and able to ask for help or make adjustments as the work progresses. Target setting and scheduling is handled by AI while descriptive analytics allows end-to-end visibility across configurable hierarchies. Some companies allow peers to see each other’s scores to gamify performance excellence, even allocating regular prizes where appropriate. 

One cannot improve what one doesn’t measure. Transparently monitoring in-run performance of one’s job allows for better employee morale and performance especially in times of crisis when employees of varying tenures may have to work side by side. 

Imperative 6: Analytics to identify cost drivers and innovative ways of redesigning products/services to improve financial performance 

When companies have high demand but are supply challenged, the imperative is to do more with fewer resources and in less time. A data-driven, cross-functional approach is needed to understand all the drivers of production or service delivery to redesign them. An analytics-driven “Design to Value” process helps support companies not only meet high demand in less time but also to improve their overall margins by deploying resources against areas with the most value. 

In business schools, almost all students are taught Michael Porter’s classical model of the Five Forces. This model holds even more true today but can be seen in a different light. The imperative for companies that are supply constrained is to embrace the five different forces and see them as opportunities to change the way products and services are delivered. Analytics can help identify substitute products that a company can produce to meet high demand in markets outside their own (for example a car manufacturer producing ventilators). Equally it can use for example knowledge graphs to map supply chains – both their own and of competitors – to identify weaknesses and look for alternatives. 

Demand and Supply Resiliency 

Imperative 7: Supply Chain visibility including supplier & geographic network vulnerability analytics for goods but also people supply chains 

At their most mature, global supply chains are characterized as highly efficient, globally integrated networks designed to optimize price and cost in an environment of secure supply. Less mature procurement organizations still rely on human intelligence from top-tier and a select few lower-tier suppliers and almost no intelligence at the organizational level on other suppliers. 

Buyers and suppliers face unprecedented challenges in the current environment. A dramatic shift in consumer buying behavior and bursts in demand exceed actually or requested supply . This has become additionally complex as enterprises face geographic & political disruptions to the supply chain. Tactics in advanced procurement have so far been very one dimensional at cost and price optimizations assuming just-in-time(JIT) supply considerations are not only adequate but necessary. Companies have advanced JIT structures further with commercial terms – such as consignment – to balance ownership with value chain partners. In organizations where information is collected via personal relationships where decisions are often based on conjecture, knowledge comes at a premium when procurement personnel leave, change roles, or retire, are ill.

We see three views becoming critical enablers to product availability in the current environment and we see linked to organizational sustainability in the post-covid environment:

Refresh sales and operating procedures; setting the stage for AI and machine learning in the supply chain. Enterprises are on a vertical learning curve in developing new sales and operating procedures as they deal with the crisis. However, even as crisis management actions persist, focus needs to shift to formalizing new processes to help employees cope faster with new realities and balance customer care and service while managing inventory and operating costs. New ways to approach pre-existing challenges are quickly emerging as talent across the organizations proves able to adapt to new ways of working. However, truly enabling innovations evolving from the crisis means that labor-intensive and highly suspect processes for data aggregation and cleansing deemed adequate in a pre-pandemic world are no longer sustainable. Manual analysis and personal judgement need to be quickly paired with hypothesis-based AI initiatives that are mobilized quickly and transparently across the supply chain. Nimble organizations will establish improved data management using data and templated architecture-based framework for quick concept development. These architectures can be established in simple proof-of-value analysis that IT or a data and analytics supplier can run alongside supply chain professionals. A new partnership between supply chain and IT should identify processes that can be automated or prioritized for deep learning algorithms. Enterprises need to accelerate use of open-source technologies to train advanced AI/ML algorithms and inject deep learning in demand planning to innovate through to a post-pandemic new normal. Enterprises need to be wary of packaged providers who may offer point solutions and focus more on concrete views of how their new supply chain data platform strategies are both enabler and competitive differentiator – you don’t want to be dependent on a packaged provider’s feature road-map for your competitive differentiation. The optimistic view is that the supply chain will become more people-centric, and more transparent giving employees the opportunity for analytical decision making as well as consumers who build confidence with their favorite brands to fulfill their demand as it shifts. 

Goals for longtail procurement will shift from delivering new procurement savings to reducing risks in geographical procurement. The long tail of spend has been viewed as reasonable opportunity for web-based procurement systems to deliver new savings to the enterprise. In a supply-constrained business environment, securing the value chain is leading to stronger emphasis on partnerships with local food and manufacturing sources – moving local products into the supply chain and streamlining to bring that capacity to the consumer. To support local suppliers, we see supply-constrained organizations taking a step further to secure supply, with favorable payment terms and other mechanisms to support business continuity. However, ultimately the challenge will still return to data. Only 20% to 40% of the data needed to tender new sources is centrally stored. It is often scattered across an organization in non-standardized formats or through category/department purchasing departments. This results in poor visibility and often unique specifications. Centralizing product data can help help overcome this, which will require enterprises to adopt big & cloud data platform strategies in supply chain with urgency. 

Innovation is surging, creating new analytical challenges to supply as industry seeks to address the onslaught of unmet needs due to social isolation. New industry partnerships are evident in the people and employee dynamic, but product availability in the last mile is also stretching into new operating principles for the supply chain. We’re seeing:

-        The idea of autonomous stores could gain momentum: in a recent Capgemini survey, two-thirds of Millennials were willing to shift purchase to a store with automated technologies. More highly automated stores will also help drive greater efficiency throughout the supply chain, not just at the retail storefront. While the report predicts that increased store automation can help retailers avoid “shrinkage,” stock-outs, and store operating costs, the applicability in the current crisis of providing consumers access to product without putting employees at risk will certainly be tested.

-       Gig-economy workers are stepping in to support the delivery to home service for almost everything – putting additional strain on those businesses deemed “essential,” such as grocery and convenience stores, to plan a new kind of store inventory and putting pressure on traditional formats and even in-stock items.

-       Micro-space plans for merchandise will be stretched into new formats: certainly pop-up stores in hospitals, as well as new partnerships from closed stores to retail in grocery will create new demands on supply chain systems and analytical models.

-       Collaboration tools are being used – by absolutely everyone – in markets in ways previously not considered “addressable.” As schools across the world move to online classrooms, data is being used in ways that is not secure, putting the onus on cybersecurity.

Imperative 8: Dynamic demand planning analytics with a focus on alternative forecasting models and adverse event scenario modeling 

The imperative for companies with a high demand in times of volatility is also to identify what is needed and, more importantly, when. Forecasting analytics can help model future demand patterns and, where necessary, artificially constrain supply in some areas to be able to meet additional demand further in time to ensure continuity. 

In seamless “just-in-time, just-enough supply chains” the vulnerability of such systems is becoming clear in times of great disruption. The current situation makes clear that new models are needed to build more resilience into our systems to adapt to disruptions – whether a pandemic like the current situation or ecological disasters. More black swan analytics needs to be included in forecasting models to ensure sufficient robustness. 

Imperative 9: Analytic transparency around managing customer physical and psychological safety within your span of control  

“If everyone else on the Titanic is running for the lifeboats, you’re going to run too, regardless if the ship’s sinking or not” – Steven Taylor

People are also more likely to engage in hoarding when they are predominantly driven by their intuitive, emotional side — fueled by anxiety, fear, and panic. Even though rationally, most people know via historical data that such shortages will be short-lived, emotionally we simply don’t believe that. The anxiety and worry about food supply shortages is more readily transmitted to others nowadays, due to the immediacy and increased reach of social media. Even if that anxiety and worry is misplaced or irrational, it spreads like its own virus throughout our social media networks. 

Just as in bank run scenarios before the US instituted the FDIC, countries show a surplus of supply to assure customers that there is plenty of supply around. “A visit to one branch showed tellers had stacked bricks of yuan notes immediately behind the glass, piled above head level, and assembled cash piles the size and height of a double bed in the back to show there was enough to go around.” The same philosophy should be used by retailers to assure customers and prevent a run on already hard-to-deploy supplies. Showing shelves full of needed supplies is key, as is anticipating what supplies are needed to show adequate supplies.

Examples to support customer confidence are developing daily across most organizations and currently include:

-       Making people feel safe and secure walking in or interacting with you in the store by putting areas and track-ways (such as one-way aisles in grocery stores); one-metre tape to keep distances in line

-       Doubling down on online commerce capabilities – best practices are to run analytics on eCommerce data to better understand what features and searches are dominant in online shopping behavior and responding to support store inventories

-       Contactless payment and delivery: showing significant growth in the media to give customers confidence that food is being prepared safely and delivered without requiring touch.

-       Active demand forecasting linked to behavior and news, but also supported with active restocking.

Demand and Supply Resiliency 

Imperative 10: Analytics to identify new channels/approaches to interact with customers to gain/maintain their trust and loyalty in times of changed or no customer interaction

In times of abrupt change, customers’ ability to adapt lags behind the changes taking place. Thomas Friedman calls this “dislocation.” His advice is that people need to “reach a state of dynamic stability where we can adapt to a constant rate of destabilization.” Companies need to find a way of both keeping up with accelerated change and keeping pace with their customers. This means using new technologies to remain in touch with clients to build trust and strengthen loyalty.

Many companies have used sentiment analysis tools to understand how customers view their brand. At times when customers still have a high demand for products or services but face constraints in procuring them, sentiment analysis can be repurposed to identify “frictions” that customers are experiencing and adapt to them. Similarly, with constrained supply chains, sentiment analytics can be adapted to serve supply chain insights to understand what challenges your suppliers are facing so companies can proactively reach out to offer support and anticipate bottlenecks. 

Companies have for a long time used analytics to identify their most loyal as well as their most profitable clients. In times of crisis, this insight should be leveraged to go above and beyond the service customers expect. This can mean, for example, ensuring continuity of service through alternative delivery methods, additional discounts to specific customer groups, extended loyalty programs, or perhaps even small signs of empathy through targeted engagement activities. 

 In conclusion… this too shall pass

For customers in Group B that are Demand Rich and Supply Challenged: we see ten imperatives across the four key priorities of Employees, Performance, Demand and Supply Resiliency, Enterprise Communications & Brand Management

A. Employees 

Imperative 1: Employee analytics to protect, preserve, and manage front-line employees 

Imperative 2: AI-based people analytics to understand current skills, cross-skill ability, and skill-based adverse event scenario modeling and vulnerability 

Imperative 3: Analytics to enable workforce boundaryless working optimization

B. Performance 

Imperative 4: Graph analytics to understand interactions between employees and customers for physical contact traceability 

Imperative 5: AI “in-run” performance analysis to provide AI-based assistance for job performance – especially for new hires 

Imperative 6: Analytics to identify cost drivers and innovative ways of redesigning products/services to improve financial performance 

C. Demand and Supply Resiliency 

Imperative 7: Supply Chain visibility including supplier & geographic network vulnerability analytics for goods but also people supply chains 

Imperative 8: Dynamic demand planning analytics with a focus on alternative forecasting models and adverse event scenario modeling 

Imperative 9: Analytic transparency around managing customer physical and psychological safety within your span of control 

D. Enterprise communications and brand management 

Imperative 10: Analytics to identify new channels/approaches to interact with customers to gain/maintain their trust and loyalty in times of changed or no customer interaction

Please reach out to Dinand Tinholt, Goutham Belliappa, Lynn Lang, and Kathleen Junkinessto discuss these or other ideas

Other Image Sources

1. https://techcrunch.com/2019/01/10/resilience-tech/
2. https://www.theworldofchinese.com/2014/01/the-great-money-wall-of-china/
3. https://www.teamsoulsports.com/wp-content/uploads/2016/12/shopping-cart-1.png
4. https://www.hawaiipublicradio.org/post/hawaii-supply-chain-continues-uninterrupted-why-are-store-shelves-still-empty#stream/0