LLamasoft Supply Chain Blog

← Back to Blog

LLamaCon EMEA 2018 Day 1 Takeaways: Digital decisioning, Brexit, and more

By Dr. Madhav Durbha  November 15, 2018

Day 1 of LLamaCon EMEA 2018 was a packed, content-rich event. Here are some key takeaways captured by Dr. Madhav Durbha, LLamasoft’s Group Vice President.

  1. A case for a Digital Decisioning platform: The proceedings kicked off to a great start with a keynote by Razat Gaurav, CEO of LLamasoft, wherein he made a passionate case for speed and agility in supply chain decisioning. Whether it is external factors such as rising nationalism, weather-related disruptions, currency fluctuations, changing consumer behaviors, or internal factors such as M&A, shift to Omni-channel delivery, or rapid product phase-in phase-outs – organizations are faced with more uncertainty than ever. This calls for an unprecedented speed of decisioning driven by scenario planning and tradeoff analysis. However, Razat pointed out that most organizations have a mixed bag of ERP and advanced planning technologies which reinforce the organizational silos along the plan-source-make-deliver-return paradigm. Asking these organizations to rip and replace these systems, however, is unreasonable. Razat talked about a Digital Decisioning platform LLamasoft is building, that can sit on top of these systems to enable unprecedented visibility and decisioning. Such decisioning layer takes advantage of an end-to-end supply chain data model, data harmonization capabilities, visualization, proven algorithmic capabilities, and a visual environment in which business users can author apps in a zero-code environment. Razat said that this capability will democratize the algorithmic horsepower. This will enable LLamasoft’s partners and customers to create intellectual property in a rapid manner. This will close any decisioning gaps in the traditional underlying systems, essentially enabling a digital twin of the supply chain.
  2. Brexit was the hot topic: LLamacon’s timing interestingly coincided with British Prime Minister Teresa May gaining cabinet approval on Brexit, moving the fight to the parliament. So it is no surprise that Brexit was the hot topic drawing quite a lot of discussion at the event. In a highly provocative presentation, Peter Ward of UK Warehousing Association talked about the challenges ahead for the logistics business in Britain. He said uncertainty around Brexit and devaluation of British Pound is resulting in highest levels of EU citizen exodus from Britain in a decade. He cited that about 35% of the warehouse labor in the UK was of Eastern European descent before the increase in the exodus. With the recent exodus, now the Easter Europeans percentage is down to mid-20’s in percentage terms, of the workforce in the warehouses. This is causing labor shortages and a rise in wages. He also said Britain currently processes about 90 million customs declarations per year. He said the projections are that these numbers will go up to 210 million declarations per year in a hard Brexit scenario. About 4% of these declarations go through inspection, further straining the customers’ systems, resulting in longer lead times and more unpredictability for companies. Several attendees commented that they are ramping up on inventories in preparation for Brexit. This is calling for additional storage resulting in rising real estate costs. This will put a severe strain on supply chains. This also calls for more agility and speed of decisioning to Razat’s point.
  3. Demand Modeling – a practical use-case for AI/ML: LLamasoft’s DemandGuru product created quite a buzz at the conference. Jim Wilson, Sr. Director of Platform solutions made a compelling presentation with numerous customer use cases of demand modeling. Forecast accuracy improvements of double-digit percentages were obtained by using features engineering powered by AI/ML. Additionally, Wilson spoke of causal modeling using external time series data such as GDP, unemployment levels, promotions, CPI, weather, and such. The AI/ML models powering demand modeling are now proven in a variety of industries such as automobile, consumer electronics, CPG, apparel, and metals. The forecast accuracy gains were rather staggering in each of these projects. What I found intriguing was when Jim asked for a quick show of hands on how many people in the room model the demands being fed into their network optimization models, barely a hand or two shot up in a roomful of the audience. My offline conversations with several of the companies in attendance highlighted that most organizations use historical sales as indicative of future with no regard for external causals. Demand modeling can unlock significant value for these organizations in terms of improved fill rates and working capital reduction.
  4. Unique challenges in SCM in Russia: In a highly entertaining presentation filled with fairy tales and caricatures, Dmitry Krasilov of KORUS consulting spoke about the unique challenges in SCM in Russia. It is the largest country in the world in terms of area, occupying more than an eighth of the earth’s inhabited land area and spanning eleven time zones! An interesting network design problem Dmitry spoke about is that of port selection wherein a retailer had to select a port between Baltic sea and Black sea ports. In conjunction with this, the problem of receiving DCs is particularly compounded by the sparse population density of Russia. While the availability of labor is better in large cities such as St. Petersburg or Moscow, the total transportation cost by centralizing the DC in one of these cities could be extremely high. The decision involved trading off the higher cost of labor by being located away from the urban areas and balancing that with the warehousing and transportation spend. All while accounting for special economic zone incentives in certain pockets of the country. A fairly complex problem was addressed very well by LLamasoft’s algorithmic capabilities, beyond what could have been managed in excel spreadsheets. Another example of an SCM challenge caused by vast terrain is a shipment from Moscow to Vladivostok can take as long as 19 days to reach and temperature controlled shipping is a must due to extreme temperatures leading to additional logistical challenges.
  5. Building global design competency: Lee Botham of Schneider Electric made a compelling presentation on their journey of building a global design competency. He talked about their journey to date, lessons learned, and opportunities ahead. He pointed out that when they first embarked on their network optimization journey, they realized that the quality of their master data was extremely low. From there on they built data harmonization capabilities which brought together the data from about 30 different ERP systems (primarily acquired through inorganic growth) and provided a sustainable and repeatable process for onboarding data to enable end-to-end visibility. He also talked about the challenges with analyst turnover and the initiatives they put in place to improve retention. One thing to note is the investments the organization made into associate trainings such as the one on Analytics from MIT’s online learning platform. Even though the analysts are scattered across the globe, Lee mentioned that they have communication mechanisms in place so they operate as one team and also host biannual face-to-face interaction and knowledge sharing sessions for these associates.

That’s a wrap for day 1! It was an action-packed day with lots of energizing discussions and networking. With breakout sessions happening in parallel, my only regret is I had to pick between several compelling presentations. Time to get a digital twin for me!