/
APICS 2018 Chicago Trip Report

APICS 2018 Chicago Trip Report

September 30 - October 2, 2018 in Chicago, Illinois, USA

Top Take-Aways

This list is based on themes discussed at the conference and my assessment of the software solutions and processes/practices represented at APICS.

  1. Machine Learning & Artificial Intelligence (ML & AI): These were hot topics mentioned by most speakers and at most vendor expo booths. That said, I sensed a general agreement that they were talked about a lot but rarely put into practice currently. Presenter Pradipa Karbhari, from Microsoft, asked her audience, "Raise your hand if you are using ML or Blockchain," and in a room of 50 people there was 1 hand raised. It's definitely an area to watch, but it may be some time until OpenLMIS has sufficient volumes of data that would benefit from ML and AI.
  2. Blockchain: Similarly, many presenters and a few vendors talked about Blockchain. My opinion is that Blockchain is more bleeding-edge than ML and AI. It is similar to self-driving cars, where lots of companies are working on it, and there are lots of corporate acquisitions and publicity. That said, blockchain is not widely adopted in supply chains in OECD countries, and even less so for global health supply chains. Chris Richard from Deloitte also surmised that "People talk about blockchain for end-to-end visibility a lot, but blockchain has more potential for smart contracts to speed up payments." It's clear that the benefit of blockchain for supply chains–and applicability to global health supply chains–is an unsolved area that we need to watch.
  3. LMIS? What's that? In conversations at APICS, I was surprised to learn that LMIS--Logistics Management Information System--means nothing in the supply chain world. Nobody I've met knows what an LMIS or e-LMIS is, except the few Gates and USAID folks here. Furthermore, the Requisition process is a requirement that is unique to global health and does not match the processes in supply chains outside of global health. I have not seen any commercial tool that does all of what OpenLMIS does, particularly the Requisition features. On the other hand, leading supply chain systems support re-supply through different means using different terminology that we need to learn about and understand. One example: supply chain tools for the "middle" of the supply chain, from a national warehouse to service delivery point, would use a DRP, Distribution Resource Planning feature set. (See an APICS article about how DRP fits into "push" and "pull" supply chains.)
  4. Offline support? No. The software/solution vendors at the APICS expo do not have any software for offline transactions. Many of the tools are going to the cloud, and all of them require users to be online. Most of the tools are desktop and there are not many solutions here showcasing tablet or mobile capability (other than for drivers). The ERP solutions seem to target business workers sitting in an office at a desktop/laptop computer using dashboards and clunky UIs. When I explain that we need to serve offline health facilities, every salesperson here says we can’t use their product. OpenLMIS's offline support is a unique value proposition which will be needed for as many years as the target market has health clinics and facilities without solid internet.
  5. User-friendly? No. Many attendees complained about the usability of leading ERP solutions from SAP and Oracle. They reported that they were clunky, not user-friendly, and slow. Based on the software demonstrations in the expo hall and the demos shown during presentations, the software systems on display at APICS 2018 are generally very clunky, full of huge grids and tables, and visually overwhelming. Many attendees here say they use Tableau to make charts and graphs look good. OpenLMIS currently has a UI that seems better designed than some of the user interfaces I see here. OpenLMIS users also have unique UI needs--language/translation needs, low-English skills, offline interface needs, and particular workflows in global health supply chains that we optimize for. This is another unique value proposition of OpenLMIS.
  6. Control Towers: At APICS, many people talk about Control Towers–this seems to be the mainstream industry language similar to VAN, Visibility and Analytics Networks. Joseph Roussel noted at the subsequent Gates Supply Chain event that, "There was some confusion about the terms VAN and Control Tower which cover analytics and data visualization on the one hand and demand/supply planning/collaborative planning (E2Open’s sweet spot) on the other hand." The Control Tower model also appears to be a more centralized model where the focus is one central view of everything happening in the supply chain. In global health VANs, it seems that our work is predominantly decentralized, because we need each actor in remote reaches of the supply chain to have a view to take action in their "neighborhood". To meet that need, it may be possible to use off-the-shelf solutions (such as OneNetwork or E2Open) or to apply the new OpenLMIS reporting and analytics stack or other solutions. More research and analysis is needed.
  7. Interoperability & EDI. Are people using electronic data interchange (EDI) of any sort? What standards are leading? How mature is it? How widely adopted? What EDI applies to supply chains in global health? Based on many conversations at APICS, it appears that EDI is considered antiquated and is not widely adopted. One presenter said, "We only use EDI with a supplier when we make 1000+ orders to them a year." EDI takes a lot of custom work to set up each time. Another presenter said, "EDI takes 6 weeks and $50,000 to establish and is surprisingly immature." He suggested we look at OneNetwork for a "new Facebook like model of how you interact with suppliers." Another suggested we look at Coupa Pay. Also, all of the conversations about EDI were for making orders to suppliers electronically. I did not hear about people implementing EDI for inventory reports from warehouses. (Mostly it seems people have an ERP for all of their own warehouses or Distribution Centers, so their ERP already knows what they have in each location they control. And they don't get a lot of automated information about what their 'customers' downstream have in their facilities.) One presenter did suggest we should look at the X12.852 Product Activity Data standard and 867 Product Transfer and Resale Report (see X12 Document List). I did not hear any discussion of GS1 EDI (see the GS1 EDI site also).
  8. GS1. Most people I talked to at APICS had heard of GS1, and understood its importance, especially for barcoding at point of sale. However, whether it is adopted and supported well in MIS systems was not as clear. Standards such as GS1 SSCC (the serial shipping container code) were not mentioned either. Moreover, it sounds like many of the supply chain products that integrate with each other have messy data mapping between systems--similar to my note above on why EDI setup costs $50,000 each time you need it. OpenLMIS may be ahead of the curve in MIS systems for supply chain in terms of having a data model to effectively support GS1. More research and analysis is needed, and obviously there is a drive to implement GS1 in global health supply chains for end-to-end visibility and for preventing sub-standard and falsified products.
  9. APICS has launched ASCM. APICS is launching a new organization called ASCM, the Association for Supply Chain Management. ASCM will focus on organizations while APICS will remain focused on individuals and their professional development. They are also announcing a 3-year Gates Foundation grant to allow ASCM to support public health supply chain strengthening. The team from APICS says there is new hiring for this project, plus collaboration with the Africa Resource Center (including Azuka Okeke), and Glenda Maitlin based in South Africa along with SAPICS (the South Africa APICS chapter), but the details are still getting worked out. The project starts with 3 countries including Kenya.

Presentations

Building the Next-Gen Innovative High-Tech Supply Chain - Microsoft

 Click here to expand notes...

Building the Next-Generation Supply Chain
presenter: Pradipa Karbhari, Microsoft, Seattle
- "Smart Modes, Smart Nodes, and the Smart Web"
- industrial revolution 4.0 = IT Wave 5.0
- smart mode:
-- means bi-modal chain innovation and transformation
-- optimize for rapid response (mode 2) on top for optimizing for efficiency (mode 1)
- smart nodes:
-- a node used to be one whole business entity; now a node can be a single truck (it has intelligence)
-- sense->communicate-->respond
- smart web:
-- how is our network performing?
-- factors: environment, labor, political, social, weather
-- social graph -> commerce graph
-- geocoding, natural language processing (know you customers sentiment!)
-- map-based visualization, track & trace
-- layering order info and supply chain data
-- rapid response by dynamic association & prediction
- ABCs of innovation
-- bots: introducing live agents based on sentiment analysis
-- edge innovation: download the ML models built in the cloud to run offline at the edge
-- final mile visibility (trucks coming and going to optimize; know traffic and weather)
-- final inch visibility (location of every good in the warehouse to optimize pick-pack-ship)
-- 'automatic FAQ maker' you feed in documents and text to ML?
- case studies:
-- IoT, AI, ML for insights and actions
-- smart factories
--- eg, pump vendor is managing pumps in all factories as a service
-- JABIL; mass customization; use MS Azure for predictive analytics;
--- prediction of defects ; also "iFactory" software managing manufacturing end-to-end
-- smart automation
-- smart operations
-- smart distribution
--- JJ Food Distribution; MS Azure and Dynamics; apps for ordering, delivery, tracking trucks

- How many people using ML or Blockchain? 1 hand in a room of 50.

The Sustainable Supply Chain: Leveraging Digitalization and Sustainability to Build a More Competitive Supply Chain

 Click here to expand...

"The Sustainable Supply Chain: Leveraging Digitalization and Sustainability to Build a More Competitive Supply Chain"
presenter: Al Zucco, USG Corporation
- Lean Supply Chain Management
- use process mapping to help move tools and processes nearer to goods
- examples of Lean: 5S, Lean Six Sigma, Kanban, Standard Work, Value-stream mapping, Error-proofing, 5 Why's
- Lean Management System: True North, Standardized Work, Visual Management (reflection boards, clear objectives, performance awareness); People Systems; Leader Standard Work (Gemba Walks)
- Digitalization (video from McKinsey, "future of commercial mobility") --> SEE VIDEO 


-- "cobots" at Amazon warehouses
-- automatic reconfiguration of the network or change of what the vehicle will be used for --
-- predictive shipping
- Examples of digitization:
-- Plan - advanced and automated analytics
-- Source - chatbot data querying; also bots to extract data from silos and put into BI tools; spend analytics; cost driver modeling to know "should cost" during vendor negotiations
-- Make - (manufacturing optimization)
-- Deliver - transportation cost modeling (network optimization); UPS (URION?) system optimizes routes on the fly
- Sustainability
-- gypsum wallboard is infinitely recyclable because you re-grow the crystals every time

Supply Chain 2.0

 Click here to expand...

"Supply Chain 2.0"
presenter: Steven Melnyk (slides are online plus white paper on supply chain cybersecurity)
- Haier applicances: by 2020, 50% of them will be make-to-order and done in the same time and cost as traditional
- old model: customer only talks to marketing who talks to supply chain
- new model: customer now talks to supply chain
- "customer centric supply chain"
- business model conversations: key customer; value proposition; capabilities
- "amazon effect" = 24x7 customer service; convenient;
- the millennials: fast is more important than cost; care about company values - 2 hilarious videos! #WokeMillenial
- social media, IoT, RFID: sensors, IoT,
-- eg, Pirelli, sensors in tires to prevent theft, plus performance/control, monitors usage
-- eg, WeatherTrak HydroPoint data systems
-- eg, Gossiping Cars and "hum" from Ford? - Waze on steroids
-- eg, IoT allows you to short-circuit the bullwhip effect by shortening frequency of data
- changes in Supply Chain management: Measures, People, Moments of Truth
-- measures: use your customers measure as your own; but also primarily for communication _not control_; including communication of what is important
-- moment of truth: if you expect something and it does not take place
- Experiential Supply Chain
-- personalized supply chains; "a market of one" - matches with lean concept of "a unit of one"
-- example: Starbucks Reserve video
...stopped taking notes about CyberSecurity...

Digital Industrial Transformation - How GE Enhances Operations with Cutting-Edge Technologies

 Click here to expand...

"Digital Industrial Transformation - How GE Enhances Operations with Cutting-Edge Technologies"
presenter: Gerald Jackson, GE Digital plus two colleagues including Dr. Young
- AssetAnswers tool that GE developed and now offers externally
- "bad actor list"
-- sort by opportunity cost
-- compare equipment reliability (eg, "how are all my 45,000 pumps across my factories performing compared to industry average?")
--- use this to optimize my maintenance, or to choose to replace my pump vendor
- the "digital thread" - digital thread for product design and production
-- when you can take information in a digital fashion from design of the product, through manufacturing, to customer, that is an end-to-end digital thread
- Machine Learning
-- challenges and solutions
--- Bayesian + RF=Random Forest
- GE is testing the following:
-- drones for inspection
-- augmented reality for the shop floor
- using Machine Learning/AI for data cleaning
-- reduce the master data management team (which is a huge team)

They use machine learning for cleaning data. "Clean Metal” is one company (I could not find a company by that name). His advice is get a good Machine Learning person into your team to start doing the work–there is no turn key solution today.

Digitizing the Supply Chain - PwC and Boeing


4th Industrial Revolution and the Supply Chain - UI Labs

  • Kevin McDunn provided great insights on UI Labs' model for bringing multiple partners together to accelerate R&D. He suggested the following resources:
    • A book called Inspired by Marty Cagan. 
    • A book called Jobs Methodology by Clayton Christianson. Do not focus your product or solution on the pain points. Nobody buys pain points. Instead, focus on each 'job' that needs to be done.
    • An agency called Strategyn publishes a lot of information about this.
    • Dual Track Scrum, a process where scrum is used for innovation, but also used in a parallel track for managing the portfolio of projects. Kevin's team has a daily standup and conforms to a cadence where a management point person joins each projects' scrum meeting and comes back to the management track to help advance that project. (It's a different model than scrum-of-scrums for coordinating between scrum teams). Divide and conquer and go to their sprint meetings. They also have "agile tech team calls" that APICS members may be able to join based on APICS' membership in UI Labs DMDII.
  • Katie will send contact info for Ashish Banjeree (spelling?) at UW. Maybe his tool can be applied or VillageReach can pick his brain about what is possible in our supply chain work. UI Labs is hosting a research project on Supply Chain Visibility with this UW researcher.
  • Brandon Bowersox-Johnson check out these resources and send Kevin a thank-you note.
 Click here to expand...

"4th industrial revolution and the supply chain"
presenters: Kevin McDunn, UI Labs CTO & Katie Tillery-Merk, Director of technology strategy & programs, DMDII (Digital Manufacturing and Design Innovation Institute)
- 3 industrial revolutions exploited mechanical leverage (Steam, Electricity, Computers)
- 4th revolution ushers in cognitive leverage (Cyber-Physical augmenting of human labor)
- role of digital tech:
-- digital thread
-- IoT & Mechatronics
-- AR/VR
-- AI + Blockchain + Cloud
- Continuous Improvement (from Toyota system) can now be shrunk to "accelerated learning"
-- vision: "every part better than the last"
-- real-time improvements to manufacturing
- mass customization
-- shift from commodities to goods to services to experiences to transformations (Pine & Gilmore, 2011)
- Examples:
-- Shao Mi cell phone company in China puts out a new cell phone variation every single week (SW + HW)
--- they put out small batches and respond quickly to customer requests
-- Zara
--- quickly respond to fashion trends and customer requests - cycle time is down to 3 weeks from a new design to in the stores
--- RFID tracking so they know very quickly what items move off the rack, use AI to plan store replenishment mix and to iterate on designs --> find and share a video or link about this! Very cool!
-- EGP (engineered glass products)
--- went from 1 hour change-over to 10 minutes; gave all employees a bonus check if they could do 15% volume growth
- Katie, DMDII, shares example projects:
-- real-time welding feedback: uses computer vision and ML to feed back into control system
-- Caterpillar integrated part variation management
-- Lockheed Martin smart printed circuit board digital factory
-- UW increased supply chain visibility through predictive analytics (Walter Young?)
--- analyzed and cleaned data, trained ML algorithms,
---- EG, what are chances you will receive parts on time? results: 40% improvement for 90% of POs
- "commercial readiness" scale goes from 1 to 9; they work on 1 to 7
- how they work:
-- monthly agile cadence to look at projects, kill off projects that have market solutions, redirect projects;
-- they spend a lot of time "grooming" the problem statement and that is really helpful to participants

From Linear Supply Chains to Digital Supply Networks - Deloitte

 Click here to expand...

"From Linear Supply Chains to Digital Supply Networks"
presenter: Chris Richard, Deloitte (knows something about EDI)
- AI: "There are a lot of acquisitions, but there are fewer applications [meaning uses]."
- Deloitte is moving from an "augmented SCOR model" where SCOR has linear steps >> over to an integrated network model