NEW DATA SOURCES DRIVE INVESTMENT OPPORTUNITIES IN MANUFACTURING AND SUPPLY CHAIN

Synapse Partners is actively looking for investments in the Manufacturing and Supply Chain and Logistics sectors because we believe it is an area where the combination of rich new sources of sensor data and a maturation of AI technologies will drive tremendous efficiency improvements. Key to realizing these opportunities will be the introduction of new Insightful Applications – software applications that translate complex data sets into actionable business insights. This blog post will be the first in a series of posts that will dig into some of the more interesting trends as well as profile specific companies we think are innovators and leaders in the space.

Why We Believe This Market Is Ripe For Disruption
We view the Manufacturing and Supply Chain and Logistics markets as prone for disruption because they exhibit several key criteria:
1. Complicated problem to solve – optimizing the raw materials, manufacturing, labor, transportation and supply of goods across a disparate set of retail options is an incredibly complex task where models look across a dizzying array of variables.
2. Complexity is increasing – consumer requests are making the supply chain more complex. Omni-channel commerce requires retailers to coordinate inventory across online orders and in-store returns and “clicks to bricks” strategies from historically online-only retailers are also on the rise. Retailers are demanding shorter response times from manufacturers as Amazon pushes the envelope with “Anticipatory Shipping.” In a recent survey of supply chain executives, only 31% considered their capabilities as “excellent. Grocery stores and restaurants are dealing with consumer requests for locally sourced meat and produce that require smaller more nimble supplier relationships and ingredient transparency that also place demands on the supply chain. Those that question the difficulty need look no farther than Amazon’s decision to purchase third party supply chain logistics software.
3. Increasing amount of data – Data from connected devices on the manufacturing floor and through the supply chain is reaching a critical mass. Manufacturing connectivity has been driven by continued slow progress from equipment upgrades and swapping out control systems as well as improvements in retrofit systems. The supply chain has been phasing in technologies like connected pallets, autonomous mobile robots (AMRs) in the warehouse as well as regulatory tailwinds like the ELD governance mandate which mandates digital logs (essentially phones) in commercial vehicles. This will only increase as we get closer to autonomous commercial fleets.
4. Enormous potential for cost savings – A study by DHL and Cisco in 2015 estimated that implementing IoT technologies would generate $2.1 trillion and $1.9 trillion in productivity improvements in asset utilization and supply chain and logistics respectively over the next decade. There are very few areas where technology can drive this kind of ROI.

Making Data Valuable
Companies will find it difficult to extract value from sensor data with legacy teams / BI tools. Collecting sensor data from the supply chain will be expensive (both transmission and storage) and will almost certainly be too overwhelming in volume for any manual analysis. Consider an example outlined by Teraki, an IoT data management company: A sensor in a manufacturing plant that uses ultrasound waves to inspect components generates 3GB of data every minute.Assuming the plant runs 24/7 (many do), that is 180GB every hour 4,320GB every day. Now multiply that by the number of sensors on the manufacturing floor and you can start to see the scope of the data volume problem. When you consider that the data is most valuable when synthesized with data from other sensors as well as the MRP/ERP system you can see why the problem is an exponential one.

Analytics and visualization tools that businesses have historically used to tackle data problems are not capable of managing the scope of data nor providing the predictive analysis that businesses are demanding. Analytics and visualization tools have a fundamental flaw in that they rely on humans to extract insights and make recommendations. As the size of the data sets and the number of dimensions expands, the ability of humans (even with visualization tools like Tableau) is fundamentally limited. The other limitation of analytics and visualization tools is that they are focused on descriptive (what happened) and diagnostic (why did it happen) as opposed to prescriptive (what can we do to improve things going forward) . That ability to offer business insights (optimizing the business through prescriptive actions) is beyond the scope of these applications.

Insightful Applications can apply machine intelligence to large and complex data sets, understanding business context and making or even implementing operational recommendations. My colleague Evangelos Simoudis has written extensive about this new class of Insightful Applications in a three part series on the O’Reilly website. In the article Evangelos lays out the five key criteria for Insightful Applications: flexible data management, rich knowledge representation, reasoning capabilities, machine learning and the ability to synthesize actions to drive insights. We are seeing a new generation of enterprise software where the ability to deliver business insights will be the key purchase criteria (as opposed to 1.0 “systems of record” or 2.0 “systems of engagement”).

We believe that this new generation will be far more important and far more significant than previous generations because it represents an enormous step function in value to the business unit (and therefore will have the ability to command higher revenues per customer). Consider for example Celect, an inventory management software company focused on retailers. Historically the majority of their customers used large ERP solutions to keep track of their inventories. These solutions represented a significant step forward in efficiency versus paper based solutions, however they offered no insights on how inventory should be optimized based on business trends. For that retailers turned to spreadsheets or perhaps visualization tools that allowed them to begin to see how overall sales might impact demand for specific items. Celect’s solution, which is based on machine learning technology out of MIT, now provides retailers with the ability to look across inventory data, CRM systems, POS data, and web traffic in order to provide recommendations for inventory optimization at individual stores. The result is top retailers showing revenue gains of 5% – 10% and decreasing inventory writedowns by as much as 25%. From a venture capitalist’s perspective, the pricing power of a software application that can deliver this kind of value to customers is mouthwatering.

The proliferation of new data sources and the potential for Insightful Applications to deliver tremendous value to customers is what gets us excited about the Manufacturing and Supply Chain & Logistics sectors. We look forward to continuing our discussion with startups, corporations and other VCs in the space and sharing our insights in future posts. Over the next several weeks we will be covering topics such as how new data sources are creating opportunities in trade finance, how robots are changing the economics of e-commerce and how the rise in AI are creating entirely new value chains (to name a few). Stay tuned!