Siena Analytics captures sensor data and provides analytics to help customers become more efficient.
In the past, distribution centers have used software that issued static reports every 24 or 48 hours. A problem such as inaccurate labels or faulty packaging could quickly snowball and companies could find themselves way behind in accurately sorting and delivering packages.
Siena Analytics, in contrast, offers dynamic real time alerts about automated sorting issues. “It is really modern software that has come to be in the past few years,” owner John Dwinell says. “The performance and analytics are analyzed continuously. At the same time, we are storing information – depending on the customer’s needs – so they can look back and see how the information has trended. Or if they are interested in something that might have happened yesterday or find a package that went through the building two weeks ago, they can find it.”
Dwinell founded Siena Analytics in late 2013 after having worked as vice president of engineering for a manufacturer of sensors. “I started up a software business there with the idea of capturing all that information,” Dwinell says. “It was very hard to find a home for a software company in a hardware company, so we agreed to split off.”
Initially, Siena Analytics created software that would allow a client to see data across its many buildings. Next, Siena Analytics created an automated monitoring system to support hundreds of tunnels. “Tunnels” refers to an array of mounted sensors on the top, side and bottom of the path that packages and parcels go by. The tunnels are where sensors gather both data and images.
Distribution centers have conveyor belts with sensors. Those sensors capture a lot of information. Until recently, however, that information was mostly being deleted. For example, a conveyor belt moving at 500 to 600 feet per minute in a distribution center would have cameras that could take photos of all six sides of a box passing by. The cameras would capture the bar code, which was used to determine where to send or sort the box to. Then those images were deleted.
“I started the software company originally because I understood we were throwing all this data away,” Dwinell says. “Computers were finally getting to the point where we would be capable of capturing it all and keeping it.”
This was a couple years before the term Big Data came into vogue. “People were realizing it was possible to capture data and analyze it,” Dwinell says. “That’s what we do. We capture sensor data and provide analytics.”
In 2017, Siena Analytics developed Siena Insights, an artificial intelligence platform that analyzes all the information previously being deleted. “It is extracting features beyond the bar code,” Dwinell explains. “You can understand whether it is an apparel bag, golf bag or cardboard box – what type of object it is, the packaging, dimensions and what type of condition. Before that, we couldn’t collect that information at scale.” It allows you to virtually “see” all of your sorting throughput.
Being able to capture and analyze a lot of information at scale is important. “”It is not out of the question that a single conveyor line would move over 10,000 packages per hour,” Dwinell says. “In one building, you can have many lines and move hundreds of thousands of packages per day.”
Among the customers of Siena Analytics are some of the largest parcel and retail companies in the world. Siena Insights software is installed on more than 1000 systems, across well over 100 facilities.
Siena Insights software can spot and alert customers about label information that is missing, incorrect or damaged. The software can also evaluate packaging. That includes the markings on a package, whether the packaging is up to the customer’s standards. “Each retail company, the standards are a little different,” Dwinell explains. “All of them have standards for what they expect of the packages as they come into a building to be able to efficiently store them and ship them.”
For companies without such software, that requires physically looking at packages, a labor-intensive process. Siena Insights, in contrast, will alert the customer if packaging is noncompliant. The images are fully searchable, allowing clients to find a particular package or type of package.
Different departments in a distribution center gain different benefits from Siena Insights. Engineering employees, for example, are experts on materials handling and sensors but are typically new to analytic software. “We show them how to fit supply chain analytics solutions into their entire automation strategy,” Dwinell says.
Operations staff are experts on their inventory but typically don’t have a lot of knowledge about sensors. “When they have a question about inventory, they can see the specific package and better understand their need,” Dwinell says. “They typically ask for a lot of analytics. We create customized dashboards so they can measure many analytics.”
Operations personnel might want to find a box with a particular bar code and then can see the other information about it. Or operations might want to know something about the packaging or labels. “They can compare from one building to another and see across an entire distribution network,” Dwinell says.
Maintenance people are generally responsible for making sure that the sensors are working. “We provide them with simple software to see overall performance,” Dwinell says. “They can look at pictures of packages where bar codes weren’t read. It helps them understand there might be problems with the cameras or mirrors so they know what to repair.”
Siena Analytics offers remote monitoring of their software – Siena Insights – along with the sensor equipment. “That’s what’s runs the building,” Dwinell says. “If that equipment is not working, they are not sorting packages. We can alert the site when there are issues. We can contact the site directly or we can contact the vendor of the equipment if they need to be involved. We can monitor anomalies much better because we have monitoring across all their locations.”
Technology and Supply Chain
Siena Analytics has little competition when it comes to supply chain analytic software. While there are lots of analytics companies, in the supply chain world people aren’t experts on the technology, Dwinell says. And those who are knowledgeable on technology don’t understand supply chain. “There is not really another company we are aware of that focuses on this type of solution in supply chain,” Dwinell says. “There are some hardware vendors that have low-level software that helps the maintenance guys with the health of the hardware. We take a much broader view by linking it up across the entire enterprise.”
Siena Analytics offers a pilot installation program. As of mid-October, the company had done a handful of short-term installations. “I think it is helpful for them to be able to touch and feel it and to have enough data so they get a sense of the value of the software and its flexibility,” Dwinell says.
So far, all the companies that take advantage of the pilot program end up buying the software, he says. “They are then interested in bringing it across their entire distribution network of many buildings. We generally think if a company has five or more distribution centers, this is a good fit.”
There are a handful of companies that manufacture sensors for distribution centers and many of the customers of Siena Analytics use sensors by multiple manufacturers. Siena Insights is vendor agnostic and can work with any of those sensors, Dwinell says. Siena Analytics offers customer support during business hours or 24/7 for 364 days a year (except New Year’s Day).
As more companies consider their digital transformation strategy, those companies need to figure out how information can be made available to all departments. Dwinell explains that in the past a company would have its hardware, sensors and materials handling in one area and operations and IT in another area. “In the past, that structure was fine,” he says. “Today, you need to join the information from the sensor side to your operations side. You are bringing together a business structure that in the past didn’t have to happen. That’s a big change and there are a lot of challenges.”
Dwinell compares it to speaking different languages. IT staff at a client’s company is not comfortable when you bring new IoT software into their world, he says. “We need to speak the language of IT to succeed with any of their customers.”
Likewise, with operations personnel. “Operations has deep knowledge about the information they have but they don’t, typically, have deep knowledge about software and not at all about sensors,” Dwinell says. “You need to be comfortable talking to operations and understand the hurdles they need to address. Sometimes, for example, the volume of data is so large, they need to understand how to handle it. We work with our customers closely, and we want them to be a part of the solution”
Speaking the language of engineering is also important. “We are trying to extend our domain knowledge to speak across those three different dialects,” Dwinell says.
Siena Analytics has been careful to grow at a deliberate rate. “We started with one major customer and built out the platform and picked our next major customer.” Dwinell says. “We are self-funded and wholly owned. We don’t have a venture capitalist to push us to grow 100-fold.”
Siena Analytics first attended the ProMat trade show in Chicago in 2018 and plans to attend MODEX in Atlanta in March (booth 910), and RILA in 2020. “It is a small industry that is tight,” Dwinell says. “We work with some materials handling vendors so they understand we have a product that can provide value to their customers. That has afforded us some very nice introductions along the way.”
End-To-End AI Platform: Learning Insights
Siena’s AI solution is one of a kind, in that it is capable of performing all the steps in the AI image recognition and analysis process namely capture images, label images, train images, and deploy the trained models for production. Currently there are no solutions that cover all of these steps.
Almost 80 percent of time is spent in capturing and labeling the data, and companies often have multiple vendors for each step. Siena provides a holistic AI solution, which ensures consistency across all stages of model development and deployment.
For the future, Siena Analytics plans to keep developing strategic partnerships, attracting talented staff and further developing its AI solutions.