Setting Up An Inventory System Is As Easy As 1-2-3…4

Written BY

Emily Friedman

May 1, 2018

Written by Special Guest Bloggers Robert Seward and Steven Lewis, Co-Founders at Rendered Perception

Computer Vision (CV) and Augmented Reality (AR) coupled with Artificial Intelligence (AI) will create that step change for inventory improvement that you are looking for. A good inventory system is a byproduct of having a deep understanding of your customer. Through our years of experience, that means you must be relentless about delivering a great customer experience. If you truly understand your customers’ jobs to be done, you will innovate and hire the correct product. More importantly, it is about progress not just a static product or service. There are plenty of off-the-shelf inventory management platforms for purchase but creating the best-in-class experience cannot be pulled from the shelf. When building an end-to-end inventory system, it should be set up in a way to collect insights, learn, teach, predict and understand customer circumstances. We will describe how setting up an inventory system is as easy as 1-2-3…4.

Step 1. Pain points: Understanding Customer Friction

First, you must dig into all the pain points, rationalizing customer friction points. In creating any solution, we first fall in love with the problem. We employ a 5-D method: Discovery, Define, Design, Develop and then Deliver. Whether you have been in the distribution business for years or are looking to increase your value proposition by adding warehousing to your transportation outfit, pain points exist. Inventory is a function of storage and flow.

Pain points on the flow could be as simple as needing better coordination with your vendors on the receiving side.  Implementing something as simple as advanced ship notice (ASN) provides visibility to the transported inventory. Couple ASN with the product type and engineering standards provides earned hours for scheduling within a workforce management system. It is common to use inventory buffers against the variability associated with customer demand. It is critical to know what you have, where you have it, where you are going to need it and how you will provide it.  Yesterday, you needed inventory correct at your edge node (local unit). Today, you need inventory visibility across the enterprise in real-time to accommodate the different purchase channels as well as provide vendors visibility to collaborate. Even outside the normal business, when a weather event like a hurricane happens, it is critical for inventory precision. Having the ability to create a pop-up retail unit, dynamically shifting inventory to the nodes that matter, is a competitive advantage and delivers tremendous value to customers that need it the most.

Step 2. People Focus: Simplify Tasks and Activities

Inventory systems can be cumbersome, frustrating and complex.  We have known operators within the business that have a successful track record garner more influence on how the inventory flow and processes should work. The challenge can be having a holistic viewpoint of the pain points and job to be done. The operator has tremendous domain knowledge of the business and expects everyone to have that level of experience and execution. Reality is the system should be designed to the lowest common denominator. You cannot assume much of the workforce will be able to execute a system designed by and for an expert operator. The balance is how do you leverage the person with operational expertise with professionals in the inventory space and a sprinkling of tasteful automation.

If you get this wrong, you spend good money and time, yet employees end up fighting with the systems and the inventory is not correct. The natural inclination is to automate everything and hope the problem will go away.  There is no shortage of use cases that people can speak to that would make their professional lives easier. The challenge with that is automation requires very complex calculations, multiple streams of data and backend processes. You do not want to automate bad behavior. The cost to automate something that changes frequently is a waste of capital. Before automation, you should have very solid controls as a foundation when creating business requirements. People that design the processes have great intentions but cannot aggregate the complexities thus creating a mess. There are several case studies that illustrate the importance of identifying the right systems to automate versus enhanced workforce. There is a sweet spot on the automation curve that leverages labor expenses intelligently versus spending capital on automation.

Fun Math

If your inventory system is only 60% accurate, what is the math of your secondary systems and how accurate are they? If they are not perfect, you start to talk about fractions of fractions and your system collapses. Where you want to start your calculus is with a near 100% for your foundation (which is Inventory by the way people!), and then your fractions can start from there, preferably 99% or 100%’s all the way down so that it runs smooth. If your foundation is secure and running great, your secondary systems will take care of themselves so that you can focus your support and attention to more important things… like the customer.

Example: You order an item online and you don’t get it. Does the problem end there? The company has a 95% ship rate. If the customer did not get it, where is it? Where was the real-time alert identifying a break in the supply chain? Proactive versus reactive. Find and fix the problem before the customer realizes anything less than superior service has occurred. If there is a problem in the supply chain that cannot be addressed in a timely manner, the customer should be updated and informed before unpleasantly surprised.

Step 3. Process Focus: Standardize and Streamline Routines

You have heard the expression, “what gets measured, gets done.” Companies understand the value of simplifying, standardizing and optimizing processes. Creating routines and standard operating procedures (SOPs) aligns large-scale labor forces. The challenge is not in the set-up of engineering standards, working data sheets and frequency studies. The challenge is in the delivery of the training material!  No one appreciates the series of 4-inch binders containing outdated instructions on how to perform a task. Maintaining the binder content has evolved to basic interactive training videos. Would it not be easier to use Augmented Reality (AR) to do the training while the employee is performing the task? We have seen training that normally takes several weeks down to a few hours.

Once you have an AR-assisted solution available to help employees complete the task, you need to have a follow-up mechanism. Yesterday and today, you would have a small team of auditors or managers audit a sample of tasks. Does that audit team need to exist tomorrow? Could you build AR tools performing system-assisted inspection? You still need to inspect what you expect. Instead of auditing a person, you would be validating the results… a modern version of trust but verify.

Building an AR-assisted solution will not happen overnight. Good news, though, is you get notable incremental benefits along the way. Most approaches today start with taking existing systems to mobile. Then from mobile to head-mounted displays. Lastly, head-mounted displays to basic AR. We believe in starting with basic AR and rapidly iterating to more value-added AR. A 3-year roadmap could look like the following:

Step 4. Platform Build: Innovation, Automation and Analytics

Building the platform is the fun part. Integration is simply a function of inputs, outputs and transformations. Most people see and judge a solution on the merit of interface. The secret is not in the interface, it is in the data capture. Identifying the source(s) of data, building real-time systems to ingest the data and build a system to intelligently understand and then apply the data are some of the most important parts. This is not sexy but pays tremendous dividends. Please note we did not get this right the very first attempt. What kept us on track is we had a motto for when we reviewed our “final” solution design– Hate your design, continue forward and iterate tomorrow. We did a 3-month proof of concept that we could have easily spent a year on, but we would not have gotten through all our test-fail-learn cycles.

The diagram listed below is an oversimplification to the actual architecture design. A few notable jobs to be done based on our experience:

  • Capture lost sales opportunities – what, when, where, why, how
  • Workforce planning – based on routines, SOPs, engineering standards and dynamic tasks
  • Connectivity throughout the supply chain – anchored in the retail unit and worked upstream and downstream
  • Predictive insights – decision options, consequences, pros, cons

Streamlining everyday tasks, performing wildly complex computations, and having a personal assistant to talk AND walk you through exceptions should be a staple. There is a lot of work that goes into building out the technology stack, software configuration and use case prioritization.


The business should be made as simple as possible. We have built algorithms to create calculations to redesign direct labor out of the system as well as add capacity and increase accuracy. As part of the journey, we built backend processes to remove non-value-added time associated with set-up and wayfinding. In the end, we have always maintained the customer vantage point.

Inventory management powered by CV, AR and IoT creates an intelligent inventory solution. AR technology is here and unlocks a wealth of value added opportunity. If you truly understand your customers’ job to be done, you will innovate and hire the correct product. We fall in love with the problem. If you strive for the best-in-class customer experience, building an inventory system really is as simple as 1-2-3…4.

If you have additional questions, feel free to reach out to us on LinkedIn:

If you happen to be attending the following AR conferences, stop by and chat:

Further Reading