Will Your IoT Project Investment Breakeven?

"Show Me The Money!" the demand from the film 'Jerry Maguire', applies just as well to IoT projects. The first step for your IoT project should be to determine the breakeven point, before getting too deep into the technology.

Here's a summary of the analysis we completed for a client considering whether IoT could lower costs by optimizing when trucks were sent to empty clients' dumpsters. It includes the business justification, breakeven calculations and technical details on how we built the prototype (using Texas Instruments, IBM Blue Mix, Google Analytics and Twillio).

Background

The waste from commercial enterprises such as restaurants varies considerably based upon how their business is doing, but trucks are sent to empty the dumpsters at regular intervals regardless of how full or empty their dumpsters were. This increases transport costs without improving customer service. This project was designed to test the feasibility of using sensors in the dumpsters to better coordinate the waste pickup service and lower costs without impacting the service level. Another requirement was that existing dumpsters be adapted for this project, rather than having to purchase new ones.

Business Analysis

We first researched the operating costs of waste haulage, the revenue generated per pickup and the potential savings from better coordination of the pickups. This analysis was intended to determine the viability of the project, rather than a rigorous calculation of all the costs and savings. The business benefit of IoT projects usually results from either:

1. Reducing costs
2. Increasing revenues from existing customers
3. Incremental revenue sources from new customers or providing new services

In this case, the benefit was determined to be in reducing transportation costs while not materially impacting the level of customer service for waste pickup.
(Publicly available information from the American Transport Research Institute has been used for this blog, rather than client specific details).

The major saving was from better utilization of expensive trucks and labor by reducing the number of pickups that needed to be made. Preliminary research showed that by reducing the number of pickups by one per dumpster per month could yield a breakeven point of under a year. These findings were going to be validated with further field research and interviews with the truck drivers.

Technical Implementation

Having determined that the project was 'potentially' justified from a business perspective, the next step was to evaluate a technical approach to meet the objective and the associated costs. The goal was to test the technical feasibility of the solution, and not to make the final IoT technology selection.

There are several ways to determine if a dumpster is full through sensors, but for this proof-of-concept, we decided to count the number of times that the dumpster lid was opened. The lid was opened when additional waste was placed in it, and this would give us an approximation of how full the dumpster was. A count of 15 was set as the threshold for when the dumpster was ready for emptying. In order to minimize disruptions to the existing operations, sensor readings indicating how full the dumpster was, would be used in an 'exception handling' mode. This meant that truck drivers would maintain their regular pickup routes, but based on alerts, skip those dumpsters which did not need to be emptied.

Our engineers picked the Texas Instruments (TI) SensorTag for its flexibility, for its range of built-in sensors and its connectivity options. The accelerometer in the SensorTag could detect when the lid of the dumpster was opened and relay that count to an iPAD app via Bluetooth which transmitted the count to the IBMBluemix cloud via MQTT. This method of communication, though adequate for the proof-of-concept would obviously be impractical in real life and we used cellular connectivity in our costing analysis. Think of the prototype as a very minimum viable product !

(Note: One of the variables we identified for further research was the cost for a ruggedized sensor that could accurately detect the amount of waste in the dumpster and not be impacted by signal interference from the metal container.)

IBM Bluemix IoT Services was used to store and process the incoming traffic and we set up the Cloudant (noSQL) DB to run a data warehouse (archive backup) on the IBM dashDB service which is a SQL database. Despite many tries, we still had challenges in getting data from DashDB into the IBM Analytics engine and finally decide to use Google Analytics.

From there we set a trigger which would notify the truck driver via an SMS alert when the dumpster was ready for emptying. We chose Twillio for the alerting mechanism given its extensive APIs and its ability to deliver alerts by both SMS and voice calls. Details on integrating with Twillio were described in an earlierblog.

Summary This project took nine days to complete and was split into two steps. First to determine the business value of undertaking the project and second to test a technical approach to test its viability. The results from the proof-of-concept showed that there were enough potential savings to warrant a further in-depth business and technical analysis.

Just as in Jerry Maguire, make sure that your IoT project will deliver the returns you want before commiting! "Some me the money $$$"!

Skilled Analysts is a San Francisco based consulting firm that helps clients grow their business by leveraging IoT technologies. We help with recommendations on how to meet your IoT objectives cost-effectively and also determine what the ROI would be. We build IoT proof-of-concepts (POCs) quickly and at very affordable costs so that you can see how your idea would work in real life, before making a major commitment. These POCs are built using industry leading IoT solutions (such as IBM Bluemix, Texas Instruments and Google) and can be integrated with your current enterprise systems.

Reference :American Transport Research Association report