Three Technology Dimensions for Reliable Physical Retail Customer Data

Deciding on the right customer analytics solution for physical retail or malls is challenging these days. Such solutions are provided by a variety of technology companies that offer substantial differences in features. Furthermore, not all technology providers deliver the same quality and spotless device maintenance and thus the precision of service and reliability of data. How is this reliability achieved at VIVIDI?

The effectiveness of VIVIDI resides in its own locally manufactured hardware, and thus short supply chain, topped with in-house software development. All in all, this leads to cost-efficiency, quick turnaround on support, and high precision, which so far has been the biggest concern of most of our clients. Let’s see what lies behind AI-powered VIVIDI sensors.

1. Czech-made hardware

VIVIDI is among the very few companies that do not outsource the hardware. We design our own sensors and continuously develop them, currently working on Generation 3 prototype. Sensors are manufactured with the help of local providers such as Jablotron and assembled and programmed in-house. Consequently, quick delivery and much faster access to support for time-sensitive issues are secured in comparison to other companies, given that support for technology manufactured in one place and programmed in another is more complex.

Each VIVIDI sensor undergoes deep scrutiny with the automated pre-programmed tests, to ensure their 100% functionality. Before it goes live, we test, as an example, if it recognizes people well, if the device can disconnect and connect again, or can see the setup scene well.

2. In-house monitoring platform

When the sensor is live and operating, each is then monitored by VIVIDI’s proprietary BOX MANAGER (VBM), which makes sure all sensors function properly.

Once each new sensor has its ID number entered into the VIVIDI monitoring system, it consistently and automatically communicates with VBM. If there is a connection issue, after every 60 minutes that the sensor is offline, the error ticket is sent. VBM makes sure the sensor reconnects and creates the log of errors which we analyze internally. If the connection is not enabled after 24 hours, VBM restarts the sensor. The data history is never lost even when the sensor is offline as it has its internal storage.

Among the most typical errors we see is that the sensor is without electric power or an internet connection. The biggest advantage of the VIVIDI sensor is that it transmits the data using LTE. During the initial development stages we found Wi-Fi to be unreliable, and POE, used by many, to be difficult to install. Hence, we built in LTE connectivity and created a plug-and-play installation. The LTE connection is again constantly monitored. Wi-Fi connection works here as a supporting alternative.

Currently, Generation 3 VIVIDI is being developed with an even stronger antenna and a more sensitive sensor for imaging.

3. Continuous data consistency monitoring

While visitors’ demographics are traced by VIVIDI proprietary AI Software, the passerby borders, which define bounce rates and conversions are set by specialized team members based on best practice scenarios and experience gained over time. Then we continuously examine our data to spot inconsistencies. When we see unusual numbers, for instance, 70% bounce rate at a particular store entry, we deep dive to resolve the problem. Store designs, particularly at the entry ways are dynamic and are often changed for seasonal needs, which also alters the scene the sensor is watching. Keeping the settings up to date and KPIs consistent is a part of the VIVIDI service. When new borders are digitally optimized we continue watching data to ensure its correctness.

VIVIDI’s smart sensors are set up to watch particular scenes where visitors are counted and analyzed. When various sensor shifts occur, for example, due to store maintenance, and scenes stop being visible, sensor shift detection issues an error message. If the shift is minor, we readjust the scene, redraw the borderlines/zones and recalculate the data. When the shift is large, we contact the client to organize immediate repair.

Conclusion

Many companies deliver somewhat similar sensors and services, however, the level of reliability and precision differs dramatically. VIVIDI sensors are highly durable, they work inside and outside in rain and dust. They also take little energy and store the information in case of internet outages. Unlike most sensors on the market, they are not dependent on a Wi-Fi type of connection as they connect via LTE. No need to say it is a win-win situation for any physical retail operator looking to get high-quality data on its visitors.