I am an event planner and a company is trying to sell me a plug and play sensor and analytics platform that can do things like estimate attendance, show crowd flow trends, average dwell time, etc based on passive MAC address scanning. Is it bs?

I am an event planner and my company puts on a fairly large St. Patricks day parade every year. We struggle getting good data and numbers around attendance, popular areas along the route, average dwell time in areas, and basically any consumer behavior trends that can sometimes be gleaned from access point connection data. This company approached us about accomplishing those things via a sensor that passively scans for MAC addresses and sends the address, signal strength, and time stamp to their back end system for processing and analytics. While they are not explicitly claiming they can tell us exactly how many people showed up they are promising "data backed estimates," along with heat map analysis of peak times and dips, average dwell time spent in range, and if a MAC was seen at one or more other sensor locations. The sensors and service are fairly low cost when compared to other people counting tech like facial recognition so I'm not worry about getting completely ripped off. My concern comes from conflicting reports trying to research this type of data collection. With MAC randomization it seems like it could be bullshit, but at the same time the randomization doesn't seem to completely expel their claims around showing peaks/dips, but some other things like dwell time and route trends seem to be more difficult to accomplish. However there does seem to be a lot of doubt around how effective the randomization even is on IOS/Android, and even claims that Android manufacturers rarely implement MAC randomization in the first place. So tell me is it bullshit or does these guys seem to have a sound product?