Do cheap GPS trackers work? A review of the GF-07, GF-09 and GF-22.
I wanted an easy way to track my bike. This article lays out what I’d like a tracker to do, and my results from testing 3 cheap, widely available trackers: the GF-07, GF-09, and GF-22.
TLDR: No, these trackers either don’t work at all or are very inaccurate (e.g. >250m off).
A brief intro into tracking
My bike got stolen recently. Someone broke into my flat’s bike store overnight, and took it plus some other bike equipment. The store is covered by two CCTV cameras, but the police didn’t want to pursue it, and the company owning the cameras won’t release the footage - so I’m at a dead end.
This got me thinking: could a GPS tracker help me recover my next bike? Both directly, and by pushing the evidence available for police over the edge to where they’ll investigate.
I think this is probably worth 8-18% of the value of the bike. 93% of London bike thefts reported to the police are never solved, giving a base rate of about 7%. I couldn’t find good statistics on recovery of bikes with GPS tracking, so my finger in the air estimate is that a tracker raises the chance of recovery from 7% to maybe 15-25%. I’m going to ignore the recurring benefit (a recovered bike can be recovered again) for simplicity.
Given I value my bike at about £400, this makes it reasonable to spend ~£30-70 on a tracker - including the cost of my time purchasing it, fitting it to the bike, charging it, etc.
Some products cost quite a bit more. For example, the Invoxia GPS Tracker Pro costs £150, and the location has to be accessed via a proprietary app with a £8 monthly subscription.
But there are a few trackers within budget. These seem to be sold under a range of brand names with a 'GF' model number. I bought one of each: a GF-07 from Amazon for £6.99, a GF-09 from Amazon for £12.99, and a GF-22 from eBay for £16.99.
They are at least small and light - the GF-22 fits easily in a palm.
Also in this price range are tags that work on the Apple Find My network, such as Airtags. These don’t use GPS: they ping out an identifier via Bluetooth, which other Apple devices send back to a central database with the location where they were spotted. Unfortunately, they require pairing with an iPhone, which I don’t have. In future there might also be a Google Find My Device network, but this is apparently blocked on Apple implementing some standards for detecting unwanted trackers. I don’t explore either of these here, but have bought some Apple Find My compatible tags for a future blog.
My requirements
So, what makes a good bike tracker? I want one that is:
- Accurate and reliable. If my bike gets stolen, I want to track it to a particular building in a city (multi-storey blocks complicate things though). That means an indoors accuracy of about 5-10 metres, while the tracker is inside my bike. It should work reliably, ideally without proprietary communication protocols or apps.
- Small and lightweight. I want to attach this to my bike easily, somewhere not so obvious that it will be quickly removed. It also shouldn’t add much weight.
- Long-lasting, at least in standby mode. I don’t want to be charging this all the time. But I don’t need tracking unless the bike is stolen, so it just needs to sip battery while waiting for a text. A good target: two months of standby, and maybe 24 hours of GPS updates (one every 5 minutes or so). It would be clever if it texted me when it was low on battery.
- Inexpensive. For the reasons above, £20-£60 is about the right range. I’d also want to avoid an expensive monthly subscription - although a couple pounds a month would be reasonable.
How reasonable are these requirements?
At this price point, and based on the product pages, I was expecting basic cellular devices with a GNSS chip (such as GPS, GLONASS, BeiDou or Galileo). The tracker acts like a mobile phone: when it receives a SMS text message, it responds with its coordinates.
Fully-fledged phones with GNSS and SMS exist not far from the target price point:
- For £30 you can get a TTfone TT240: GPS, GLONASS and SMS, plus WiFi, Bluetooth, cameras, a screen and a Linux-based operating system for apps. 11 days of standby.
- For £40 you can get an Alcatel 1: GPS and SMS, plus WiFi, Bluetooth, cameras, a 5” touchscreen, and Android apps. 13 days of standby.
A longer-lasting battery also seems possible. For £30, you can get a Nokia 3310 with a 30 days standby time, or for £75 the Nokia 2 with a 56 days standby time. One or two 18650 cells would fit neatly into a bike seat post, and would likely give the device plenty of battery.
With just the core GNSS and SMS functionality, it should be possible to build a compact yet long-lasting tracker for about £30 that meets all the requirements.
The GF-07, GF-09 and GF-22 all advertise roughly this, with GPS. However, they all skimp on the battery life, and we’ll see about GPS in the next section…
Comparing tracker performance
I tested all the trackers on 2024-03-24 at 51.52565, -0.089443 (near Old Street station, in London).
GF-07
Location tracking did not work at all. Instead of coordinates, it just returned a Chinese message: ‘【未获取到地址链接】’ - ‘Address link not obtained’, probably meaning it struggled to find a signal to determine its location. I gave it another 5 minutes of searching and tried again, but got the same error.
The device also advertised audio monitoring: you can call it and hear what’s going on around it. This did work, but the microphone quality was poor and only picked up noises very close by. I reckon this would be useless across the room from somebody, and unlikely to help retrieve my bike.
GF-09
The exact same location problem as the GF-07. It again returned ‘【未获取到地址链接】’ (including after waiting 5 minutes). Given the same message and formatting, I suspect the firmware behind these things is all roughly the same.
It also advertised audio monitoring, but this didn’t work: the device simply declined all incoming calls.
GF-22
It wasn’t looking hopeful after the last two, but this time I got a location! The device responded:
BAT:93%,http://fcstgps.com/smap.php?lac=316&cellid=31743&c=234&n=15V=7100
The link took me to Google Maps displaying the location 51.528194, -0.089806 - about 280 metres off. I tried a few times, waiting a couple of minutes between each attempt, and got back the exact same URL pointing to the same location.
The red pin is where the GF-22 reported it was; the blue dot is where I actually stood.
I tried once more at a different location. This was even worse: over 400m away.
I’m not certain what’s happening here. I think it might be finding the GSM cell tower location rather than actually using a GPS chip. In any case, this would be useless for finding my bike in a crowded urban environment like London. At a stretch, it could maybe help track a larger object like a car or van in a less dense environment - but even then I suspect there are better options, given weight and size become less of an issue.
Like the GF-09 it advertised an audio feature, which didn’t seem to work - again declining all incoming calls.
Conclusion
These all clearly fail at requirement 1 (accurate and reliable tracking). I was somewhat expecting the GF-07 to perform poorly, given its manual claimed to only use LBS positioning (looking at nearby GSM cell-towers, generally only accurate to the nearest few hundred meters). But I thought the GF-09 or GF-22 would provide reasonable accuracy - their packaging does claim to have A-GPS.
These devices do perform well at requirements 2 (small and lightweight) and 4 (inexpensive). But a small, inexpensive device that doesn't work still isn't very helpful.
On requirement 3 (long-lasting), they generally only advertise 1-2 days of battery life, so likely fail here. I didn't bother testing this given the devices didn't work well in the first place.
How are these still on the market?
It's surprising that these products, which clearly don't work as advertised, are still sold so widely on Amazon and eBay. The situation seems similar to sellers listing dangerous USB-C cables or fake flash drives and SSDs: it’s simply cheaper for marketplaces to refund customers than properly monitor these product categories (or hope that the refund process is annoying enough that customers give up).
Marketplaces could still do more here: the default is that these products do not work, and they keep using the same model numbers - so clamping down on these in particular (or at least how they’re advertised) seems relatively low lift.
Finally, I am confused by the positive reviews on some of these products. Maybe these are fake reviews, or people who are okay with 400m accuracy? If you bought any of these products and had a good experience that differs from the above, get in touch.