Mister Beacon Episode #28

Geofencing 201

February 05, 2017

We continue to part two of our master class on geofencing. Patrick Leddy CEO of Pulsate builds on the last episode which explained the “what”, “why” and “when” of geofencing. In this episode he discusses “how”, reprising the chapter he wrote for our book (Beacon Technologies). We talk about the different techniques that are applied, when they should be used and what to look out for when selecting a geofencing solution.


  • Patrick Leddy 00:03

    Geofencing is when the app needs to be open, generally speaking on the device, and people are using GPS to kind of figure out where a user is past geofencing realize, you know, more so on Wi Fi, a lot of people think you have to be a navigation app to run GPS in the background. You don't, there is situations that you can run us in. And we know that you've been extremely grinding their polygons and go right down and let's say, you know, and meters before, it was like a 100 meter circle now for me like a 10 meter triangle. And the way we've done that is through density, using the default geofencing API over the phone, there's major occasions accuracy, finding that and we're friends with the monitor, or if you're looking for a vendor that can really, you know, monitor an unlimited number of locations globally. You want them to be able to do this in the background and possibly when the app was terminated, and you want them to be able to do this with different shapes and sizes.

    Narration 01:03

    You're listening to The Hitchhiker's Guide to the Beacosystem with Steve Statler.

    Steve Statler 01:11

    Welcome to The Hitchhikers Guide to the Beacoystem the podcast for people designing solutions and entrepreneurs who are location aware my name is Steve Statler of Statler consulting. I have the great pleasure to have Patrick Leddy back with us to talk more about geofencing. Patrick, welcome back.

    Patrick Leddy 01:33

    Hi guys. Thanks for tuning back in. So we're back in geofences again today. So I am ready. able for any question that you can throw at me.

    Steve Statler 01:42

    Very, very good. And so. So time has passed is probably a week I kind of now measure things before Trump and after Trump. So this is kind of one way or another week towards the either the beginning of nirvana or the apocalypse depending on where you are in the political spectrum. But if I had a boss, he would be talking in my earpiece and say, don't talk about politics. So we weren't talking about Donald Trump anymore. But apparently it is very good hits on social media. I did write an article for geo marketing once that had Donald Trump as a as a as a hook. And apparently it did very well. But unless you want to say anything about Donald Trump.

    Patrick Leddy 02:22

    I would not I'd prefer to get into end of the video.

    Steve Statler 02:29

    So this this session, last session, we talked about why? Why would you use a geofence. And this week, if you insist, we're going to talk about how and what you need to do. And so one of the things that you talk about in our book, The Hitchhiker's Guide to the ecosystem, is this concept of active versus passive geo fences. What's the difference?

    Patrick Leddy 02:53

    Yeah, great question. I like active versus passive. This is really my own terminology. And there's various solutions that are out there. And people may not even label them this way, or understand what they implemented. But broadly, I divide geofencing into two main groups. Active geofencing is when the app needs to be open, generally speaking on the device, and people are using GPS to kind of figure out where a user is kind of like on the last show, he talked about, you know, regional campaigns and figuring out whether people are in New York stage or whether they're in California, and you probably want to know where people are. So that can be a great way to do it. If you only care about where people last use your app, you can use a form of active geofencing to just run the GPS, one time when the app first launches on every subsequent relaunch of the app, and then shut it down within a second. So every time they open that app, you're just taking a snapshot, this is the current location, and you then do something with that information actually does have some benefits, you know, in that it is GPS based usually, which means you can be very, very granular. And you can figure out where users lost use your app data to incredibly granular locations. GPS can go down to sometimes eight meters of accuracy. And that depends sometimes if you're in a building with lots of steel and concrete, which will attenuate the signal as we know, after geofencing, probably requires less technology to make it work. Because really, your mobile app is just firing up GPS, finding out where it is, and it's telling your server or you're putting down in a database. And then you can use actually MongoDB has some built in functions. To do this. You can do geo queries of people within a catchment area of latitudes and longitudes and figure out, you know, where were they last using my app? One of the drawbacks of course of this is if you want to know where users go when they don't have your app open, as in when they walk into a competitor, or when they come near one of your own stores. You can't rely on customers being helpful enough to have your app open. I'm on the screen all the time, and at the right time. So really, for a lot of people that we work with a lot of companies, active geofencing isn't really an amazing solution, it doesn't really blow anyone away. But the Affer mentioned use cases that I talked about it is good for generally knowing where your customers reside, when you need to take immediate action. And I mean, when users step right into that to you events outside your store, you want to build that attraction and that attention, and then hopefully, to get them to come inside and inspire action, they kind of need to do it in the moment. And in the moment, the the mobile device, it's probably a screen is off, it may be in their pocket, they may have another app open, but it's probably likely that your own app is not open on the screen. And that's why pass the technologies around geofencing are useful. Paths geofencing relies, you know, more so on Wi Fi. And that's why I mentioned in the previous episode, Wi Fi is potentially one of the drawbacks, because it is a prerequisite, you do need to have Wi Fi on for most passive geofencing solutions, where we use the Wi Fi and when you cell towers to figure out usually within about 100 meters of accuracy, maybe a little bit lower, in some occasions where you are, we can do this with the app completely in the background, and then wake up the moment you set foot inside that fence. So once you have Wi Fi on we can do with your app can be closed, the app can be terminated on the device not deleted or uninstalled. But I mean, what I mean by terminate is when you swipe apps out of your app drawer, a lot of people think that that actually saves your battery. But it doesn't, because all apps are put into a hibernated state. But nevertheless, a lot of people do like to kill apps that have their app drawer for good app hygiene or maybe they think they're saving some battery power. So you do need to be able to restart in the background of the tech that geofence. And that's why passing sorry, passive geofencing is more advantageous than active geofencing. But it requires a different a different technical set of skills and a different approach and a different set of challenges as well as doing it in the background. You know, Apple doesn't like a running GPS in the background. That's why you got to bring in the Wi Fi and the cell tower stuff. And of course, even if even it could run GPS in the background on iPhones, you know, your users aren't going to be happy because your battery will be gone in a number of hours if you're not careful.

    Steve Statler 07:21

    So passive is less accurate than active. Is that correct?

    Patrick Leddy 07:25

    Yeah, I mean, it can be had because active uses GPS and uses minimal GPS, usually the last time they opened your app, you can be more accurate around where someone last used your app, but not necessarily where they are. Now while the app is in the background. Passive is kind of like where they are now when the app is in the background. But there is a trade off because the app isn't open. And generally we're not we're relying on less accurate technologies, namely GPS, sorry, Wi Fi and cellular, we pay a little bit of penalty in terms of accuracy, I think the trade off is quite like it's worth, the accuracy is still pretty good, like down to like maybe between 50 and 100 meters. And depending on the situation on the app and the environment that the user is in.

    Steve Statler 08:09

    This can be challenging, isn't it? Because some places is just got loads and loads of Wi Fi access points and others don't. And then there's the kind of like the position of cell towers. And yet is that? Is that uncertainty an issue? Or can we just take it for granted? You're writing an app for light Island, and you've got reasonable level of consistency. But I think about like the states and there's this huge variety, whether you're in New York or in the back of beyond in a less populated state. Is there? Do you just kind of is that an issue that I should worry about? Or or not?

    Patrick Leddy 08:46

    Yeah, I mean, it's definitely something we need to consider. I mean, if you think about you know, urban areas with a, you know, a lot of cell tower sites, and a lot of Wi Fi access points, your resolution is going to be and your your reliability geofencing based off those datasets is going to be it's gonna be very robust and reliable. I was at Burning Man. Back in September, which is in Nevada in the middle of the desert. There's maybe one or two cell towers, not many Wi Fi routers, but you got a geofence in the middle of nowhere. It's not going to work so well. So generally though, what what is helpful around this is the companies that we work with their geofencing urban areas, or geofencing, where they have stores, and we haven't really seen too much of a problem with it. And certainly in the use cases we've explored very few of our customers and prospects are looking to geofence rural or remote areas. So the cell towers kind of get you broadly down where the customer is. The Wi Fi is what zooms in and it gives you that kind of I wouldn't say laser like accuracy, because that's what GPS is what the Wi Fi really shortens it down to a very small area. And the way this works is every iPhone and Android actually works. You may not know this for people watching at home, but it works with a crowdsourcing device or location data. So if I buy a new TP Link router, or Belkin, or whatever it is, I rip open the shrink shrink wrap, and I upload that riser into the wall, within, you know, probably within 30 seconds, that router is now an apple database. Because of my iPhone, which is constantly scanning many Wi Fi networks to join the detective in a new SSID. It hadn't seen before the iPhone in the background irregardless of any app, this is the OS of those days, it fired up GPS, it got a latitude and longitude, high accuracy you want to where this Wi Fi access point was seen in the real world. And then it goes back to an apple database. And every iPhone and Android in the world do this on a credit score settle on all the Wi Fi access points. But why did they do this? The reason they do this is so the apps have pretty decent pretty good location awareness, whatever the battery spends constantly having to fire up GPS. And it works. And it actually works very, very well. That's one of the reasons that we rely on this. And a lot of companies rely on this for geofencing, the Wi Fi data, your phone's already out there, it's doing its thing, phones being a phone, it's connected to a cell tower, so it can get text messages and receive phone calls. It's also constantly sending out a flood of Wi Fi ping requests to say, is there any interesting Wi Fi networks around here I might like to join. So our SDK and anyone who's building their own geofencing technology and the route, you just piggyback on these pings of the phone already doing a thing being a phone. And then you get access to some of this highly accurate location data with minimal battery expense is kind of an interesting implementation.

    Steve Statler 11:39

    So is there any more to be said about this battery life issue? Because it seems like GPS, we have this wonderful constellation of satellites we can use, but we're really draining the battery. Apple doesn't like us using GPS a lot. So we have to think about that when we want to get our app accepted. Is there anything else to be said in terms of battery optimizing for battery.

    Patrick Leddy 12:05

    I think you want to minimize GPS completely wintered an act of fencing and using GPS in the foreground, you kind of just want to do it the first time they open the app. And when you want to do it every time they open the app, you don't want to ruin the GPS radio for the whole duration of that app session. But just for the first second, you know, where are they where was the obsession Bay, that's kind of good enough. When it comes to background. GPS usage, Apple is very careful about how they let you use this and run this and you can actually other people think you have to be a navigation app to run GPS in the background. You don't there is situations that you can run it in. And usually if you're running GPS in the background, and the whole time, it does significantly damage the battery, not permanent damage. But you know, it lowers the battery power on the device. And you know, iOS and Android, you know, come with a setting screen that show you which is the offending app that's causing the battery drain. So if you're doing a lot of battery drain users are going to figure it, there's some app on their phone draining their battery, they're going to look up the summary battery usage and click that X on your app and delete you for the good. And also if you are going to GPS permanently in the background, to do what I would call a very crude version of positive NC There are companies out there that are built this themselves or even solutions on the market offering geofencing, which they painted as extremely accurate. And of course it is are they using GPS the whole time, possibly when the app is in the background, with the battery drains is off the charts ridiculous or unusable by users. And you also have put a warning on your app if you're using GPS in the background. And that warning has to go on the app description of the App Store which says Warning, this app uses background GPS and may excessively draining your battery. Yeah, you actually have to put that on your atmosphere, that's a little bit of a turnoff, you know, doesn't really set the stage, what we do, and we've done you know a bit of a marked change between you know, pulsate geofencing well and and multi geo to dot O is we now allow you to draw extremely granular polygons and go right down and let's say you know, 10 meters before it was like a 100 meter circle, and be like a 10 meter triangle. And the way we've done that is true sensor fusion. And that means that we're looking at a range of different things under the device. So first of all, how someone entered, you know, a region around a smaller fence so we have a small polygon range for us and when people have reached out, so we put a bigger area around that we'd step up our monitoring sooner. As soon as people enter the general catchment area of the smaller fence, then sensor fusion accelerometer data, and other sensors on the device jars. So data being alive, there's lots of things that you can look at, give you clues whether the phone is in motion or not. If the bones connected to a power source, probably not in motion, as well as API's to check these things. This is what I call sensor fusion. And this allows you to decide when and where you want to use GPS. So when you use GPS, you Even versus open data in a certain pattern or algorithm, winning assured either a lack of emotion. So if you're using GPS to arise, be careful with that pattern.

    Steve Statler 15:12

    All right. This is a great one of these things where your head starts to spin, because it's just so there's just so much to this. And we've talked predominantly about iOS just say a few words about the difference between geofencing in iOS and Android.

    Patrick Leddy 15:29

    Yeah, I think you know, the differences. The differences were more apparent, we're seeing that a convergence and the way these geolocation services, and if you look at core location and region monitoring, which are the iOS API's, Android has its equivalents. So we, they are very similar. One thing that I would say, in terms of the differences is very much tied with Android economy. As with everything else, as a developer, you're forever getting ready to do what you like. There's a lot of flexibility and a lot of freedom to launch in some situations where you can kind of just go a bit of mass, and how you implement the solution. And you can really cause a lot of battery drain, you can do all sorts of things. So it's a lot easier to do this Android, but it doesn't mean necessarily any user experience is going to be good. Apple, kind of like in their walled garden, they're protecting that user experience on their sacred device. That's a good thing as a developer, because I guess it those constraints force you to innovate with the solution. And make sure that you're not sacrificing the user experience, a bit more of the Wild West in terms of what you can do.

    Steve Statler 16:33

    Any last comments on just selecting a geofence product, I designed it, I just don't want to learn all the stuff that you described. I just want someone who's going to solve the problems for me, what do I look for?

    Patrick Leddy 16:50

    Sure, you know, if you're using the default geofencing API as a phone, there's major limitations, accuracy, timing, the number of fences that you can monitor for, so you're looking for a vendor that can really you know, monitor an unlimited number of locations globally, you want them to be able to do this in the background, and possibly when the app is terminated. You want them to be able to do this with different shapes and sizes, you know, circular regions and polygons and probably, you know, this age to go down to maybe, you know, probably 10 meters of accuracy, and maybe even a little bit lower. In some situations, you also want to do this, but at the battery drain, you'd probably want to see some 5% bathroom utilization across the day for your app running about geofencing SDK, you want to make sure that the provider is not using GPS in the background the whole time. And you should probably ask them whether their SDK requires the warnings to be put on the App Store description. I bet the background GPS usage. Usually a lot of our competitors answer is yes. Yeah. So I think there's some work questions that are worthwhile asking.

    Steve Statler 17:53

    Very good. Well, Patrick lidding, CEO of pulsate, always an interesting discussion. Thanks so much.

    Patrick Leddy 18:00

    Cheers. Thanks, guys.