Advanced Subscriber Segmentation and Targeting

Level up your use of OneSignal's segmentation by making use of advanced subscriber segmentation and targeting practices.

OneSignal's Customer Success Manager Eyra Dzakuma goes through how you can level up your use of OneSignal's segmentation to make use of advanced subscriber segmentation and targeting practices! The transcript of the webinar can be found below.



Okay, just to keep things on time, we'll go ahead and get started.

Thank you all for joining me today for the OneSignal webinar on Advanced Segmentation.

My name is Eyra and I'm a Customer Success Manager here at OneSignal. I work closely with our clients, helping them to get the best out of the system and figure out what they can do with OneSignal to help them achieve their goals.

For our agenda today, we're going to do a quick overview on segments and filtering, and then we'll actually get into some more advanced segmentation. And then finally, we'll take everything we learned and actually walk through the platform to make it a little bit more real-time so you can see how to set up these advanced segments. And then we'll also leave a little bit of time for some questions and answers at the end.

Segments Overview

So getting to the segments overview, the first segments are essentially just a group of users that you are targeting based on the data that OneSignal collects and any data that you send to OneSignal as well. These groups or segments allow you to target devices with more personalized and engaging messages to hopefully have a higher click-through rate. Essentially, you can divide your subscriber base into specific audiences or groups based on multiple filters that we have within the OneSignal system. These filters are based on data that OneSignal automatically collects via the SDK, as well as any custom data that you send to one side.

One thing I wanted to make clear is that segments are dynamic. Segments are updated automatically once they are created, which is great. Once you create a group, anyone that meets the criteria of that group will drop into your segment. As users interact, they may come in and out of segments without any additional work or set up needed from your end.

I also wanted to highlight that segments can be created through the OneSignal dashboard or via the API as well. Today, we're going to be focused on creating the segments via the dashboard, but I do want to highlight that you can create segments, specifically the API. We have a lot of great documentation about that as well.

So getting to filter types, essentially, the filter types are shown in the dashboard. This is basically how you can group and create specific segments. You can combine lots of filters and do this by using an AND clause or an OR clause so you can get as specific as you'd like. As I've highlighted here, essentially, AND clauses mean the device must match all the filters to be included in the segment. The OR clause means a device can match either case of the filters to be included in the segment. If a device matches both OR clauses, they will still only be included in the segment once so you don't have to worry about perhaps reaching people twice. Essentially, even if they meet both clauses or filters, they were only being included in the segment once.

Segment Filter Types

So here's a quick overview of the filter types. Again, we will go over this once we hop into the platform, but did want to at least list them here visually. You can see, you can create filters based on the first session, so the first time someone logs into your app or website; the last session, the last time they visited your website or open your app; session count, it's the number of times they visited your app or opened your website; as well as the usage duration, just the amount of time they've actually spent in your app or on your website.

We have a lot of others here such as amount spent, purchase item. Those are going to be for mobile apps. You can even filter by language of the person's device, the app version they're on, the specific device type, user tag, which is a great one that we're going to get into a little bit more, location, and then a few others as well, such as country and email.

So now that we've done a quick overview of what a segment is and the filters, let's get into advanced segmentation because this is why you guys came today. Let's get to the good stuff. Advanced segments. Essentially, what we consider advanced segments are, it's combining a lot of filters to create one segment. Again, this just lets you get a little bit more targeted and specific with your group to hopefully engage them on something you know that they're interested in or based on a specific event.


I do also want to be clear that for advanced segments, some of the advanced filter options do require additional code to be added to your app or website. Essentially, a lot of the custom data that you would send to OneSignal is going to require a little bit of code to be sent.

One is the form of tags, what we call tags. Essentially, these are just user attributes. They are very important for advanced segmentation, but it does require a bit of setup work and a small bit of code first to have that custom data or custom tags sent to OneSignal. We do provide good documentation on this. We won't get too in-depth in tags, but we'll show you how these tags are going to be crucial to creating advanced segments.

Again, after you have added the code, this will allow you to provide more personalized and just even relevant targeting with your messages, which is what we would like to stress here today.

Some common applications of advanced segmentation are grouping your users by interest, so things you know they definitely want to get information on. Incrementing by the number of times actions are taken, how many times they may have visited a specific page on your site. Abandoned cart scenarios, which are going to be particularly useful for e-commerce sites. And then any other event-based triggers that you might want to set up.

To go a little bit further, I wanted to include a few examples here that we can walk through quickly. And then we're actually going to hop into the platform to go through these examples as well.

To build on those common applications, first, a good example is, let's say you're a mobile gaming app and you want to send all of your subscribers who have completed level 10 or higher within the last 24 hours of free coins promotion so that you encourage them to keep playing the game and stay in your app longer.

Another example is if you're a news or media website and you want to send, let's say breaking political news article to all subscribers who have visited the politics page of your website more than four times. Again, based on that activity, you know that they're probably going to be interested and want to read a new article, a new political breaking news article.

Next, we get into an e-commerce website, and this is essentially prompting subscribers that put items in their cart within the last 24 hours to actually go ahead and make that purchase. A quick reminder saying, "Hey, you left something in your cart, don't forget it." You're prompting them to spend money on your site.

Finally, a restaurant mobile app. This is where you might want to send a 20% off dinner coupon to subscribers that are currently within about five miles of the restaurant, so building on that location data and who's around and so that they can quickly see that notification and be like, "Hey, sure, I'm going to go buy some dinner quickly since I know I can get it a little bit cheaper," and spend money with you.

Those are the examples that we are going to go to. Let me quickly exit out of this presentation and then we're going to hop into our demo app. We have a few demo apps here. I'm just going to go into DEMO Testing. And then actually, I'm going to click on Audience and then Segments. This is where I want to create the segments.

As I mentioned in the examples, the first one I highlighted was a mobile gaming app. That's where they want to send a notification to anybody that has finished level 10 or higher and within maybe the last 24 hours to send them free coins.

We're going to click on Create New Segment. As you see here, these are all the filter types that we ran through in the presentation. You can see them here. The one we're going to select is going to be called User Tag. Again, this is where I've already set up my site with the relevant tag information to then be able to segment by.


Using User Tags

For user tag, you actually get to type in the tag that you have directly. As you can see here, I've typed it in before. It's finished_level. And I do want to highlight this is case sensitive. If you use caps, you will have to use them in at the exact time and place that you did. A good recommendation is maybe just to keep them all lowercase, so that way you never have to use caps.

We're going to type in the tag directly. And then you see here on the dropdown, I want to reach anyone that is higher or greater than level 10. I'll select greater than, and then put in the value directly, 10. You'll see here, again, this is just our demo site, so there's not too much data. But right now, you'll see that anybody that has finished level 10, there are 11 people that are 10 or higher. This is a great way to send them a free coins promotion so that they can continue playing in the game.

What we can also do is even add another filter here if we out AND, and you can easily add the last session filter. Let's say I want to do anybody that's within the last 24 hours. Oops. It looks like we don't have any data of that in here, but let's say I'll change it to a week, 168 hours a week. Again, this is just our demo site. It looks like nobody's playing on our made-up game too much within the last 24 hours. But you can see how you can combine these filters to get a little bit more granular and specific.

Again, you can gauge who's been playing the game a lot and then reward them for that, and then also prompt them to keep continuing to play your game. We're combining the user tag in the last session filter here. You can always name your segment like Free Coin Promotion Level 10+, whatever makes sense for you. I'm going to X out of this, but let's go into the next example and create another segment.

The next one is, we are a news media site. What we want to do is we have some political news articles that we want to send. Again, I'm going to go to the user tag and I have created an interest. Let's see, it's called politics. We're going to say politics. And then in the dropdown, I'm actually just going to do IS here, and we've created a value of 1. Anybody that has opted into politics, here you'll see with a value of 1, we know they're interested. Again, it looks like just three people are showing up here, but then we can easily send them the latest political breaking news articles to drive into their interest and probably get a higher click-through rate.

One thing I also want to highlight is that you can also do IS or IS NOT, or even just EXIST. If we do exist, this is just anyone that has the tag. You can also do, DOES NOT EXIST to understand all the users that don't have this tag. This is also a great way to just understand, okay, these people aren't interested in politics, so I'm going to send them less. Or maybe every once in a while, I will send them some news around this to see if they might be interested in that topic as well. But you can even just exclude them from any political news just to make sure you're not sending them that information. It's a great way to use these additional options, whether there IS, with a value or EXIST or DOES NOT EXIST to understand your subscriber base and send various segments based on that.

Next, let's go into the other example I highlighted, which was an e-commerce website, wanting to send abandoned cart notifications. Again, this one is also going to be using the user tag. This time, the tag is, I believe it's cart_update. What we want to do here is cart_update. The cool thing is that with tags, you can also add a UNIX timestamp to get a gauge based on time or when they've done something or when an event has occurred.

For this one, we want to do time elapsed is greater than, and we're going to put in... This is in seconds. So, 3,600 seconds is essentially an hour. Essentially, we want to reach anyone that has added an item to their cart for over an hour. Those are people who, for whatever reason, didn't complete their purchase. We want to send them a notification to say, "Hey, you left something in your cart," or "Hey, these sunglasses are waiting for you to purchase them." That's a great way to take advantage of the UNIX timestamp and adding that to a tag to then create an advanced segment here.

Obviously, if you wanted to do 24 hours, my math might be showing, but you get the idea. It looks like it's the same amount. But you can change this. It doesn't have to be an hour. You could do 2 hours, 24 hours, whatever makes the most sense for you and your subscriber base and just your even site activity.


Using Segments Based on Location

And then next, the other one I wanted to get into is the location segment. I do want to be clear that the location one is going to be for mobile apps only, but as you see here, there's a location filter. In this example, we are a restaurant app and we want to send a notification for 20% off dinner to anybody that's within maybe five miles of our restaurant at any given time. What we do is we'll basically enter in the latitude. In this example, I'm actually just using the latitude and longitude of our OneSignal office in San Mateo, California. But what you want to do is enter the latitude and longitude of your restaurant, and then you'll put in the radius that you want to get within.

For instance, let's say within about five miles is about 8,000 meters, give or take. I'm probably missing a few. But you'll see here, even in this dummy data that we have in our site, there are about five users that are within this location, within about five miles of the restaurant, for example. This changes. It does get updated. OneSignal SDK will actually update the new coordinates about every 5 to 10 minutes. That's great. Again, it's giving you a fairly good accurate idea of who is within that restaurant or your store at that moment so you can send them a notification and encourage them to come buy some food or come purchase some of your products. This location filter is really great. Again, it is going to be mobile apps specifically.

Those are the examples. maybe the other one I wanted to highlight, since I did mention that combining segments is a lot about just adding a lot of segments, I'm going to hop into one of our other demo apps. We have a few here at OneSignal.

I'll create a segment here, and let's say, this time I am a food recipe site and I am providing different recipes to a lot of my subscribers. Again, I'm going to use a user tag filter. But in this example, I know that I have users that want recipes like slow cooker recipes. I'll do the IS value. You'll see we'll have 10. But I have a recipe that would be good in, not only slow cooker, but also in Instapot. What I'm going to do is hit the AND filter again, and I'm going to select User Tag, and then type in Instapot again is 1. This is actually a good example. You'll see that it doesn't look like I have anybody that meets both of these criteria because I use the AND filter.

What we might want to do is instead of doing AND, we can actually just go to the OR filter. Again, I'll go to User Tag. Type it in. The tag is correct. I'll do one. Again, maybe in this dummy data, or if I perhaps typed the tag wrong, you'll see that there doesn't seem to be any users in the Instapot one. But again, this is great. It's a good example of using the OR filter as opposed to the AND, because again, AND would mean they would have to meet both criteria, but OR means they can fit into one or the other filters.

I just wanted to highlight that maybe in this recipe... I think here, let me actually get an idea of what the other filters are. I think I just typed in the filter wrong. That again is a good understanding of how you can see how the case sensitive filters are. If you type them incorrectly, you're generally going to get a zero. But here's a good example of a segment already created where we're just sending people recipes that fit a lot of different segments. As you can see here, we use the OR filter for these. Maybe it's a recipe that you can do in Instapot, it could be a freezer type meal as well. That's an example of adding in lots of different filters to get a segment.

Again, we could even add another filter to this. Let's say we added an AND or another OR, like we wanted to do the last session. We want to make sure we're getting anyone that's been to the site within the last 48 hours. Again, you'll see that this changed slightly, but this is a good way to take advantage of people that have recently been active in your site. You can do this within the last 48 hours, the last week, the last month, again, whatever makes sense for you and your site and users.


But I've walked through a few examples. Again, it highlights where you're using both the automatic filters that come with the OneSignal SDK, but also the custom data filters of data that you send to us and how that can really help you get targeted and granulated with your segments, and hopefully increase your click-through rates with your notifications.

I know I've been gabbing quite a bit now. I wanted to save some time for questions. Feel free to just either raise your hand. What I can do is actually I will allow everyone to speak. If you have a quick question that you'd like to ask, feel free to do so

You can always just type a question in the chat as well, if that's a little bit easy for you, but hopefully, you should all be able to unmute yourselves and ask directly if you like. Any questions at all?

Perfect. Well, one thing I wanted to highlight is, last year, I did a webinar on Segments as well. That was the intro to Segment. If you need a refresher, that's always a good place to start just to create some basic segments. And then maybe you can review this Advanced Segmentation webinar to get a little bit more targeted and personalized with your messages. I also recommend reviewing our documentation on data tags and sending that custom data information to OneSignal.

Again, any questions at all, feel free. Don't be shy.

All right. Well, hopefully, that means I did an okay job and I've covered everything. We'll definitely be following up hopefully with a short video of this demo. I encourage you guys to visit the OneSignal YouTube site as well, where we have recapped a lot of our webinars and it holds a lot of information, helpful information, as well as our documentation is going to be super helpful. Finally, you can always reach us directly. If you come here, you can click this little button once you're in the dashboard and actually reach out directly to our Support Team for help as well. We'll typically try and get back to you ASAP.

Great. Well, if there aren't any questions, I'll go ahead and end the webinar, but thank you all so much for attending and for being OneSignal customers. We look forward to helping you engage your own customers with notifications, as well as hopefully in-app messages and SMS as well.

Thanks so much for being here, everyone. You guys have a great day. Take care.