Zero to 7-Figures w/Push Traffic, My Biggest Profit Producing Insights | Andrew Payne, AWeurope 2019


Thank you for the introduction.
I’m Andrew Payne, based in Dallas, Texas, and I’m an affiliate. So I only caught
the last part of the previous presentation. This might be a little
polar opposite because I’m an affiliate. I have done, currently, do, a bit more of
the aggressive push notification advertising, and I was at the beginning
of the trend. So just backing up a quick second though. About three years ago,
I got into pop traffic and I was really interested in building stable campaigns,
trying to grow revenue, but I was having a difficult time finding strong offers. I
would have a great campaign, great offer for a short period of time, and then
things would just fall flat, and you go one, two, three months before finding
another offer. So at the end of 2017, I took a break and I went to Affiliate
World Asia, spent time there, went down to the STM Island party. Just, kind of, you
know, cleared my mind a bit before coming back and I discovered push notification
traffic. So today I’m going to share with you how I discovered it some of my best
practices that’s allowed me to produce several million dollars in revenue,
across a variety of offer types, and a small peek into some of the tools I’ve
built that allows me to manage that. Over the last 18 months, I’ve delivered more
than 6.5 billion messages to over 36 million users and I
primarily focused on the US market and tier 1 countries. In the beginning,
my initial focus was to work with countries that had a lot of offers, that
allowed me to have a variety of opportunities to make things work. So
what happened was I got back from my trip, cleared my mind, and I’m trying to
understand what’s my next step. Do I want to explore native, Facebook, what’s what’s my
next direction? So I’m on AdPlexity, and I spent quite a bit of time
researching competition, what’s going on, and I found there was four landing pages
at the top of AdPlexity that had been running pop traffic, of all things, for
nine or ten months consistently at high volume, but I wasn’t sure what they were
doing. This is around January of last year. So I visited the URLs, subscribed to
their push notifications. I downloaded the landing pages, and I sent them to a
friend of mine, who’s a little more tech-savvy with the code because I knew
absolutely nothing about push notifications. So between him and I
collaborating, in minutes, we got a good idea of what was going on, but what
confused me was the competitors that I had seen. They were sending promotions
that were not very targeted, very poorly written, weren’t promoting very strong
offers, and they were sending the traffic a lot of times to like native ad pages.
Just general generic articles that have ads for native websites. So I was
confused with how they were monetising things, but I wanted to figure out what
was going on to see if this was really a viable idea, because they were obviously
running a tremendous amount of traffic. And my thoughts were like, if they’re
running so much pop traffic, which is what I specialise in, at this point, how
come I’m not doing this? What are they doing that I’m overlooking here? So
I set up a solution tool called OneSignal. which is a very popular push
notification platform because my goal was to just to quickly set up a test to
figure out what the results would be, and if it’s worth pursuing more. So I got my
developer, he spent a couple of days putting together a little bit of code so
that we could properly track everything. Sending the user data notification
information and all that stuff to OneSignal from our custom landing pages, we technically ripped landing pages at that time and just
modified the code. Over the course of about three days, we set this code up, ran
a few thousand dollars of media spend and my initial results was -35%, and for any affiliates out
here, who’s an affiliate? Raise your hand. Few of you guys. Any pop affiliates? Push
affiliates? There we go. So the first results that I had was -35% and for us, that’s exciting because I knew absolutely nothing about
this, I just did a very basic test, so I knew there was a lot of
potential to improve the campaign and to grow from there. So it’s a winner, cha-ching!
I felt there was a lot of potential and opportunity there, and over
the next week, things continue to grow pretty rapidly, and I was sharing this
information with a friend of mine, and he was like, you know, bouncing ideas
back and forth with me and because he wasn’t familiar with this traffic type
either, but as I started to show him my results and things are ramping up quite
quickly, I told him, I said this is gonna make us millionaires and it later did,
but going back to what I said, I knew that as this was scaling over the next
several months, that a long-term solution, I needed to pick a custom platform. I
couldn’t rely on a third-party tool like OneSignal. The competitor that I had
found, he was using his own custom tool, as well built and based on Firebase,
that’s the name of the solution that Google offers, and so I started to
continue growing and scaling my efforts on OneSignal, acquiring new subscribers,
promoting to those subscribers, and we’ll get more into that in a moment, but at
the same time, we were developing our own custom platform on Firebase. Eventually, I
had to switch completely over to Firebase, kind of on short notice because
we kept scaling our traffic so much — as one example, overnight we had launched a
bunch of new campaigns and the servers at OneSignal crashed. I don’t know if
it’s me, but the timing on multiple occasions was too coincidental to not be
me. So we quickly switched over to Firebase and continued scaling from
there. A couple of ways or the majority the ways that we acquire subscribers
and almost every ad network out there, who doesn’t already have their own
publishers, this is how they’re doing it as well, and this is how the competitor
that I discovered was doing it in the beginning. There may be other competitors
that started before that. I’m certainly not the first. He may not have been the first,
but certainly in the earlier trend of this. So my primary way of generating
subscribers was through pop traffic. That was my core strength anyway. So it really
aligned with what I was doing, so that made up the bulk of the subscribers and
currently how we get the majority of our subscribers. Monetizer was also another
format that was being overlooked, that a friend of mine and I started getting
into and it worked really well for getting subscribers. We would set a CPA
and we pay a specific cost per subscriber, per country we were targeting,
which was really helpful because then we didn’t have to manage it as much. Another
format that we get subscribers from are select affiliate friends of mine, who
also do pop traffic. We organise a specific CPA for them and we basically
pay on a per subscriber basis. They send us traffic as well, so the idea behind
all this was how do we rapidly scale things? How do we diversify where we’re
getting traffic from? And that was the top three ways. Banner traffic is
something we’ve used at times. It’s not as effective, but it does work. More
generic thing, so you certainly want to work with ad networks that allow for
simpler creatives. And direct deals. There’s been, on occasion, I’ve
purchased a flat rate buy from a whole website. I’d buy all their Android, Chrome
traffic, as well as all of their back button traffic for Android, Chrome and
they would send it to me and I would turn that traffic into subscribers, and
we’ll get more into doing that in a minute. Just before we get too far, is
anyone in here not familiar with push notifications? Raise your hand, if you are
not. Alright great, so no one is living under a rock these days. So what’s really
important when you’re running an affiliate campaign, obviously, I’m
approaching things from an affiliate side because unlike a lot of these other
ad networks, I’m essentially my private ad network. My goal from the
beginning when I discovered this, was not to just be an affiliate. How can I
take things to the next step? How can I own something? How can I make something
more stable for myself? And push notifications, I had noticed right from
the beginning, that I had that opportunity to have an asset, the
subscriber database. So I’m essentially my private ad network, myself. I’ve never
ran a single push ad from another network because my goal is to acquire my
own subscribers and monetise my own subscribers. I didn’t have a great list
of publishers already working with me. So I had no one to reach out to. A lot of
other ad networks, they can say, hey we have this new ad site, let’s set it up,
we’ll pay you a rev share. So they’re not really paying upfront for the
subscribers, whereas I have to pay upfront, therefore, my goal is to monetise
those quickly. That’s one thing that separates why some of us have to be more
aggressive and others do not. Not to mention, if you’re scaling things quickly, as we started getting to $5,000 / $10,000 / $12,000 / $15,000 a day in ad spend I was a little guy at the time, I was not able to float that cash. So I had to
monetised to get the revenue coming back into my pocket quickly so I could continue growing. So we had to go through and optimise our Firebase code several times because I needed to reduce my costs. Right from the beginning, my game
plan was, if I can reduce my cost per subscriber, that allowed me to buy more
traffic at a higher cost, higher CPM, or whatever bid it is for that traffic type,
and that would allow me to be more aggressive to acquire more of the
traffic, and to make things more difficult for you guys if you tried to
step in and take some of my traffic. So right from the beginning, I spent a
lot of time optimising every single step of our entire process, up to the point
where we developed and redeveloped our Firebase platform three times, taking
what we’ve learned every time, rolling it into the next step, and keeping it really
lean, really organised, and tracking just the most essential data. Certainly we
serve our code and our landing pages from Amazon S3 and the CDN they have, Cloudfront. That allows us to have very quick and very
fast-serving globally. Another thing that became important was
to track your performance. As a marketer, we should all be doing this, but it
seemed to be really important in this case because we’re tracking how much
we’re paying per subscriber from all of these different sources. Both pop sources
and otherwise, and down to even the placement level in some cases, that if
there were certain placements that we’re delivering a lot of traffic and a lot of
subscribers, because we found a huge discrepancy from getting subscribers
from one source, at a certain price, and the revenue that those group of
subscribers would produce would be great, but that another source, we’re
paying the same price and we’re losing 50% on that revenue from those
subscribers. So there’s a difference in quality, certainly. In the beginning I spent most
of my time focusing on very adult demographic, male demographic type
of traffic, and so we were doing a lot of dating offers at the time, when I started
to mix that and started going more mainstream, that’s where we started
having a lot more issues with the quality of our traffic, and so it became
more important to really focus in and narrow down to what was working, what was
not working, and to adjusting so we’re profitable on all levels. Landing pages,
obviously, are important. The process for getting a push subscriber is we’re
buying a pop-up traffic. We send that to a landing page. That landing page asks
the user to subscribe. They have the option to subscribe or deny. I’ve tried
and tested everything from simulated video pages to graphic images of movies
to adult-themed or a specific-themed type of promotions. Testing all different
styles and everything you could possibly you think of and imagine. We’ve ripped pages,
modified pages, created pages, tested hundreds of different pages. At a
certain point, you stop seeing a lot of change in costs, and so we just
kind of stuck to a group of landing pages that we found to be performing the
best after a period of time, but they did have a significant
impact on reducing our costs. There’s a thing that I call my subscriber backflow,
and this is where the last speaker might not like what I have to say a bit, but
when we send a user to our main landing page, if they chose not to subscribe to
our database, or they click the back button to leave our page, we would send
them to a new domain. That new domain would ask them again, would they like to
subscribe? Some would say yes. Some would say no. Some would click the back button
again. If they said no or click the back button again, we would send them to
another domain. We would do that five times. So in total, some users saw up to
six times, a subscription page asking them if they would like to subscribe.
They had the option to, but the only way they could really exit the page is if
they closed out of it. So yes, that’s a bit of an aggressive tactic. It’s not
malicious, just that we didn’t want them to leave. That helped us to recover around
30% of subscribers that we would have lost. So that really, really reduced the
cost we were paying for a subscriber on average, and it also grew our database
much quicker and there’s a screenshot there that shows you just a
small snapshot. This is our backflow snapshot. Five steps, and you can see the
number of visits and the number of subscribes recovered from those steps.
Quite interesting. At the very last step, we would send them
to a smart link like Monetizer or some other promotion that I had set up,
like a sweepstakes offer or something like that, and we’d make very small
change out of it, but in general, recovering these was really important
because these were unique users. Users that did not subscribe to our original
pages, but we wanted to keep them. The next thing we had was what we call
sleeper apps, and I’ll explain why we call it that. It’s really my friend of
mine that I collaborated with and I introduced him to push when I discovered
it. He’s a big car guy, like huge, and we were thinking of the same backflow setup
and how can we use that to our subscribers that have
already subscribed, because initially, someone that would successfully
subscribe and choose yes will choose we would like notifications from you,
we would send them to a smart link like Monetizer and we might split test
that by sending them to a regular landing page with a sweepstakes or
dating offer, or whatever was performing strongly at the time. We’d split test
that, so it’d be successful, subscribe, redirect to one of those two, but I don’t
want someone else to have the opportunity to monetise my traffic,
before I had the opportunity to do that. So I decided to test out setting up the
same format of multiple domains that someone could subscribe to, but on the
front end. So they would subscribe, get redirected to another domain, ask them
the same thing, if they would like to subscribe or not. Anyways, they would
get brought through that same process, but on a forward sense. So anybody that
would subscribe to these sleeper apps, they are not unique users, they’re users
that we’ve already had subscribed. So these are the duplicate users the
previous speaker was talking about, however, we approach these differently.
This is why we call it our sleeper. These are not something that we
send constant promotions to, these are people that for some reason, more than
39% subscribe more than one time, and in many cases, it was the exact same landing
page they’d seen. I don’t quite understand this because a lot of these were in the
US. So maybe that’s our intelligence level, I don’t know, but anyways, so we
would take these users and they’re kind of like dormant. They just sit to the
side, we don’t do a whole lot. We do occasional promotions to them, or we send
promotional messages to them for our strongest offers. So if we have an offer
that’s super hot, is really killing it, then we’ll send out a message to these
guys, because we know they subscribed before. So we don’t want to bombard them
with a tremendous amount of messages, but we don’t want others to monetise our
traffic. We want to monetise it ourselves. So we would set up a certain schedule of
sending out messages to these guys and we would rotate them. There’s five
sleepers, each sleeper will always have the same duplicate people subscribed to them. So we might send a message to the first
one today. Second one tomorrow. Third one, you know,
next day, and just rotate things. Not even that often, actually. Much less frequently
than that, but it really helped because over a period of time, like 7 to 14 days,
we would make more money doing that than if we redirected that traffic to a smart
link or to another promotion. So our goal wasn’t to just be overly aggressive,
although, it’s a bit aggressive a technique. We weren’t trying to be overly
aggressive and hound these users with ads and promotions, it just made the best
economical sense and made us the most money, and that was my goal. So we also
set up a backflow monetisation process, very simple, and this kind of information
has been shared on the STM Forum before, in the past, and so this is a tidbit that
I learned a couple of years ago, and this was the first opportunity that I had to
actually implement it, or that I felt was worthy of implementing it. So what we
would do is, we send out a message to a user, they click on the ad, they go to our
landing page, some of them would convert, some of them
would not, some of them would exit. For those that click the back button, we would send
them to a new campaign that that rotated all of our top offers, all
of our best landing pages and best offers, and we’d set it up to where they
would infinitely see these offers, for as long as they continue clicking the back
button on any of these. So we just essentially, cycle them through all of
our top offers, and after doing some small tweaks and improvements, this
accounted for roughly 12% of our overall revenue. We’re generating 6-figures of
revenue per month. So when you add 12% of that coming from just the back flow, that
really adds up quickly. So again, this whole game is about economics, it’s not
about being aggressive or malicious. The techniques are a little just divisive though.
So messages are certainly one of the most important factors with your
push notifications. There’s a lot there in the beginning, you know, we’re working
from scratch. The person that I had discovered doing this before me, they
were poor language, misspellings, didn’t make sense. So we certainly approach things a bit more properly. We initially set up
and started doing split tests of hundreds of headlines, and hundreds of
descriptions, and hundreds of icons, to try to find what type of messages worked
best, which ones don’t. There’s a few main important parts of a message. The
headline and the icon is extremely important, because the icon is the most
visible aspect of the push notification and that catches their attention. Your
headlines are next or equal in the most important aspects. So those two
pieces, you really need to focus on. The message copy or the smaller text,
typically, we put more of a call to action there, but that’s usually
not as effective as the headline and icon. Another important part is the time
of day that the message was sent out, and we’ll talk a little more about that here
in a moment. There’s a few other elements that I added here that says what’s
not important. There’s an icon item called the badge, this is what shows up at the
very top of an Android phone, when a user has a message from an app or from a
website. It’ll show an icon, we split tested dozens of different icons.
They’re so small most people don’t notice them. We never got any significant
improvements and results. So we have just some default ones that we use and just
leave it at that. Images, same thing. There is the option to
add a larger image to your push notifications, but in most devices,
especially the newer devices that image is not visible, unless the user actually
selects a down arrow when they receive a notification. Most do not. Most just open
the message. So it’s not that important. We have tested adding the
little click buttons, the different options you can add for
open or whatever additional buttons you can add. We tested pretty much all
the features. Nothing really made a significant improvement. So that’s a
little contrary to what the previous guy said. It’s a cool feature, we just never
saw any real results from it, cause I tested everything. Not saying it doesn’t
help in some cases, but most of the time at the volume we were doing,
we never noticed. So we never put a lot of time into it. There’s
two main types of messages that we’ll send out. We have what we call our
general messages. So that’s something that really appeals to everybody. It might
be more generic, simple, but a lot of times, we can get higher open rates from
that. So if you look at the example I have, one new message,
generic icon, click to read. The user who receives that might not know where it’s
coming from or who is coming from. You can certainly get more creative. You
could say one new message from Jessica, or whoever and it might be, more
relatable to the person. So those typically generate a higher open rate,
but because it’s so random, and the person is just out of curiosity opening
it, the conversion rates vary on the offer you’re promoting. So it’s not
consistent at all times, and because of that, the quality is hit or miss. So it
just depends on the type of offer, but the one great benefit of using this type
of format is you can really test a lot of offers quickly by using this, because
you can get a lot of opens, a lot of a lot of traction, a lot of traffic rather,
and you can test a handful of offers really quickly, just to see if there’s
one that sticks out, then you can move over to the tailored message. So be very
focused, very specific. You get lower open rates typically, but you have a higher
conversion rate because you’re pre-qualifying the user prior to them
opening the message. So in this example, if you’re a Call of Duty fan,
and you have a something related to a Call of Duty icon, excuse me, that would
at least catch your attention and you might open it out of curiosity, but
that’s very likely that you’re gonna be somewhat in that demographic of a user.
So that helps improve the quality, as well as have better retention. You get
less unsubscribes from your database because of these very focused messages.
Something else that we tested and explored, was translated messages. So even
though most of my attention has been on English-speaking countries, for example,
in the US we have a large group of Spanish-speaking users, and so at the
time they subscribe to our database, we would log and retain
what browser language setting they were using, and so we began to
translate these messages that we’re sending out and delivering them to the
user in their language that their browser is set to, that way they’re
getting their default language, and it’s not mixed and matched. Now, this doesn’t
have any huge improvements in most cases, but it does help. You can get higher open
rates for those segments in certain countries and get better results from
your conversion rates. We later on, as we started to develop our custom tools,
we added this feature as an automated way working with Amazon Translate, so we
simply load in our messages in our custom tool, we click translate, it’ll
translate to all the available languages that Amazon supports, and so it’s pretty
automated. Any languages that we have a lot of traffic for, or that we see a lot
of results with, or we want to be a little more fine-tuned, we go back and we
manually update those, but we don’t spend a lot of time on the languages that we
don’t work in too heavily. So not a huge bonus and benefit, but definitely helps,
and it’s one way that we differentiate and separate ourselves. Like I mentioned
earlier, time of day extremely important. Early mornings, late evenings, seem to be
really strong for the US, but holidays always varied for some reason. Some days,
some holidays were great, other holidays like Black Friday, we just had like
really pathetic results, and my assumption is that users are being
bombarded with emails, text messages, other push notifications, ads everywhere.
So they’re just more blinded to things on those days. Maybe others have
different results, but that was my results from those things. So I like to
focus our messages around the stronger times of the day, because the
results greatly vary. Like first thing in the morning, you’ll send out a message,
say we sent out a message like 4am, someone wakes up at 5am, the
first thing all of us usually do, or the majority of us, is open up and look at
our phones. What’s new? What’s the news? Who texted me? What did I miss? What’s my campaign
stats for the day? So that’s the same thing the general consumer is doing as
well. So first thing in the morning is really strong.
Sometimes early afternoon, like around lunch hour, and then evenings, and
weekends are certainly quite strong as well. So good thing to do is to send on
the best times. Figure out the best times for whatever country you’re
working with and go from there. We also did a lot of split testing, like I mentioned. A lot of icons, headlines, body text. So we’ll load in just a large number of icons and all these other
other features, and we split test them over multiple days. Sometimes a week,
sometimes two weeks, and we test them going out at different times, because
sometimes we have one message that goes out, performs really well right now, and if we
split tested it with three other messages, but they had a clear winner,
tomorrow we can send it out at the exact same time or a different time and that
message is not the winner. So you can’t judge too quickly.
At the time, I was not aware of any split testing capabilities. OneSignal
didn’t have it. Our custom tool wasn’t ready for it, but eventually we
found our own process of getting this set up, but I suggest testing messages
over many days, and many different times, before making your judgement, and we
typically keep the top 20% of our performing messages, and then we add new
ones in for testing. So we constantly have messages in a rotation. A user won’t
see the same message too often. So to keep everything fresh, everything
unique, same with the icons, we like to keep things going in a rotation fashion.
So keep things less exposed and less drowning of the same information.
Some of the top performing verticals I’ve worked with over the last 18 months.
Dating. Dating is my biggest one for sure. I’ve done over $1.2 million
in revenue just from dating. Sweepstakes, such as your iPhone, Android,
type of sweepstakes offers. Streaming offers, I didn’t list that one here.
Streaming offers have been really big as well. Branded SOI. So things like your
Visa gift cards, Walmart gift cards, those type of things do really well because
the user can relate to it, and apps. So utility apps, games, social casino
apps, lots of different options there. These are just some of our top
performers, but we’ve also ran nutra, that works really well, especially in
certain countries, and there’s a lot of other things you can run as well. So keep
in mind, there’s a great number of flexibility with what type of
verticals you want to run. These have been really consistent and very strong
for myself. What I want to show you now is a peek into the tools that I
use. These are tools that my team and I have developed and that we currently use
today, and I hope you’re generating some questions because there’s some things
I’d like to mention. I mean, I’ll think about it, but I hope you guys have some good
questions at the end, because I’d like to tell you my predictions, and things like
that as well. So someone bring that question up later. So we have, opening it
back up one second here, we have a few core things that we work with. Message
scheduling, it became very difficult to constantly manage how we’re
sending out all of our messages. So we needed to build and develop a tool that
allowed me to automate sending the messages with some flexibility.
We wanted to send during specific times. We wanted to make sure
that we’re not over-exposing messages to users. We wanted to make sure that we are
rotating messages so they’re seeing a fresh or more unique messages ever so
often. We wanted to make sure that we’re able to split test and get a variety of
different creatives in front of them, so that we can look at our results and cull
out the stuff that’s not good. So frequency capping. Both on which users
were sending a message to and how often we’re sending it to them, as well as
frequency capping the messages. We’re not too focused on being very
specific with a frequency capping of each individual message. We typically
have dozens of messages and probably hundreds of unique combinations for
every category of messages that we send out, and then automating the translation
part. So this is a screen of our main campaigns page. So pretty basic, simple
set up, which country you want to target, what sources. Sources is where we’ve got
those subscribers from. So if I wanted the target from a specific partner that
sent us traffic, we can do so. We can select which apps. Apps are the domain
that we have acquired that subscriber from. Obviously if we want to get more
granular, we can do device ran, and so forth. Timezone you can set up. So we
typically set up multiple campaigns for a country. One country or bigger
countries might have ten campaigns, and we will target morning, evenings. We’ll
stagger which groups of users that we’re targeting during those times. So we break
things up very granular to make sure that we are maximising what we’re
sending to, and that we are managing things properly. So even though my techniques of
acquiring subscribers are a bit aggressive at times, the way we’re
delivering our messages are not always aggressive. I will confess, in the very
beginning, I was testing everything. I didn’t know what to expect. As a
marketer, it’s my job to test and figure things out because there’s no place to
go to read this stuff, no one’s gonna tell you this stuff, except for if you
come here. So I was sending anywhere from 1 to 20 messages a day. So, too much, I
will tell you that, but as I started to scale and buy more traffic, my goal was
to get the revenue back as quickly as possible, and while the case study earlier was nice. I appreciate the case studies. Any
success is great. I was dealing with 10, 15, 18, that was the range. $18,000, between
$10,000 and $18,000 was very consistently the revenue we were producing every
single day. We’re not doing quite that strong right now, but the majority of
last year was quite strong, and because you’re doing so much revenue, like I said
earlier, I was a smaller guy at the time, I couldn’t float $200,000 / $300,000 a month in
revenue or cash. So yes, 20 messages a day, if that’s what it took to get my
money back, I did so. Our quality was strong, eventually I learned how to
manage things more appropriately. I began to reduce how aggressive we were running
things. I began to split with like the sleeper apps and some other information.
If you look at the bottom four options on the screen, it says auto reset
grouping, sending frequency, collapsing, and sending limit. So the auto reset
grouping, we have different groups, or different types of messages. Like
sweepstakes messages, dating messages, utility app, social casino, whatever.
Inside of those groups, we have dozens or hundreds of combinations of
messages for those types. So we have a certain order that we send things. We’d
set things up to where our best performing groups or best performing
verticals would get sent out first. Every single day. That way we maximise our
fresh, new audience or the beginning of that day for that offer. We know these
offers are our strongest performers. We want to make the most out of it. So if I
turn that feature on, at the end of the day, things reset, and we send out
the next day starting with the most important things.
Sending frequency, we can set how often we want to send, because in the beginning
I was sending quite aggressively, anywhere from 60 to 120 minutes. As time
went on, we increased that to once a day or once every other day, and different
campaigns, we have different settings for different purposes, but we set things in
minutes. Collapse is something that we use, so if I send out a message that’s
related to a dating offer. If say, that user never opens that message, but it’s
still on his phone, and two days later I send out another message to that user
that’s about a dating offer, I don’t want him to see two dating offers or messages.
I want them to see only one. So this one will push out the oldest message, and
only show him the most recent, most fresh offer or sorry, most fresh message.
Our sending limit, if it’s set to zero, it means it’ll send to everybody in that
apps database, or I can set it to 20,000 / 50,000 / 100,000 whatever,
and so, we’ll only send to that many people at one time. So if I set this
to every 60 minutes I’m sending out a message, and I set the sending limit to
10,000, that means every hour, I’m sending a message to 10,000 people.
By setting things up this way, we’re able to consistently produce revenue every
day, all day, for months. Since I got into push traffic, I’ve not had a single
unprofitable day since the first week I started. We’ve been doing very consistent
revenue since day one, and that’s something that really, really changed my
view because with pop traffic, it’s very, very different. It’s very up, very down,
very up, very down. I hated that. This screen is a quick snapshot of how our
message batching system is set up. We can set up a single message if you want do a
quick test, put something in there, do tasks, check it out. We have groups of
icons badges and images that we can set that rotates different images in those
groups. We have the batch message. We can set a batch name, description, and so
forth. We can put a list of headings and a list of messages. So this is where we
would set up like 50 messages for the heading, 50 messages for the
description, and they would do random combinations, and they wouldn’t send the
same thing, until all unique combinations have been sent.
We can look back on our stats in our tracker, we’re tracking all this stuff, we
can see which combinations were the best performers, which ones weren’t,
and then we’d remove the bad, generate new ones, so then we’d have one batch of
our best performing messages, and then we’d have another batch that would be
like our test messages. Path allowed me to assign a certain path I wanted that
user to go through in our tracker. So if I’m promoting like a Walmart offer, one
of our giftcards, I can type in Walmart and set that up in our tracker, and we
can send that user straight to a Walmart offer or whatever type of offer I wanted
to send them to. Here’s a quick snapshot of how our translations work. We input the
messages, as soon as we save them, we can click the translations tab and auto-translate.
Everything translates. It’s certainly not perfect when you’re using
automated tools like that, but we can modify things and it works quite well.
Moving a little quicker here, we’re running out of time. Some of the things
that impacts your lifetime value is your message sending frequency, like we talked
about. The lower the frequency, the longer the user will stay subscribed.
The slower it takes to monetise a user. The types of messages matter as well. If
you’re sending out aggressive, you know, antivirus offers or things like that,
users are going to be really offended and they’re gonna unsubscribe. So just
because you have good performance from an offer like that, make sure that the
loss of value by getting a higher number of unsubscriptions balances out, because it
may not always be the best. Promoting a variety of offers, keeping things very fresh. Messages, offers, landing pages, sometimes
we’ll have one offer, but we’ll let the user see 10 different landing pages
just so that the first impression they get is always a very different look.
That really helps a lot. And send during peak times and only peak
times, that will extend the lifetime value of your users. Some additional
things that we’ve done is that I’ve run surveys to understand what brands, what
interests that users have on our database, that helps us tailor how we
promote or what we promote. We’ve also tested some of our own product ideas. So
I come up with like a digital product or a gaming platform or something like this.
We can do some tests to see if there’s any viability to it, and we’ve also been
able to do our own email list building, lead gen, that type of thing. So there’s a
lot of options you have when you have a traffic source of your own to work with.
So there’s quite a number of other things that can be done,
but that’s kind of the main things we work with on our additional side. If you
want to learn more about me, you can go to my site, it’s affiliatesuccess.com. I
will confess I’m not extremely active on this site. I do occasionally send out
information. I promise you, if I do send out something it’s worth listening to,
but there is a newsletter you can sign up to, and more than happy to communicate,
and if there’s interest you guys have, you guys want to learn more about push
notification from the affiliate side, shoot me a message through there.
I’ll get back to you guys or I’ll come up with like a blog post or some video
that you guys can learn a little bit more about things, but thank you for your time, and let’s open up some questions here
soon, but I will start since we have like 50 seconds left. I will start with my
predictions. So I’ve seen a lot of things change in the last 18 months. The biggest
thing changed around December last year. The overall user engagement dropped.
Things changed with Chrome browser updates, with how they’re served on
Android ,and it just kind of dropped. So our traffic is not quite the same. So
with pop traffic being served, it’s quite different but things are still strong.
These are all very valid and actively used techniques that I shared with you. I think that Google will also impact how
subscription works. I don’t think there’s gonna be a two-step process, I think
you’re gonna have to engage with the site in some way, spend a certain amount
of time on the site, use some type of interaction with the site. I’m very
involved with like the forums that the Google developers use. I read those quite
often and that’s kind of where the chatter is going towards, so they have
some questions lined up? Yeah, that was great stuff Andrew,
thank you very much for that. So now we have questions from our users, and the very first
question is what is your target CPA per push subscriber? Alright. That varies a
lot obviously. Every country is very different. In the US, you might pay
5, 10, 15 cents. In a country like India, for example, it might be half a
penny or a penny or less. So that varies a bit. I would start with working with
getting some tests going if you want to acquire subscribers. Start somewhere on
the lower end. Anywhere between 5, 10 cents if it’s like a tier one country. If
it’s a tier two or three country, you want to be under 5 cents. Get some
tests going, start doing some promotions, and see what kind of
revenue you can produce from that. What’s important is you can pay more for a
subscriber the better your monetisation process is. That’s why I spent a lot of
time focused on monetising my users, because I knew if I can monetise my
users better than anyone else, and if I can use every trick in the book I can
think of to reduce my cost, to acquire a user, that increases my profit
margin. So then I can pay more to acquire a subscriber. So just start testing, get
some base data, and then improve from there. It’s too difficult to say exactly what you should pay per subscriber. Okay the next question is have you tested selling ecommerce
products via push? Not really. The main thing is we’re looking for a certain
target revenue per user on average. Ecommerce products, I mean, if you’re
working with a push ad network that has a much larger database than what
we have, it might be viable, but for me we haven’t done that.
Right, the next question is — there was a question about what was your ROI?
My ROI? Yeah. Overall? Overall for push traffic. The first 6 months into the space,
my ROI was 100-125% very consistently, as I started to scale
even larger, we started averaging about 75% ROI and we’re still
doing about 60% ROI very consistently. Okay, next question is
what is the life expectancy of push database? Like, life expectancy,
you have push as a industry or push as an? As an inventory. So like a subscriber?
Yeah. So again, the more aggressive you are, you’re gonna drop
them like flies. So you have to have a way to replenish them. We’ve toned
that thing that down quite a bit. So I can tell you, we have some of our
database that subscribed over a year ago, and they’re still converting on offers.
So you can they can last quite some time. I would say the majority will fall off
within the first two weeks. Then you have a lot of that stay on
there that are dormant, and they will occasionally open up a message.
Again, the more aggressive you are, you’ll drop them faster. The less aggressive you
are, the longer it’ll last. So that depends on your own personal approach.
Okay, what is the biggest hurdle in setting up a push ad network? The biggest hurdle is
you gotta have the technical know-how or find someone who does, because there are
some technical challenges. Like I said, we had to redevelop my Firebase platform
three times. Not that it didn’t work the first time, but that we learned, and we
improved, and moved forward from there. I don’t know much on the technical side, so
I had to find a developer who understood push notifications, or at least had the
competency to learn it really quickly in the format that I needed it, and I
found a good guy, and we developed it, and went from there. So besides having the
technical aspects, if you’re getting into building your own ad network, you need
the capital to invest into your subscribers. You know, the more capital you
have, the faster you can build your database because you have to also keep in mind that you’re investing upfront to build
subscribers, and then you have to monetise them over a period of time that
could take weeks, months, or in some cases, years, if you’re being really
low frequency. Okay, does asking for push often hurt the conversion rate on the
side where the script is placed? It does impact conversion rate. So if you
have like an ecommerce product, a native landing page or you’re running
some other type of promotion, like a sweepstakes offer, and you’re just
wanting to collect subscriptions passively, it does impact conversion rate,
but usually if you have a good promotional routine set up, you can
recover that and more revenue. So there’s usually a boost. I haven’t personally
tested that myself, but I have a friend that set the test up and for his
campaign, he lost about 5% of his revenue by doing that test, and he gained 12% from the push subscribers. So net bonus was a 7% bump for him.
Okay, there was a really good question, what is the perfect frequency? How frequent you
should send messages? I would say if you’re wanting to be very conservative,
once or twice a day. You can be even more conservative as that if you have enough
capital and you’re not looking to monetise users too quickly, but once or
twice a day during peak times is a good rate. If you want to get revenue a
bit more quickly, I would say probably not more than like six to eight times a
day at the most, and that’s pretty aggressive already. Another question, in
one sentence what is push marketing? Push marketing is a way to communicate with
users from your website, if we’re talking about web push marketing, straight to their
phone, you get update, promotions, web pushes. Let’s say, it’s almost like text
messages but from your website. Someone asked, how your landing page
looks like? Can you share with us. If you go to AdPlexity, you’ll see dozens and
dozens and dozens. Some of mine, some or others, it’s very obvious, like that’s an easy question, but some of them are very generic. My
entire focus was just get subscribers. So I was not promoting another offer and
then prompting them to subscribe. Our entire design of the page
was to get them to subscribe, that was a sole focus. So if you go to AdPlexity,
do some searching, you’ll see plenty of pages that have like simulated videos, or
they have this or that, and if you click on the links, you’ll just see a push
prompt. That’s the whole point of the page, that’s the same format that we
used. Right, someone asked how are you tracking all the traffic? So we track
everything through our tracker. We’re using Voluum as our primary tracker, and
we just had everything set up through like tokens and different data that we
pass back and forth. We pass like message data, placements, we have our own custom characters that we use to track
certain private credentials and stuff. Okay, next question is how do
people get push messages? Where did they opt-in to get push messages?
Sure. So they have to be on your domain, your website, in this case and/or a page of it,
and you have to have a push prompt code set up, and they can opt-in there. So they
just go to your website or a landing page you have set up. That’s it, I think.
There are no more questions. We answered all of the questions.
So thank you very much Andrew.

3 Comments

Add a Comment

Your email address will not be published. Required fields are marked *