Home > > Newbie Follow-Alongs

---

Follow-along with Jeffff (31)

---

Hello there!

Jeff here, from Austin, TX. Just got a ticket to Affiliate World Asia 2019! I plan on making Affiliate Marketing a full career and am excited to get as far as possible over the next 4 months, before AWA.

In the process of following Amy's guide, I've finally setup my first full (sweepstakes) campaign with a LP and tracker. It's been running for 2 days now and I've gotten 5 conversions!



This campaign is with a single landing page and a single offer - next level up will be implementing multiple LPs and offers - but I was too excited to get a complete campaign up and running and I'm happy to see the offer is converting. Not sure what a good LP CTR is, but 1.08% seems kind of low to me. What is a good LP CTR?

I plan on sticking to sweeps offers. My next step is to setup more campaigns - including a couple LPs (need to work on these) with multiple offers to test (already have a handful that look promising)

---

I set up another campaign with 3 offers and 1 landing page. The top line is the same campaign from yesterday (with 3 more conversions) and the second line is the new campaign:



First (yesterday) Campaign: I cut website placements for sites that had a lot of clicks, but delivered no LP clicks - looking forward to maybe beating -50% ROI on this campaign tomorrow?

Second (new) Campaign: No conversions yet, but due to the rotation between 3 offers, I haven't gotten enough traffic yet.

I've settled on a region I'd like to focus my campaigns on. My strategy right now is to test as much as possible. Test different landing pages, offers, geos within a certain region - all while being as efficient as possible. I'm keeping a close eye on campaigns at all times to monitor spending, and cutting and moving on when it's time.

Next Steps: Add bot filtering, launch another campaign, and clean/make more landing pages...

---

Great that you started your FA!

What traffic source is this on and what geo if I may ask?

1% CTR on the lander looks low to me as well, sweeps landers usually have anywhere between 10-30%, depending on geo.

Some tier 1 geos are usually on the lower end.

---

It's exciting to be receiving conversions for the first time isn't it! Very happy for you @jeffff!

As for what's a good LP CTR - 2 points I want to make:

1)It can vary widely from geo to geo, from vertical to vertical, from lander style to lander style.

2)CTR isn't the most important metric - CR (conversion rate) is way more important. This is because CTR doesn't directly impact on profit margins, whereas CR DOES. I've seen low-CTR landers beat out high-CTR landers in terms of CR/profits.

Exception: When a lander's CTR is so low that it would take an impossibly-high CR in order for the campaign to reach profits. In this case you can exclude the lander based on CTR. Otherwise, try to cut a lander based on stats.


And like @erikgyepes has pointed out, developed countries (tier 1 geos) usually have lower CTR than developing geos.


What's the payout on that offer? If you've spent the equivalent of 10x payout in traffic cost, and the offer still hasn't converted, then it would probably be best to test another offer.

Lots of impressions but no clicks - please define "lots"? The important thing is look at cost vs. revenue. If a placement is in loss by 2x payout or greater, I would suggest to cut it.

Also: If a geo doesn't have enough traffic volume, don't waste your time - even if you're able to make campaigns profitable, you'd be making small money - not worth spending time on maintaining such campaigns.

Your plan sounds great - looking forward to seeing lots of efficient testing - that's the BEST way to success!




Amy

---

Of course, Popads and Malaysia. I have Binom setup on a server in Europe, so I think that delay may be (at least part of) the cause for such a low CTR. Going to stick to offers closer to my tracker from now on - I'm thinking specifically Eastern Europe - but considering this was my first converting offer, it's hard to forget about it!

---

^^^ Yep, on pops fast speed is crucial, best bet would be to use CDN for your landers, this way you no need to worry every time you launch campaign in a different geo location.

---

Unfortunately, that's a limitation of all self-hosted trackers - unless you deploy multiple installations at multiple server locations.

But like @erikgyepes pointed out - at least you can host your landers on a CDN to make them load faster. We need to optimize what we have control over, and ignore what we don't.

(Unless you want to switch to a cloud-based tracker - but the speed comes with steeper pricing.)



Amy

---

I’ve had some confusing results using the bot detection script by caurmen.


In Binom, when you are adding multiple offers to your campaign, they can be rotated (with a single-offer lander) or used in a multi-offer lander with links to multiple offers on the same lander. When rotating multiple offers on a single offer lander, you have the option to set weights for the rotation balance.


I set my lander to have two offers and added both my real offer and the ‘parses javascript’ offer to my campaign - after sending some traffic to it, I noticed almost an exact 50/50 split to both offers.


I spent about an hour chatting with Binom support trying to figure out why I was getting a 50/50 split on these clicks - they had no idea and could only repeatedly tell me to set the number of offers correctly on the landing page settings (which I had set from the beginning) I eventually noticed there where still weight settings for these two offers (you’d think the weights would be disabled if you were using a multi-offer lander, and this is also what Binom support said) but I decided to try it anyway and once I set the weights to ’100 and 0’, I instantly stopped getting a 50/50 split on the real and fake offer. After I made this change, all the js traffic started going to the real offer, so switching it to ‘0 and 100’ ended up fixing the problem.


I think what was happening was, my real offer link was getting switched into the "parses javascript" offer, so I was sending them to the real offer 50% of the time and google the other 50%.


It's been making me feel a little crazy trying to figure it all out - especially when I don’t see anything about this mentioned in the Binom documentation and when their support line doesn’t know anything about it… has anyone else experienced this?


I can see how if someone was running traffic to a landing page with two offers and the offers were getting rotated between two different buttons, it could be difficult to tell which clicks were coming from what buttons. I only noticed because I was expecting a high CTR on the js offer and a low CTR on the real offer, not 50/50. It also seems interesting that there is even an option to set weights when you are using a multi-offer lander. If rotation of offers was actually disabled (this is what Binom support told me was definitely the case) for multi-offer landers, why would the weight settings still be available to adjust and why would adjusting them to 100 and 0 have an immediate effect? If anyone from Binom reads this it would be great to get some clarification on this feature.


Anyway, I probably won’t be using this script again, because for every placement I cut (those with less than 70% CTR on the bot test) it was just as apparent for me to cut based on the number of clicks with 0 conversions.

---

I’ve been surprised how easily I’ve found offers that convert right out of the gate around -40% ROI. That part has been a lot easier than I expected. Is this a good range for an initial ROI to look promising?


I have a CZ campaign, with $0.30 payout that was getting around -42% ROI initially. Yesterday I cut 6 zones and in the 8 hours after those cuts, the ROI was -32% - too soon to tell, but I can’t help checking constantly!


I noticed that the traffic prediction, before and after I added the placement blacklist, looked like it dropped off more on the lower end. That makes me think the zones I cut were overall good for the campaign?


Before:



After:



After those 8 hours with the placement cuts and for the start of today, I decided to adjust the bid from 2 to 2.75 CPM, hoping to get a bit more of the traffic - after about 12 hours of running at a 2.75 bid, the ROI just for those 12 hours was -44.59%. Should I wait it out and cut continue to cut placements or adjust the bid again? Too soon to know?





Maybe I am checking too often and making too many adjustments.


In a previous campaign I ran into the problem Amy tried to warn me about…

Does the volume in CZ look large enough to be worth it? I was hoping raising my bid would get me a bigger chunk of the traffic, but traffic still seems too slow to profit much. And it has reduced the ROI.


So far, I’m 12 campaigns in with an overall ROI of -65.20% - right now my focus is on getting to 100 campaigns of good experience with this. I know I should be managing more campaigns at once, but because I haven’t invested much money into a single campaign, I want to practice trying to scale this one up first before making the same mistakes scaling multiple campaigns at the same time.


Anyway, it is really exciting to see $24 in my account! Only $476 to go haha :|





I’m open to any advice, so any suggestions or ideas would be great! Thanks in advance.

-Jeff

---

So I ended up going back to to my original CPM of 2.0 - I also made a few more cuts, in addition to the 6 placements I've already cut.

I cut iOS based on this:


And I cut all languages but cs based on this:

Good idea, bad idea?

This is my first time running a campaign long enough to make cuts like this, so I'm wonder at what point is it going to be too little traffic? The projected traffic is now down to around 40k impressions/day, less than half of the original volume before cuts.



One interesting thing I've noticed, is that the android webview browser has the best CR despite reading that a lot of people cut it from the beginning - even though it is such low volume and only 3 leads, I thought this was interesting:

---

So it has been about a day and a half since the cuts above - CR is up and ROI is down (or up since it is negative?)

Still some way to go, but I'm excited about these improvements:



Right now I'm still in the phase of constantly refreshing my stats and though I know this isn't exact, it seems like the ROI always gets worse in the night and picks up in the morning balancing things out again. I plotted conversions/time, but would this even be helpful in day-parting? Seems like CR/time would be a more useful metric.

---

If it's that significant might be worth it to run some split tests and see what's happening

---

Yesterday ended up being a new best CR and ROI on this campaign!



I've been refining the placement cuts and I think it's time to bump up my CPM - very interesting to see how steep the lower end of this projection has become.



It has been at 2, but just I adjusted it to 2.5 CPM. Excited to see how this plays out

I've been having a lot of fun with this and can't seem to help getting sucked in more and more haha

---

This is really strange - something isn't right.

Let me show you what my set up is - maybe you can compare this with your setup?

Adding the lander to Binom as a multi-offer lander with 2 offers:



Adding the "Parse Javascript Offer" to Binom:



Adding the click urls (outgoing links) to the landing page code:



And lastly - creating the campaign in Binom:



Your setup looks wrong - because there isn't supposed to be rotation between the '2 offers'. Each time the lander is loaded, BOTH "offers" will be triggered/served.

One link is "seen" by the human visitor (registers in the tracker in the "Clicks" column) and if the human clicks, the click will be registered in the tracker under "LP Clicks".

The other link is "seen" by the browser of the visitor and should register in the tracker under "Clicks" as well - and IF the browser IS able to parse javascript, then it should be registered in the tracker in the "LP Clicks" column.



You shouldn't need to change the rotation weight - I didn't.



Amy

---

This is a good indication, yes!

But then it would also depend on how much traffic you're getting initially.

If you were only getting a bit of traffic initially, then there may not be enough for you to cut all the way from -40% ROI to +20-30% ROI - and THEN you'd need to have enough traffic left to profit from, that amount to more than peanuts.


That's very creative thinking!

However - you'd be assuming that placements with lower average bids are less-desirable than placements with higher average bids - which isn't necessarily true.

Cheaper placements may not convert as well (which is often why competition is lower and advertisers tend to get traffic from them when bidding lower), but - well - they're CHEAPER.

Same thing applies to expensive placements - they tend to convert better, but then you'd be paying more as well.

So in the end, higher-quality/more-expensive placements won't necessarily net you more profits. It's a case-by-case thing.

What you DO want to make sure is that you're not paying $2 CPM for traffic that's making you $1 RPM. Figure out how to do that, and you're golden.


The slower you adjust the bid, the less money you'll waste on cutting placements.

Why not spend the time to come up with more offers and landers to test? It's always more satisfying to come up with better landers/offers that can make a bigger portion of the total traffic profitable, than to cut your way to green before you've tried your reasonable best to optimize landers and offers first.


For newbies CZ is a good geo - especially if you scale it to multiple networks.

But make sure you've made a reasonably good effort to test landers and offers before cutting - the more lander/offer testing you do, the more traffic you can stand to make profitable, and the less cutting you'd need to do, = the more profits you can keep.


Agree with you!


A case of so far so good!


Agree on cutting IOS and EN, but it may be good to leave other languages.

They're not costing you much at all, but can give the traffic network the "illusion" that you're targeting broad as opposed to narrow.

On many traffic networks, the broader you target, the more "importance" the algo assigns to your campaign, and the more traffic they'll send to you. But this isn't a set rule - just keep this in mind as a possibility and do your own testing.


Interesting indeed! And yes it's baffling that sometimes webview would convert well. But we know that more often than not, it doesn't convert well. So what we can do is exclude it from the beginning (especially for larger geos - to curb spend), and then throw it back in once the campaign is profitable.

Or - you could just include it from the beginning and keep an eye on it, and cut it later. Either way works.

BTW I would also cut the safari browser - looks pretty bleak at -80% ROI (unless you want to test more offers/landers, in which case I'd leave that running to get more traffic and conversion data).


The improvement is definitely something to be proud of!

Dayparting would definitely be a good idea if you're noticing a trend, but conversions/time wouldn't be a good metric, because the traffic volume will vary. CR/time would be better - but why not just cut say the lowest-ROI hours? For example anything lower than -70% ROI.

One optimization approach is to drill down into all variables and cut a couple of the lowest-ROI segments, wait to see the impact on the overall ROI, then maybe decide to cut a couple more - until you're in the green. Those segments may include just placements, or other variables as well.


The increase in ROI is encouraging! I'm happy for you!

Just a reminder here though to always focus on DAILY PROFITS rather than on ROI. At the end of the day, that's ALL that matters.

It's fantastic that you're taking your time to do a lot of testing with all the cutting and bid adjustments. As a newbie there's nothing more valuable than to do a lot of testing and observing results to further your learning. However, remember that one day when you're running campaigns "for real", it would be wise to keep this type of optimization to a minimum because it's just so damn time-consuming AND won't result in much increase in daily profits.

In the future when you have experience and want to be making sizeable profits, you'll need to focus more on testing offers and landers and on scaling, and less on optimization. This will be particularly true when you already have a blacklist of the worst placements in a geo on a given traffic network.

You're definitely on the right track with your 100 campaigns goal! You're making great progress - it will get you somewhere.



Amy

---

@jeffff We found you.

---

Great! Can't wait to get some traffic flowing from ZP now

---

Thank You Amy! I don’t know how you find the time to write such detailed responses

My bid adjustments must have been too drastic, I’ll have to dial this in - each time I’ve adjusted it from my initial bid, I’ve ended up changing it right back



Yes! I need to be better about this - and managing time between different tasks in general. But this is also why I am really loving AM, so many new things to learn and work on and the learning never seems to end



This is gold Amy! I love learning these little tricks



It hasn’t been converting since and has become quite an unprofitable browser overall. Must have just been an anomaly haha



Good catch! Just fixed this



Thank you, but it’s all thanks to your tutorial and this forum



Haha idk why I was so confused about this… it finally clicked after reading this and seems too simple NOW that I get it


Thanks again Amy! This is invaluable information

---

Just made it to Bangkok! PM me if anyone else is here a day early too and would like to meet up

---

It was GREAT meeting you in person Jeff!

Thanks for stopping by the STM booth!



Amy

---

Thanks Amy, AWA was awesome, would definitely go again! Been pretty busy since getting back, but I'm getting everything back up to start running campaigns and pick up on this follow-along, with a better budget this time. Will have updates soon

---

In PropellerAds, if using the Distributed option on 'Ad delivery method' I noticed that all the traffic seems to pour in at the top of the hour and then cut off. For example, if I set a daily budget of $24, propeller seems to set an hourly budget of $1 and spend it at the top of the hour until that dollar is spent, and then cut off traffic until the top of the next hour.

I was using this distributed option while testing out some offers in a bigger GEO, because buying the full amount of available traffic was a bit excessive.

Seems like this method of distributing the traffic would make the top of the hour more competitive and the least likely time to get quality traffic for testing.

Has anyone experienced this or have more insight on this topic? Would be nice if there was a different method for throttling traffic, such as only bidding on every 10th bid, or something like that.

I've been thinking of work arounds to distribute the traffic myself at different times, without manually starting and stopping the traffic each time... still thinking though.....

---

Good question, I never paid attention to that myself.

But another tactic to control the voume a bit is to just block several/many high volume placements so that you are left with more medium and lower volume placements.

That way you can decrease the volume a bit and when you succeed with your tests you can still test the excluded placements with your working funnel again.

---

Nice, I like this idea. I hadn't thought of re-including cut placements - it's usually the simplest solutions that works the best

---

This is basically the first step I do in a fresh campaign: Get rid of high volume placements so that the traffic for the test is distributed across more placements.

When 90% or so of the traffic is coming from only few big placements then these placements have way too much impact on the campaign performance.

Then you don´t really see the real average campaign performance but rather the performance from only these big placements.

So get rid of them as soon as possible and then run the tests with traffic that comes from many different placements where each single placement doesn´t have that high impact on the overall performance.

But always keep in your mind, you can only cut placements because of bad quality when you have a working funnel.

So when you exclude placements during the test period you have to test them again when you found a good converting funnel.

---

Hello Jeff, GL with your FA and hope to see some stable profit soon. I am at the beginning of this journey as well. Gonna check on your progress now and then.

Seems that there are hidden golden nugets everywhere so thanks to everyone sharing the knowledge with us!

---

Hey I know I haven't posted in a bit, but wanted to wrap this follow along up because I finally had my first $XXX day this month. Learned a ton with pop traffic, but ultimately applied that to other traffic sources to finally find some green. Took me a long time, but proof that if you stick with it, you'll eventually find your first $XXX day too!

---

Congrats! Those are some very nice green numbers. Those payouts too... $50... $100+?!

What traffic type did you end up on?

---

Very nice and great ROI on the green days, congrats!

I would also love to hear some more details on where you ended up... traffic, offers ... without revealing any sensitive data of course

Congrats again!

---

Thanks for the update, but same as jaybot and matuloo said: would be great to hear a bit more about what you changed and such stuff.

---

Wow @jeffff congratulations!! Always nice to hear an update from members that have disappeared for a while!

Looking for your response to the same questions as well (asked by the others).



Amy

---

Home > > Newbie Follow-Alongs