Follow Along with Binom (33)
I'm creating this follow along to both document my progress and help new people out that are using the 40-day tutorial. I am very technologically inclined in this space. A few weeks ago I could barely explain the difference between a server, a domain, and how hosting works. I know practically zero coding, though, have spent a few hours doing the beginner HTML, CSS, and Javascript courses at codeacademy.
I've been following along with this thread https://stmforum.com/forum/showthrea...g-(with-Binom), and I'm going to post updates using the 2018 version since the 2019 is still in progress. However, i'll skip the parts on direct linking as its not that viable of a strategy anymore.
Day 2 Picking Offers: Signed up to nearly every affiliate network I could find. Even today there are still some I need to follow-up with as they haven't approved me and haven't even responded. I suggest signing up to as many as you can and if you get rejected by one still keep note and apply in the future when you're profitable. Picked sweeps offers from several tier 3/4 geos. Logged into a spreadsheet.
Day 3 Set up Tracking: I had no idea how to set up the tracker, fortunately
It actually took me a TON of time just to understand the global postback url, tokens, etc. The way I understand it works is this:
Traffic source sends user to the tracker (along with information regarding city, IP, carrier) --> Tracker assigns a "click id" (individual identifier) to the user --> Tracker sends user to landing page with click id information stored somewhere --> User goes to offer page --> User converts and the click id (a tracker identifier) is sent to the affiliate network --> Affiliate network records a conversion and sends back to the tracker the click id that converted.
Hope this helps people understand it because it took me a long time to conceptualize.
Day 3 Setting up a Campaign: My first campaign I selected "CPM" instead of "SmartCPM" in PropellerAds, and as result I blew through my budget very fast. Make sure you select the "SmartCPM" function. I also made the mistake of trying to scale to other traffic sources to quickly. You will just burn through money. Try one traffic source and nail down the process of picking offers, finding landers, optimizing landers, cutting placements, removing bot traffic, testing bids, before scaling to other traffic sources. I also didn't set up separate traffic campaigns for carrier and wifi traffic, which I think should be done.
Day 21 Hosting and CDN (skipping the direct linking days): I had a lot of trouble with this step as I don't have a background in web hosting/servers, etc. Initially I couldn't get my domain to appear in the URL of my landing pages. They still had the "S3 Amazon" prefix. This may have hurt some of my initial campaigns. I have a more technical friend and he helped me set up a VPS with DigitalOcean and a CDN using Cloudflare. I understand this is not what's included in the tutorial, but things were not working out for me. This setup seems to work very well - loading time of a sample landing is 8ms. I'm sure the recommended way works very well, I just couldn't get it to work for me. If you need help setting up the DigitalOcean --> Cloudflare set me a PM and I'll help you out.
Day 22 Downloading to Landing Pages: I used Adplexity and downloaded 5-10 landers per my GEO and prize. This part is fairly straight forward.
Day 23 Fixing Up Landers: Definitely not my forte as I know barely any coding. I may consider using a full-time person to rip landers, change code, input my click URL, and do other optimization. May post an ad like Amy recommended. I think you can save a ton of time and efficiency by outsourcing this unless you are a proficient coder. This is a key part of the process as landers have HUGE impact on your ROI. If anyone has other recommendations on how to find a quality full-time VA that can do some of this, please let me know.
Day 24 Fixing up Landers: I ripped a lot of landers across several GEOs and keep track of them in a spreadsheet. I describe the lander, the GEO, and its language. This helps me keep organized. This day is a very important and actually may take a day or two. Do not skip this step and make sure you optimize well and all your links work. Even after following the optimization steps in the tutorial (Used kraken, moved JS scripts to the bottom of the index.html, Used google Jquery libraries, moved CSS to index.html) some landers loaded over 2s. I'm not sure why, and am considering cutting them for that reason. Definitely take your time and do this right.
So this is about where I am in the tutorial. I've probably launched 2-3 test campaigns none had much success. I initially didn't separate out WiFi and carrier traffic in the traffic campaigns, which I think now should be done. I also don't think I spent enough time for those campaigns optimizing landers so I went back and optimized landers and are re-running the campaigns.
Here is the process I'm following:
1) Get 3-10 offers from various affiliate networks.
2) Get 5-10 landers that appear to be working well.
3) Fix landers and do optimization steps (Kraken, move JS code to the bottom of the index.html, etc.).
4) Run a sample campaign with all offers and all landers. Make the budget 10x the payout for all offers (Not sure if this is optimal). Make two campaigns in the traffic source that go to the same campaign in your tracker. One for web traffic and one for carrier traffic, with different bids. Both campaign budgets 10x the payout of all offers.
5) Stop the campaign after an offer converts 2x.
6) Run that single offer with all landers (Again at 10x the payout).
7) Cut placements that have over 70% bot traffic.
8) Find the optimal offer/lander combo. Cut unprofitable placements.
9) Test bids on traffic source then if profitable scale to other traffic networks.
I mentioned above that I'm learning the paid traffic with sweeps offers using Pop ads.
I selected around 25 offers across 4 GEOs and originally ripped 40 landers. I narrowed the landers down to about 17 as either I couldn't figure out some of the features of the lander, or the page loaded very slow (over 2 seconds). Before I started my tests I had at least 3 offers per GEO and 5 landers for that particular GEO.
My test run consisted of testing these GEOs with all landers to find an offer that converted 2x. The budget was 10x the payout of all offers. Pausing the campaign if a single offer converted twice, running more traffic and seeing what was the best lander. I created two campaigns in the traffic source (yet only one campaign in the tracker). One traffic segment was carrier and the other wifi. My results are below.
https://imgur.com/a/C3N269Y (Initial stats)
For the top two campaigns, I was able to identify the top lander and will now test all offers. For the 3rd campaign, no offer converted twice so I'm going to send more traffic to see if one converts twice. The last campaign had no conversions at all. I'm going to send more traffic for this campaign and
I've been simultaneously cutting placements that appear to be 70%+ bot traffic using this bot script guide.
https://stmforum.com/forum/showthrea...live-campaigns
Is there a way to double check the bot script is operating correctly? Besides just looking at the zone IDs for the "Parses Javascript" offer?
I find it quite difficult to find enough offers for this segment. I've already looked up around 6-7 affiliate networks and applied for all the sweeps offers for this prize in these GEOs. I'm awaiting approval from a few more networks. Should I start testing random GEOs for the landers I already have?
Hello bravich!
You started a nice follow along, so let me share a few tips and answer some questions.
Traffic source sends user to the tracker (along with information regarding city, IP, carrier) --> Tracker assigns a "click id" (individual identifier) to the user --> Tracker sends user to landing page with click id information stored somewhere --> User goes to offer page --> User converts and the click id (a tracker identifier) is sent to the affiliate network --> Affiliate network records a conversion and sends back to the tracker the click id that converted.
2) Get 5-10 landers that appear to be working well.
3) Fix landers and do optimization steps (Kraken, move JS code to the bottom of the index.html, etc.).
6) Run that single offer with all landers (Again at 10x the payout).
7) Cut placements that have over 70% bot traffic.
8) Find the optimal offer/lander combo. Cut unprofitable placements.
9) Test bids on traffic source then if profitable scale to other traffic networks.
The attachment with your results is not loading, can you please upload it to imgur and then post the share url here? It's a vbulleting bug we don't know how to fix
Cheers,
Matej.
your first two offers looks fine, try to send more traffic to them.
LeadCloak
Hey guys thank you for following along and reading my ramble. Since my post above I did the following:
1) Isolated the top lander for the MY & KE Sweeps campaigns and sent more traffic across all my offers (I think $30 for MY and $15-20 for KE).
2) Sent more traffic to the TH campaign and cut one offer. This campaign only had 3 offers (smallest among my 4 campaigns) as I couldn't find that many across the affiliate networks I'm signed up for. At this point, I think its a combination of poor offers and poor landers, regardless I think I'm going to stop this campaign and offers.
3) For SG Sweeps campaign, I sent more traffic attempting to find a good offer and cut two obviously bad offers. Doesn't look like I got more conversions to identify a decent offer. I'll send another $10-20 in traffic and see what happens.
https://imgur.com/a/NBZns9h
I got a little excited because in the MY sweeps campaign one offer converted 15x and I appeared to have a good lander. I took that single offer and lander and created a new campaign at PopAds. I may have jumped the gun here and didn't get any conversion.
I'm concerned that I may have a ton of bot traffic and am not running the bot script correctly. The LP CTR% rates seem to be in the 45 - 55% range - and excuse me if I'm incorrect - but I believe this means 55% of the visitors "click" the landing page and go all the way through the to the offer. This seems like a lot in my limited understanding of POPs and could potentially be a lot of bots.
I will read the bot script post again, but essentially I created a new offer called Parses Javascript. The "url" of this offer is google.com. In each of my landing pages near the </body> tag I pasted this code:
<script>
u1 = "https://mytrackingdomain.com";
u2 = "click.php?lp=1";
document.write('<iframe src="' + u1 + "/" + u2 + '"></iframe>');
</script>
I also posted this in the head tag of the landers
<style>iframe {
visibility: hidden;
position: absolute;
left: 0; top: 0;
heigh: 0; width: 0;
border: none;
}</style>
As explained in this post https://stmforum.com/forum/showthrea...live-campaigns
I'm not sure it is working because 1) the LP CTR% seems high, 2) it appears there is a lot poor placements in the campaigns, and 3) there is a "cost" for the Parses Javascript offer. Correct me if I am wrong, but the mechanics of the bot script test work like this:
1) The Parses Javascript offer is grouped in the campaign and randomly rotated with all the other offers;
2) If a bot is sent to this offer it can't read the javascript and is routed to google and stays there;
3) If a human is sent to this offer they somehow get sent to my real landers and go through the conversion funnel.
So should there still be a cost associated with the javascript test? Perhaps, I still need to pay for bots that get routed to google?
https://imgur.com/a/iLPKiYp
Today I am going to focus on the current GEOs I have (except TH, which I am discontinuing), try to find more offers, cut obviously bad placements, and send more traffic.
So I was drilling down into the offer that had 15 conversions to verify it was consistent with what was on the affiliate network, however, the affiliate network shows zero conversions. I double checked the offer url and it includes the {clickid} tag, the postback url also includes the {clickid} and {payout} tokens, so not sure what is the discrepancy. Also, double checked that the offer is still active, and it is.
Maybe I need to wait another 24 hours for the conversion to show up in the affiliate network, any thoughts?
Originally Posted by bravich
also I'll be interested to see what happens with your KE camp, that was one of the first GEOs I tried and I killed it because 1) couldn't get enough traffic 2) I needed too high of a CPM bid
Originally Posted by bravich
I won't comment on the other things, since if these leads were reported by some error, the whole data will be skewed and not usable. So talk to your AM first and make sure the postback works fine.
I haven't setup caurmen's script (yet), but for
I'm seeing about 52% bot traffic for SG when I include wifi + carrier.
I ran into some technical roadblocks yesterday, my server that was hosting my
The affiliate network got back to me and indicated that I did have 29 conversion, but they were directed to a "Smartlink offer" because they didn't meet the country/carrier/hardware specifications from the advertiser. The payout for this was much less apparently. I suppose the traffic I was sending wasn't compliant with the advertiser.
I'm still in the learning phase and happy I'm sticking to these lower payout offers while understanding everything. I restarted the MY sweeps campaign with the same offers and landers I had a few days ago. I split the traffic ($25 each) to segment wifi and carrier - see the screenshot below. Not great results. I'm going to go back to the drawing board a pick a completely different sweeps prize and geo. I'm just going to focus on one geo for the new prize. I don't understand why the costs in the Binom tracker don't reflect the $25 spent for each campaign. I imported the traffic source from template in Binom (to get the cost token) and selected "auto" costs when creating the campaign. Am I missing something?
https://imgur.com/a/JEH9wfv (Binom stats)
The traffic source and tracker also appear to have different impressions/clicks. Is this due to bots?
https://imgur.com/a/ghaMad8 (Propeller stats)
I also restarted the KE sweeps campaign and seem to have better results but appears I need to send it more traffic. I set the budget to $20 so I'll let it run then analyze the results.
I've been reading a lot of the forum and there's so much useful information. Was inspired by this thread:
https://stmforum.com/forum/showthrea...low-my-journey!
I'm continuing to test offers and geos. Here are my results from today
Looks like my first profitable campaign. I'm going to discontinue the JO geo and try to blacklist some obvious placements from the other geos before I send more traffic to it.
I'm a little confused if
Also, they may be the same thing, but can I find bot placements by doing offer drilldown then placement and cut the LP CTR % less than 30%, or do I need to drilldown by placement then see the LP CTR% for the parses javascript offer. Hope that makes sense. Below is a screenshot of the former approach.
Do you remove the parses javascript offer from the campaign after you have cut bot placements, or keep it in. Binom indicates costs are incurred for these visits, just want to make sure I'm not wasting traffic funds unnecessarily.
I'm still struggling with the Parses Javascript test (https://stmforum.com/forum/showthrea...ive-Campaigns/) and did more reading around the forum.
I read what mr black said above, and also read this thread (https://stmforum.com/forum/showthrea...on-using-Binom) and it appears that using the Parses Javascript bot test may not work 100% optimally on
I still would like to confirm my understanding of the Parses Javascript bot test:
1) You need to specify 2 offers when you create the lander, and you can't edit this after you create the lander if you only select 1 offer upon creation;
2) You have to specify the offer links in the landing pages, and these links are different for every offer. So unless you want to change the code many many times, it is easiest to test it with one offer, cut bot placements, then run a landers with no bot script test.
3) It is meant to test with 1 lander and 1 offer + the bot script test (1 lander --> offer a (real offer) --> offer ber (bot script test). If you want to test multiple landers and offers at the same time you need to set up paths. I think grasping this concept is what really confused me, as I haven't set up paths yet. This post helped me understand this better. https://stmforum.com/forum/showthrea...l=1#post345535
Please anyone chime in if my understanding of the bot script test above is incorrect. But moving forward since I am using Binom I am going to use their bot test method. Here are their instructions to set it up. https://docs.binom.org/events.php#p1
I re-ran three geos today with the Binom script, and I'm even more confused, almost totally lost. Two of the geos had very bad results so I stopped them. One geo (top one in the screenshot below) had some conversions, so after I spent $15 testing the script, I cut placements I thought were bots and will test another $15. Here are the results so far, I'll leave it running overnight.
I thought I had my first overall profitable day, but then I realized the costs from PropellerAds in
When I created the PropellerAds traffic source in Binom I did "Import from template", and it populated the PropellerAds tokens. Not sure why the costs are not correct. I certain I sent $15 in traffic for the first row in the screenshot above.
The Binom bot test seems to be working and now I am cutting placements and testing some bids to improve ROI. Also, I was thinking about the parses javascript bot test more, and I now believe that you can use it with Binom. But you need to set up paths, and make sure you use a multi-click URL in the lander you are running the bot test on. This page from the Binom documentation shows how to create a multi-click URL vs. the default click url in the Binom settings page.
https://docs.binom.org/start.php#p3 (point 10.)
Update for the day:
I think I determined an "optimal enough" approach to incorporate the bot test. I was reading around the forum and I see Amy has corroborated this approach a few times.
1) Run 1 AM recommended offer with top lander on adplexity $10 budget (include bot script test). 60 impressions is enough to determine if placements have bots. Can stop the campaign when enough placements have 60 impressions if you don't want to spend the whole budget.
2) Run all offers with all landers (optimized for speed) with no bot script test
3) Find a decent offer
4) Use that offer to find an optimal lander. But include the bot script test for these landers. In order to do this with
5) Find optimal offer/lander combination and cut bot placements from previous step
6) Test bids and cut unprofitable placements (2x payout appears to be a recommended rule of thumb)
7) Optimize bids and test previously blacklisted placements to see if they convert
8) Scale to other traffic sources if profitable.
I'm sure
For the campaign I was running the last few days, it appears I may have cut too many placements, and now the traffic is coming in at a trickle. I have a better idea of the process I am going to use going forward (identified above), so tomorrow I am going to test offers in 2 new geos.
Today I'm continuing to optimize my process and what I've learned is that in AM you definitely can't skip any steps and rush anything. Small shortcuts compound over time and I think this can affect overall success rate. So I'm going through all my landing pages again and make sure they are completely optimized and functioning correctly.
A few additional tips I think can potentially help speed up your landing pages:
1) Set your cache refresh on the CDN or server to 8 days or more. This should allow more browser caching which should speed up times
2) Minify CSS/HTML/JS code. You can use this tool. http://www.minifycode.com Even though I am not a coder, I noticed my landing pages have a lot of javascript that isn't even doing anything, or at least nothing material. You can remove this to speed up processing time.
I still sent some traffic to my Malaysian geo campaign, which had my first overall profitable day in the screenshot above a few posts back. However, unfortunately I think I cut too many placements and the traffic is coming in at a trickle even though I am bidding above the average according to PropellerAds. Only 1 conversion today for that campaign. I'll up the bid tomorrow and see if I can get more traffic.
I started targeting the ZA geo and even though it was warned in other follow-alongs that this geo takes a lot of costs to cut placements I went forward anyway. Overall for the day got 200 conversions, which is by far the most I've gotten so far, however this geo is quite expensive and I was paying quite a bit for the traffic. In addition, the best offer I identified had a small payout of $.22. Clearly, not a good strategy so despite 200 conversions I had a -70% ROI. It was fun despite losing the money to see the conversions, so I kept sending traffic.
Also, I noticed that almost all the conversions were from 3 placements, so I created a new whitelist campaign for just those placements (separate campaigns for wifi and mobile) and saw the ROI increase slightly. Still way negative for the day.
Still having issues with costs not being captured properly in the tracker. I know exactly how much traffic in $ I am sending. I've resorted to manually updating the costs in
Hey guys another update. I know I haven't posted in a few days but I've still been working at it everyday. I've decided to switch prizes I think the the iPhones sweeps offers are quite saturated so I'm trying something new. Another thing I think is really impacting my profitability is the payout of these iPhone sweeps or similar offers. They are quite low while I'm paying anywhere from $1 to $2 CPM for traffic, so clearly I am paying too much for traffic, relative to the offers, or you can think of it that my offers aren't paying out enough.
I spent a lot of time revamping my landers, creating new ones with little twists, and just making sure things are as optimized as possible. So for this new geo I ran last night here are my stats:
I ran $10 for the initial bot test and cut placements that were 70% or more in bots. Then used 4 landers to test 2 offers (most that I had approved at the time). 1 offer converted twice so I paused the other offer and continued testing landers. At the end of the test 1 lander converted 4x while the rest 1 or zero. These offers payout $.75 and I'm paying around the same CPM, so I just need 1 conversion each 1000 impressions to be profitable.
I've spent $35 so far, while cutting unprofitable segments (1000 impression no conversions, 450 impressions no LP clicks, etc), cutting browsers, targeting connection types, device types.
Any tips would be helpful. I am waiting for the other offers in this geo to be approved, but when they are I will test 3-4 offers with the single lander I found and do 10x payout of the offers.
Originally Posted by bravich
With DigitalOcean and any self-hosted tracker try to aim high with memory/cpu (as you would with any other software). I'm assuming they set you up on Ubuntu. If you feel comfortable, simply ssh into your instance and type 'apt-get install htop' then just type 'htop' at the command prompt you'll see a user-friendly graphic display of cpu and memory usage. It shouldn't be above 50% at your peak traffic times in my opinion (you want to leave room for periodic background tasks, etc). htop screenshot: http://hisham.hm/htop/htop-2.0.png
Hey Kintura thanks for the feed back and reading along. I appreciate the additional info regarding the server, since I upgraded the ram/cpu its been running smooth no problems. If I run into any problems in the future I'll take a look at the usage based on your instructions.
Yesterday, I mentioned I was going to cut additional placements, browsers, connection types, etc., anything I thought was negatively impacting ROI. However, for both of the campaigns I running this resulted in almost no traffic coming in overnight or during the day. A similar scenario happened with one of my iPhone sweeps campaigns a few weeks back.
At this point, I have removed all previously cut placements, except the ones identified in my bot script test, and set the budget to 10x the payout of all the offers in aggregate. I'm going to let it run more and collect more data before cutting placements. After this additional round of traffic I'm going to use the placement killer xlsx and see if I can cut some unprofitable placements without completely losing all the traffic.
Hey @bravich, I'm wondering how you came to the conclusion that iphone offers are saturated?
Not disagreeing with you or anything, but curious if you have data to support that thought?
edit: I chatted with my AM and she mentioned she thinks only phone prize sweeps offers convert well with pops. I'm going to do some checks on Adplexity tonight to see if that's true
Nope you’re right no substantive data, just feels to me quite saturated. Just anecdotal data, seems like everyone is using the same landers and process, maybe it just means a few select people are getting most of the conversions.
gotcha. my AM told me that for voucher prizes almost everyone runs FB (I'm guessing blackhat?), so just a heads up on that.
will update here if I see anything interesting on Adplexity in any case..
Hey guys another update.
I've been running a campaign in SG, but my traffic was coming in at a trickle even though I was bidding around $1.60. I think someone was running something quite profitable and was bidding $3 CPM, looks like he was getting all the traffic. It just doesn't make sense for me to continue in this geo for these offers so I'm going to move elsewhere (paying too much for traffic with too little conversion/payouts).
Another international geo had a very, very low CTR (under 1%). I think it had to do with improper translations. I used google translate to do everything myself but there must have been some obvious errors. I hired a translator from the geo to do the translations for me and will re-launch the campaign with those revised landers.
A question I have is around bots, is it best-practice to run a bot test once and maybe another time later in the campaign, or should you structure the flow to have bots tested continuosly?
Thanks.
The translated landers had a positive impact and improved the ROI a lot. I got some conversions and had a positive ROI for a few hours. Here are the final stats. Unfortunately, traffic was coming in quite slow again, even though I am bidding near the max rate in Propeller.
The biggest issue I am having now is optimizing the campaigns correctly. I am using the STM kill placement calculator as well as referencing this thread
https://stmforum.com/forum/showthrea...-ACTION-Part-2
Here are some of the optimization rules I've been employing:
1) Cut placements with 30% CTR to the parses javascript offer (bots)
2) If >= 60 LP clicks and no conversions, cut.
3) 2x avg. offer payout spent and no conversions, cut.
4) Over 500 impressions and LP CTR is under 1% with no conversions, cut.
5) Cut poor performing browsers, devices, and connection type.
Any feedback on the rules I am employing above or other insights on how I should be optimizing/cutting placements?
I started a new campaign yesterday in a geo I haven't tried before. I was pretty meticulous creating the landers in the local language, checking all buttons, optimizing speed, clicking through to offers, etc. Spent about $50 with only 1 conversion, so was quite disappointing. I have three other offers for this geo so I'm going to test those before hanging up the hat.
It feels like I'm much more efficient and understand the process much greater now than the beginning, yet the results are not indicating that. Either way I'll keep plugging away and troubleshooting the errors.
I've been more diligent checking the offers before running/starting campaigns. I'm using Nord VPN and go to the geo of the offer and try to click through to it. Either I get an error/blocked or it goes to a weird random offer (antivirus, VPN offer, adult) type pages.
Am I doing something wrong, how do you guys check the offer before running traffic?
I'm running into the same issue with all my campaigns: I end up with little or no monentizable traffic after optimizing.
Typically I'll run a budget of $ = (# of lander x # of offers x avg. offer payout) x 5. After a few conversions, I'll isolate the best lander then the best offer. When conversions come in I'll blacklist browsers, OS versions, connection types, devices, etc. At the same time I'll be cutting placements that are 1x avg payout with no conversions and ~2.5x avg payout and still unprofitable (if they had previous conversion).
I've been trying Sweden for my latest campaigns and was testing out both Propeller and PopAds. Propeller was not good at all, PopAds had some promise, but when I did all the above I end up with very little traffic coming in - maybe 4k impressions per day.
Any insight on why the traffic may be getting strangled? I believe I'm following all the recommended placement cutting strategies. I've even tried whitelisting placements that previously converted (set up two campaigns at traffic source: (1) previously converted placements blacklisted, (2) only previously converted placements whitelisted.)
Just replied to your PM.
Once again - many apologies for my extremely late arrival. Don't know how I've managed to miss this thread for so long. I think I noticed it last month and had a tab opened to this thread but somehow forgot to reply.
But apologies won't help - will do my best to give advice instead.
Typically I'll run a budget of $ = (# of lander x # of offers x avg. offer payout) x 5. After a few conversions, I'll isolate the best lander then the best offer. When conversions come in I'll blacklist browsers, OS versions, connection types, devices, etc. At the same time I'll be cutting placements that are 1x avg payout with no conversions and ~2.5x avg payout and still unprofitable (if they had previous conversion).
I've been trying Sweden for my latest campaigns and was testing out both Propeller and PopAds. Propeller was not good at all, PopAds had some promise, but when I did all the above I end up with very little traffic coming in - maybe 4k impressions per day.
Any insight on why the traffic may be getting strangled? I believe I'm following all the recommended placement cutting strategies. I've even tried whitelisting placements that previously converted (set up two campaigns at traffic source: (1) previously converted placements blacklisted, (2) only previously converted placements whitelisted.)
My advice would be to focus on testing offers and landers - with special emphasis on OFFERS.
The less traffic volume is available for your geo (and any other targeting), the higher ROI your offer+lander combo will need to be at before you even focus on cutting placements etc.
But even for bigger geos, focusing on finding a good lander+offer combo FIRST is important - in fact, it's even MORE important - because big geos require spending a ton of money to cut placements. So we better make damn sure we're using a lander+offer combo that converts well, before investing into cutting placements. Otherwise it could get very expensive.
In short, focus on testing offers and landers FIRST. Only when you see a promising ROI, would you want to focus on cutting to get you through that last stretch to get green.
You may be wondering "what would you consider to be a promising ROI?" That would depend on the geo. The smaller the geo is (plus whatever additional targeting, such as carrier), the higher the ROI needs to be before you do any heavy cutting.
Another key: Look for profitable traffic segments. If the profitable segments aren't capable of giving you daily profits of say at least $5-10, don't even continue cutting.
Otherwise, it would be like trying to blast through rocks where there are no diamonds in the first place.
Make sure there's a profitable OS (or close to breaking even at least), or a green carrier, or a group of medium-large sized placements that are green - BEFORE focusing on cutting the unprofitable.
If there aren't profitable traffic segments, then no amount of cutting will give you profits.
Hope that's clear! It took me a LONG time before I realized the importance of looking for profits first, before focusing on cutting. Understanding this will save you time too.
Looking at your placements stats above for Propellerads: Almost all the placements that have made conversions, are red. That's not a positive sign, especially for a smaller geo. (For bigger geos, if you run more traffic, some of the medium-sized placements can end up profitable enough to jointly give you a good daily profit.)
The popads camp does look promising! Could you please post some placement stats and other stats (OS, browser, etc.)
Also: For smaller geos, cutting a placement at 1x payout may be too strict. Perhaps resume some of the placements that got paused at 1x payout and run it to 2x payout.
Another thing to do - as mentioned above, find more sweeps offers to test for this geo. Go to Adplexity to check if sweeps offers are even doing much volume first - if another vertical is hot for this geo, can give that a shot.
Also: How many landers did you test for the current campaign? Are they optimized?
Let's start here - please don't worry about posting too much stats or information. Now that I'm subscribe to this thread, I'll be alerted to all subsequent posts - will reply as often as I'm able.
Hang in there! You're doing well - profits aren't far off!
Amy
Also worth mentioning: All your other campaigns that you've launched in other geos are NOT wasted!
You've cut some bad placements already. Next time you launch a campaign there you won't be starting from scratch - you can start with those placements blacklisted from the beginning.
Once your campaign is profitable, you can retest the previously-blacklisted placements in small batches to give them another chance if you wish - in a way that won't cut into your profits too much.
Something else I forgot to mention in the last post is the importance of testing different bids. But let's take a look at your placements stats for the popads camp first. Sounds like a good camp to test bids on.
Amy
Thanks for the detailed response, I appreciate the feedback.
Originally Posted by vortex
This may be a simple question, but how do you determine if a geo is "small" or "large", is it simply looking at the traffic estimates without anything cut at the traffic source?
I actually paused the SE campaign altogether and moved to Germany, but after reading your post I decided to send a bit more traffic to it and actually un-blacklist all placements. They are mainly red, but perhaps I need to send more traffic to accurately assess performance of the placements. And use a more lenient cutting strategy (if no conversions and 2x avg. payout = cut, if conversions but 4x avg payout and still not profitable then cut). This may give me some more traffic back.
I still cut some browsers and OS versions, but this shouldn't eliminate too much of the traffic (screenshots below).
I initially tested 4 landers for this campaign. I find that sweeps iphone/samsung landing pages are really one of a few varieties (spin, survey, open boxes, "you won claim now" type pages). Early on in my testing I was doing maybe 7-10 landers, but I think I used the types of landers (4 total) that appear to be performing the best accordingly to adplexity. I optimize images, code, etc. and test the speed all are under 1s load time according to GTmetrix, so I think they are optimized enough.
Was having some issues with the Germany campaign as well, had some initial conversions, but then when I isolated landers and sent more traffic I got little or no conversions. I just sent more today so will see how those results look. I'm also going to start targeting only 6pm - 1am in the local GEO time in order to try and get a higher ROI for these campaigns moving forward.
So I mentioned above that I was going to give the SE PopAds campaign a bit more traffic after removing previously blacklisted placements (I think I was too aggressive in the cutting). The only thing I was excluding were clearly unprofitable browsers/os versions shown in the screenshot above. I added another $20 in traffic. It took 2 days to complete, as I was dayparting (6pm - 12am in local geo time).
There were 5 conversions, and 4 from the second day, which actually was a little bit in profit. But my biggest concern is that there still is not enough traffic coming in. I was a little puzzled, as I'm barely excluding anything from the traffic source. I went back to complete baseline (nothing excluding at all), and the daily estimated traffic was only 11k. With the browsers/os versions above and desktop devices excluded I ended up with an estimate of 3k per day. This is before cutting placements, that I think should be cut in the screenshot below.
I believe I have a good lander and a decent enough offer, so I'm going to try on another traffic source.
Originally Posted by bravich
Yes. I use a the "User agent switcher" chrome extension to act as a proxy for a phone. Not sure if that's the exact name, but google user agent switcher for chrome and should give some options to choose. I also set the VPN to the location. This makes the computer appear as a phone (since I'm mainly targeting mobile devices) and in the correct geo for the offer. If it still doesn't work I don't run traffic to the offer.
Home > POP / PPV / Redirect > Follow-along Campaigns