Introduction: This is an overview of how things happen when you take a creative thats working for a similar offer in another geo. Make your own variation of it, port it to new geos and then you try to run broad 45+, with just one broad interest which is engaged shoppers. Many say oh let the pixel do the magic for you, so I’ve decided to spend $2000 to test this theory. The following raw guide is my notes, observations, stresses and disappointments. NO ONE will tell it how it is; except for me of course. This was a classic failure, failure is what you’ll encounter 90% of the time when you do any type of paid marketing; that’s life. The gurus will always show you this student is killin it a week after taking their course, and that. They will show you one student out of maybe 1000 students. Because they need to sell you courses, I don’t have a course to sell. so Here is the real deal!!!!
PS: In the process, since the results were so damn bad, I’ve decided to find the ideal customer by narrowing my targeting which is also detailed on how we did that in this case study.
I am so overly sick of all gurus teaching how to scale shit that already works, but they don’t help anyone how to start something from zero and get it to the point where you have to scale it because its making you money… so today you are in luck because I am going to document and write a step by step guide on how to promote branded ecom products (these aren’t drop shipped from ali express, nor do they use shopify – 1 lander = 1 product)
Materials & Tools & Traffic Sources We will be using
- The Offers
We are going to be split testing 3 different angles, each one has a unique video ad that was created. Two of my angles were created by my team at Anglesaurus who came up with the angle, and then the videos were done by BannersLanders, and one video was created by a 3rd party service called AdZombies whom I really like.
I didn’t want to come up with the angles myself, so I just told the product to Anglesaurus as well as AdZombies by filling their project brief and let their creative geniuses do the work for me. Why? Because I am trying to distance myself from having to do everything and to a point where I just delegate every single thing — I am getting old and really tired of sitting on the computer all day (weird how things change; considering I was a computer nerd since a young age and that was my life) Anyways..
It took like a week to get the video ads done, while that was happening, I sent in one survey style lander to B&L to create for me – I found this lander while spying on AdPlexity for ecom style offers. It was in English, so what I did was edit the text and make it relevant to my offer, then send in and get it translated to the languages we’re targeting — I use www.transey.com for translations – 50% less expensive, aka cheaper than OHT (one hour translations).
After I have all the materials so:
1 – Video Ads
2 – Landers
3 – Offer(s)
I will go and create a simple campaign in RedTrack where in this case, I am split testing the offer page. One of my offer pages is the direct order page where they select how many items they want at a discount, and enter their details/cc to purchase. The other offer I am testing is linking to the offer page, which tells more about the product, but on this offer page, they need to press the buy it now button to get to the purchase page. You never know which one will convert better so you need to test, and for that we’re using RedTrack.
Here are some screenshots of our Redtrack setup
We made one campaign in our tracker and it looks like this (click to see bigger image)
As you can see, it’s a campaign, with a single landing page, which is an EXAM style landing page that asks the users their age and shit (like those dating landers made out for our ecom product) and then says checking if they qualify for a discount, lalala you can find these while spying with AdPlexity) i had B&L just rip and clean it up and edit it to have our ecom offer basically..
We have two offers added, one of them is direct to the offer lander, and the other is direct to the check out page to buy the product. We were testing just one, because on day 1 we always want to make sure things work, before we get crazy with testing.
—- Results after 24 hours —
I’m no guru spreading happy happy bullshit to grab your money, hell this guide is free, so it’s legit. Bad News: results after 24 hours suck. We’ve spent $300, got 0 conversions. But it’s OK, for now….
The first day after we launch is always designed to check for errors and fix them before we increase the budget. We do high budgets because FB needs 15-25 conversions, so since the payout is $27 for our offer, we need to spend at least 15x that to get enough data so the algo starts working for us. $27 * 15 = $405 at the minimum per adset.
Take a look yourself:
WAIT, DID I READ THAT RIGHT ATTILA!? 4135 link clicks and ZERO sales??? Yes, it’s CRAZY – but again, it happens — what matters is HOW YOU TRY N DEAL WITH IT!
*Note: from years of doing t his stuff I know it has to be some tech related shit with the lander or something, so we troubleshoot.
I like to start out by reading the comments on my ads, especially when they got this much engagement and likes and comments… people love to complain, and by doing so they will tell you the problem… let’s see…
WOW, so they are saying my SSL is giving errors and they cannot order.. shit, this is a major problem.
I fired up my hidemyass proxy, grabbed a local IP, and went through the funnel.
- FB AD => RedTrack URL – CHECK, All Good
- RedTrack URL => Exam Pre-Lander – CHECK, All Good
- Exam Pre-Lander => OFFER PAGE, all good. Weird it works for me, but I detected that the quiz lander, after they press the button to buy, takes 2-3 seconds to actually send them to the offer page.
Was this the problem? NOT SURE.
So what will we do now.
#1) I sent a message to these two people asking them what happened, since I couldn’t re-create their errors.
#2) I hit up my programmer at BannersLanders, to get this fixed so when you click it goes right away. I also got a minor bug fixed, basically the CTA button was confusing and cut off on mobile.
We will have an adset running the old ad, and we will setup a new direct link campaign on redtrack, that will go straight to the order form page so they can buy without going through the exam. This new ad will be more ecomy.
Update 48 hours later
#1 – Heard back and received a screenshot about the SSL error/error the people were commenting, turns out it has to do with cloudflare’s SSL, which occurs sometimes. We’ve sent an email to CF see wtf.
#2 – We had one sale, so right now $25 revenue and around $500 spend all time. Which is HORRIBLE but we were going super broad our audience size was well over 24 million, normally in blackhat I’d give up and try new offers and what not, but since this is our own branded product I can’t do that we’ve got money invested in stock, marketing. etc
Brainstorming, we’ve spent $500 bucks so far, and we’ve got some data. It’s not useless data though, heck I don’t even know who the hell bought the damn thing because the pixel wasn’t placed. I also don’t know how many people who clicked actually bounced on the lander, or went through the survey. To top it off the FB pixel isn’t working in the stats, but it’s working with fb pixel helper and it also shows data is coming in.. so WTF?
To try and fix the above, we’re re positioning the FB pixel as it was outside the <head> tags, and we’re adding in Google Analytics. We’re also going a step further with FB pixel, to log when people go through the survey, and click the last CTA, and then when they arrive on the check out page, and when they actually purchase by using events in the facebook pixel and tying them to button clicks (I wrote a guide about this here before).
OK we have spent $500, we purchased data. We can use this data to generate custom audiences, as well as lookalike audiences… we are going to use custom audiences to retarget, and lookalikes to target more people that watched more of our video.
So this is what we do; we were running a video ad, according to our statistics over 130,000 people watched the video. We create TWO custom audiences, viewed 95% of the video or more, and viewed 75% of the video.
To create these custom audiences, in Business Manager select AUDIENCES, under the ASSETS menu.
Now press the CREATE AUDIENCE blue button and select Custom Audience, here you want to click ENGAGEMENT option.
It will take an hour or two to populate the audiences. After that’s done time to create the Lookalikes. I won’t make screenshots about this because its super easy, just click the check box next to each custom audience, and then click ACTIONS > Create Lookalike.
On that screen, click the show advanced options, and make 3 audiences in one go.
Are you with me so far?
OK, to continue, we need to get some new ad texts created/translated. I try and delegate everything to my team of workers, so I needed some creative texts for the retargeting ad, and I needed that translated. So anglesaurus, and transey handled that (my creative, and my translation service).
In addition to the above, I am going to try and build a ‘passionate’ or warm audience of people that would most likely want to buy my product. Audience insights is a good place to start, but I personally really don’t like it because many of the interests in audience insights you can’t target when you go create an ad. So what I do, is just make a new adset, and then play with my targeting there.
My product is for pain, so since there’s no interest, or fan groups for this on Facebook (i’ve checked) we have to take another approach.
- List of organizations that deal with this pain
- List of meds that treat this pain
- Sources for products related to health/recovery – stores and big popular brands
- Scientific names of the pain
- List of major credit cards/shopping sites to get people that have CC’s.
A few days later update.
I’ve implemented some changed, and had no more sales. It was time to shift perspective and look at things from another angle. We’re down $1200, and we got 1 sale only. Looking through the funnel all looks good, only a few people getting SSL errors due to cloudflare, so that’s like 0.01% of the total visitors that came. A lot of people are commenting where to buy, how much, i want it.. but somehow they aren’t buying, so most def either A) they don’t have money B) they don’t have CC which is a big ass problem. I’ve also made the mistake of putting all the geos I targeted into the same adset, which isn’t a good idea as FB always prefers to pick one and send most of the traffic from there.
Weird observation, I was getting disapproved non-stop because the ad had the word ‘sus’ which is ‘harassment, calling out people’s traits’ according to their stupid system. While I was troubleshooting, Facebook hit my tracker over 1500 times and sent 1500 visitors from USA to the offer/advertiser. At first I was like WTF, but then I remembered why (none of these showed in logs, so for an hour while thinking wtf it could be i was quite baffled if you can imagine.
THIS IS NORMAL, this is what real affiliate marketing looks like – a lot of problem solving (That’s why I loved the book Problem Solving 101, wrote about it here) — GURUS will show you big ass screenshots of success, they will never ever admit its a confusing and hard ass journey to get there – WHY? because they want you to buy their course, and they don’t tell it how it really is. Anyway… game plan change.
- WE ARE GOING TO MAKE THIS WORK AND NOT GIVE UP <– positive attitude, they say its important to have this so we will listen to ‘them gurus’. (Just wait..)
- We are going to be using a new strategy where we will run a separate adset for each country.
- We are going to be doing Website Conversions campaigns, and objective will be ADD TO CART. This event will be triggered when someone goes through the funnel, and clicks the buy it now button on the end to goto the order page.
- We are going to still be using RedTrack because it helped me analyze the data and find the issues I’ve explained above and help come up with this game plan.
- For our adsets, we are going to be using LOOKALIKE audiences of 95% of people that watched the video AND we will be targeting broad – ENGAGED SHOPPERS and using flex audiences where people liked Visa, Mastercard, Amex, Paypal, Credit Cards, Ebay and other top shopping sites where payments by credit cards are practiced.
- For our ads, we are going to be doing 4 ads, duplicates 2 times in each (so total 8 ads in an adset).Ad1 is going to address two problems we’ve learned from the comments 1) People are so stupid they don’t know where to buy even though it says in the description (we are going to put this info into the video ad itself!) 2) People keep asking the price over and over again, even though it says in the text of the ad.Ad1 is going to be stating price, and payment methods.
Ad2 is going to be more click baity I’ve actually found to be doing good for a similar product in English marketers, this ad will tell people they must take the survey to get special deals.
Ad3 is going to be a super simple ad, which is like a question – because in my past of running NUTRA i had insane luck with a simple 1 sentence Question getting a lot of attention, a lot of expert copywriters say the mind is wired to always catch ? marks more so than just sentences, and automatically want to answer them.
Ad4 is going to be like a testimonial style ad.ALL ADS will be using the exact same video ad.
- We are going to be retargeting the people who clicked ADD TO CART but didn’t buy by using a custom audience. Our retargeting message will also test 4 different ads and angles.ad1 -Say goodbye to pain for only $24.95 USD! We accept VISA, Mastercard, American Express and PAYPAL – ORDER AT =>>
ad2 -HURRY only 2 left!! Order now at
ad3 -FREE SHIPPING – Lifetime Guarantee – 24/7 Customer Service – We accept VISA, Mastercard, American Express and PAYPAL – Order Yours Now ==>>
ad4 – Order in BULK for loved ones and save up to 67% OFF!! Get Yours Today at =>>
- We modified our ad to help with the ‘idiots’ who despite the ad having clear LINKS and Call to action buttons still ask how to buy, where to buy, etc and added on the last frame of the video ad a HUGE CTA that says basically click on the SHOP NOW button to take the survey and claim your special discount!
We’ve setup different adsets for each country, and we followed the above system of setting up the ads and what not. In a nutshell, we had 1 adset for each country, there we targeted LLA of peeps that watched 95% or more of videos, and are engaged shoppers and like visa,mastercard,amex,paypal,amazon,ebay, and the top shopping sites in the latam region.
Then we sent it for approval. From 12PM noon, up until 8PM, we kept on editing, and resending and everything just kept getting rejected.
I talked to a facebook rep, and they told me we can’t say that your product can help someone do something, there’s no guarantee and we should focus on the product like its color, size, shape, and other bullshit – basically not to talk about benefits, sell features <——– HUH?? That goes against what they teach in Marketing 101 for fucks sake!!
Anyways, miracously all of the ads got approved at 8PM, ever single one even the ones that were getting rejected during the day. By this time instead of having 8 ads in an adset total, we had like 30, and I had no clue which ad was direct linking to offer, which was going through the landing page, and what have you.
but, here’s the good news, we had a sale!!!!!! So we spent $300, and made $25 revenue. AMAZING isn’t it? If FB wouldn’t have wasted all my time, and I would’ve had my 8 adsets, and everything working as it was meant to work, then we’d know why/how and be able to go in that direction, but since it was all a mess, and on top of the iframed purchase pixel placed via CAKE into the offer landing page didn’t fire. so no idea which ad sent the sale!!
But wait, since of Facebook’s bullshit, when we launched new variations of ads to try and pass the approval stage, we skipped in many the lander, that has the addtocart event firing on the final step (ppl clicking ORDER NOW) so damn, what a mess!!
Disorganization stresses me out so much, it’s not how you are supposed to work – but when facebook denies everything, and you start testing different things fast to see which gets approved things tend to get messy.
Today, I’ve deleted all the ads, and added the best ad into every single adset using the EXISTING POST > specify post ID function that had the most social proof.
I will also setup the retargeting campaigns and if I get to it create a manychat bot because I got a feeling that might work better with these people, they are so dumb (and I am being blunt honest here) even if you write BUY IT HERE>>> link, link, link, and have a major shop now button, and a button that says CLICK THE SHOP NOW BUTTON TO BUY, and the price – they will comment and say HOW CAN I BUY THIS WHEN YOU DON’T TELL US!!!!! I want to buy it where can I?????? << chat bot might solve this, since it tells them to their face!
Just another day in the trenches of real affiliate marketing.
An Update A Day Later
So yesterday was a busy one, we’ve got retargeting up, and tried to address the where can I buy. Sadly people are so dumb, no matter how black and white the buy it now link and shop now button is written everywhere (in the video ad included) they will still ask. So you really need to use a tool like agorapulse to manage the comments in real time. In my case, I got Aubrey from Transey on Skype to help me out, when people would comment I’d write a response and paste it to her to translate to spanish and then she’d send me the response in a couple of minutes to reply to people. To tell them what’s already there, where to buy. This is how it is in ecom, no joke.
I also got a manychat bot going, and had a bitch of a time getting it approved, they are super sensitive to images it turns out, so i had to test like 30 different images in the end to get 3-4 approved. But it was enough to get $200 spend overnight while I slept, to see if people would buy, and no dice.
Funny enough, thanks to many chat a lot of people cut n pasted screenshot why they won’t buy. Aka the shock on their face when they say that product cost $79.80 (even tho that was the price for 5 products in bulk with discount) but shows you yet again, people just don’t fucking read a damn thing the page says.
In this experience, we also had people scream SCAM because they couldn’t click a button (we had one angle that had a dead link on the lander for a few hours), and also scream SCAM if they get an SSL error due to cloudflare acting up 0.1% of the time.
So, as you can see, if the product sucks, or the price is wrong, no matter what you do, you just can’t make it work. We’ve spent $2300 dollars, and had 3 sales, generated around $300 revenue in the backend since some ppl took more some less of the product in bulk.
This is a fail, why it failed? Maybe because the check out page the price was too high and they didnt read the details. We could test another wording on the checkout page, which is more ‘idiot proof’, — or simply put the product sucked. Which in my past ecom experiene is always the case.
Remember our positive attitude? Just because we believed we can make this work, doesn’t mean shit as facts prove it. Because ultimately product and price is what matters the most. That why nutra trials are sold with fake advertorials saying Oprah and Ellen recommends it, or the ROCK used it.. and when they get to price the deal is so cheap they wanna take it for $4.95! But we’re doing WH, so we can’t really make up bullshit lies with celebs because then we’d be back to BH territory and account bans right?
I’ve now had my team prepare a survey that we will re-target to people who visited the checkout page, and ask them why they didn’t buy to get some concrete proof to share with the advertiser.
But in conclusion, based on my years of experience media buying, every time we had a winner, BlackHat or White, if it had potential, it had potential right from the start on small test. This just confirms it, even if you throw too much money at a shit product, or get the pricing wrong no matter how much money you blow on it testing different audiences, angles, ad format types, and strategies, it won’t magically spin around.
UPDATE – 24hrs later
The first survey results are in, we’ve received 66 responses from people who spent more than 2 seconds on the checkout page. (We’ve done this by creating a custom audience in FB audiences, then retargeting them with an ad that says we need your help, we want to understand why you didn’t buy our product. For your help, we have a special gift for you) **the gift is a 50% off coupon FYI…
the survey was made of 2 questions, q1 was why didnt you buy the product, and q2 was a field to provide their email address so we can send em the special gift… (custom audience opportunity yet again)
here’s the results
40.9% of people said I DON’T HAVE A CREDIT CARD OR PAYPAL
42.4% said I don’t know where to buy (blah, people are stupid yes, these guys went to the checkout page and don’t know where to buy, guess they expect a personal call on the phone COD style??)
39.4% said the PRICE WAS TOO MUCH…
at least now we know.. if this was MY product, and i had a shitload to sell in my LATAM warehouse that i don’t want to remain to collect dust there… this is what i’d do to continue the battle towards profits on Facebook…
1. I’d make a new checkout page, that doesn’t scare people off with price, I’d show the huge discounts they get vs the total price for ordering in bulk… and then display how the price per item gets smaller the more they order.
2. I’d add in local payment methods, and even COD.
3. I’d try COD style, where we take just their name & address and have a call center call em up to confirm details, the way Adcombo does it… I ran $1.7 million revenue with them in Asian countries on COD, and gave a speech about how we did it at AWE2017 .. look for it on youtube.