A newbie on Facebook! (58)
Hi all!
Follow me on my exciting journey to a (hopefully) first profitable campaign on FB. I've chosen an offer that is slightly different than your typical dating or adult offer; I will be trying to promote a golf-related offer that appears to be doing a decent EPC (0.70 on a payout of 9.95) on my network of choice. I will probably be leaking many details which will make it easy to replicate my work, so if this ever gets to be profitable, consider the archived case study my gift back to this generous community.
Activities and planning
- Stake out my demographic: I'm targeting people in the US who like certain golf stores and golfing brands. I've chosen this over the simple 'men >35 who like golf' because it targets people who are far more likely to be active golfers themselves and have actually financially invested in the sport. This is important, because my offer asks for $29.95, so golf spectators (magazine readers, are rather useless to me. I still get an audience of 1 mil. Not bad.
- Find images for my ads: I found about 10 images that I will run against eachother. Some of them are cartoon-ish, others are photos, others are humour, etc. I will be using the advice I found in the FB forum to create a small separate campaign for each of them with four exact duplicates. This way all my ads get an equal chance to 'catch fire' without FB messing up through preferential delivery.
- Image split testing plan: I will split test to find the best image first. After two days, anything with a CTR below half the top performer will go. After that, I will kill off the bottom performer every day until I'm left with one top image. I will spin this image into ten variations (brightness, borders, sharpness, mirror, upside down, etc etc) and repeat this step until I'm left with a top performing image.
- Ad copy split testing: Once the top image is found, I will create five headlines (a question, a shout, a joke, etc). These will run for a day two before I'll start pruning. The top headline stays and gets split-tested with five descriptions. After a few rounds of pruning I will hopefully end up with the ideal ad. Target CTR: 0.2%. I will share my darling ad.
- Landing pages: I've built two landing page variations to split-test. They are basically a blurred out screenshot of the offer's landing page with a fake lightbox on top with some extra qualifiers. The entire lightbox is a clickable image that takes the visitor to the offer. I will share my landing pages in due course.
- Get budgets in order: I will spend about $20 a day to gather data. That's $2 per ad group per day until the first prune. Budget from paused ads will be added to top performers to give them more run-time.
- Bidding strategy: I will bid on optimized CPM. Target conversion value is 9.95 (the offer payout). Target CPC = 0.70 (the offer EPC). This should under perfect circumstances get me to break even. This is obviously not going to happen, so I'm going to take a loss. Such is life.
Strengths
- Me: I'm incredibly driven and have a slight OCD personality that can help me get things organised and done. I'm good with numbers and have a good understanding of the digital landscape. I know what's out there in terms of digital marketing.
- The offer: the offer is a tangible product that provides a tangible benefit at a modest price of only 30 bucks. It's not some shoddy 'fuck your neighbour' dating site that ends up draining your customers wallet more than anything else
Challenges
- Hosting location: Australia is a seriously cool country and moving here was the best choice I've ever made, but you can't have your servers sitting here if you want to mine the vast riches on offer in the US of A. I'm currently running prosper202 and my landers on a shared hosting plan in AU, which obviously needs to be resolved ASAP.
- Physical location: timezone differences are quite noticeable; support inquiries take long to resolve and ads can sit in approval queues for an eternity. Reporting timezones can get messed up, especially with Facebook's inflexibily in regards to changing your ad account's location and currency. Aussies working on POF can attest to this. A currently weak Aussie dollar also doesn't particularly help me whilst I'm still investing.
Questions
- Hosting: who can point me to a good entry-level VPS deal in the states? Hostwinds looks pretty affordable ($13.95/mo for the entry level package and they are prosper202 preferred). Any tips, pointers or vouchers I can use?
Anyways, enough talking. Let's DO this! Wish me luck
Hi Jdenhaan (Dutch by any chance? 
),
Good luck! When setting up the campaign on FB why not split it into smaller campaigns right away? I've found this works best to quickly eliminate groups which aren't working.
So you could make a campaigns like - Male 35-39 who like golf, Male 40-45 who like golf etc.
Just something I've found to work on FB.
Cheers,
Stealth
Yeah I'm Dutch
I already have 10 campaigns to split test 10 images, which until FB get ad groups is the only way to reliably do this. Adding demographic split testing at this point would force me to dilute my limited daily budget across at least 20 campaigns. It would take forever to get meaningful data like that.
So image first, then ad copy, then demographic. Makes sense IMHO, because testing demos without a sufficiently effective ad to test seems bothersome.
Multi variate testing is not particularly easy either from an analysis point of view. If prefer to just split test for now.
try finding a list somewhere in a directory of golf courses and driving ranges.. see if they have a page, or have likes.. to target.
You are right. The image has the most significant effect on the CTR, then the headline. I would focus on those 2. The ad copy is not that important, you have enough room to shape the user's perception on your LP if you use one. In ieder geval veel succes!!
Originally Posted by jdenhaan
I already have 10 campaigns to split test 10 images, which until FB get ad groups is the only way to reliably do this. Adding demographic split testing at this point would force me to dilute my limited daily budget across at least 20 campaigns. It would take forever to get meaningful data like that.
So image first, then ad copy, then demographic. Makes sense IMHO, because testing demos without a sufficiently effective ad to test seems bothersome.
Multi variate testing is not particularly easy either from an analysis point of view. If prefer to just split test for now.
By the way, I'm Dutch too, maybe we can stay in touch?
Originally Posted by stealth17
The only way to give each ad a fair run is through putting them in a separate campaign.
Sure we can stay in touch, but please note that even though I'm Dutch, I do live in Australia
Question
Hosting: who can point me to a good entry-level VPS deal in the states? Hostwinds looks pretty affordable ($13.95/mo for the entry level package and they are prosper202 preferred). Any tips, pointers or vouchers I can use?
do you need managed or not?
digitalocean is good but unmanaged... or ramnode
Originally Posted by dynamicsoul
digitalocean then.. $10 a month SSD
Just signed up to digital ocean. As if they're not cheap enough, I found an $10 coupon. Boom! Setting up a LAMP stack with ispconfig will take a bit of time as my Linux is a bit rusty, but I'll get it done before tomorrow, when I want to move the whole thing over.
I bought com-jdh.com for some nice looking sub domain landing page URLs. More to follow.
Goals for today: set up server, analyse first batch of Facebook ad results
I just had this random thought that's sort of throwing me off balance:
Is it a good idea at all to promote a CPS (cost per sale) offer on Facebook? People on FB are not in a buying mood, so it seems more like a CPL (cost per lead) channel. Am I making a rookie mistake by promoting something on the wrong channel? Note that I am targeting people who like certain golf stores, so these are people who've spent money on their sport before.
It might work, but I've found that FB traffic is being interrupted by the ads, in the sense that they're not going to buy big stuff on a hunch. Maybe adding landingpage with optin to give free tips and THEN sell the product works better. But I'm not sure if FB still likes optins.
there's a boat load of tutorials on digitalocean that walk you through typical set ups.. and if that weren't enough, there is a guy on fiverr who will set up whatever for you also too haha .. time saver..
Originally Posted by dynamicsoul
Little update: I've successfully configured my VPS with CPVlab and have setup the campaign. This obviously took a fair amount of time to do, so my writeup on the stats will have to wait (its 1AM here).
Good news is, that stats are coming in and everything seems to be tracking sweetly. Some ads are clearly outperforming others, which is interesting and good to see. No conversions yet unfortunately. Tomorrow I'll start pruning the ads.
Originally Posted by jdenhaan
Stats time!
We're two days and a bit into data collection now.
Clicks: 126
Average landing page CTR: 35%
Conversions: 0 
As said before, to give all ads a fair run I've created them all in a separate campaign, each with a $2 daily budget. Interestingly, one of my favourite (for it being different) images turns out to be performing poorest of all, prompting me to pull the trigger on it. The bottom 3 (IMG06,IMG01,IMG10) ads are going on pause and the budget will be redistributed amongst the other ads. The top performer will be doubled up and four other performers will all get an extra dollar a day.
Landing pages
My backend stats are a bit messed up still and you will find that the totals won't match up with Facebook, but this is because I transitioned over from Prosper202 on one server to CPVlab on the other. From today onwards everything should start to marry up though, so I'm not overly concerned about the hiatus.
It's obviously too early to draw any conclusions, so both landers will stay on for now. I'm thinking of running ~100 clicks to each landing page and then making a move towards customising / further split-testing the winner.
If you're promoting a $29 product, 126 clicks is nothing. You need to have a few thousand clicks between different landers and ads before you have any meaningful information. If you spend $100 and haven't made a sale, drop/change the product. The reason why people start off with CPA is that you need less data. CPS/trials etc. (anything requiring financial commitment) requires much more data.
I'm afraid I can't get that many clicks for $100. Even my top performing ad is doing about $0.19 CPC at the moment. Or do you mean thousands of clicks for the lander data to be meaningful? I reckon I can at least start culling ads based on CTR/CPC at this stage?
The reason why I wanted to try this offer, is because a lot of the other niches (games, dating, etc) have already been done to death. I wanted to see if I could make something else work. Something a bit more original and fun to promote. Obviously that doesn't mean I'm going to ignore the numbers: If I don't get any conversions after putting in about $100, I will move on to greener pastures and try to give a unique spin to a gaming offer / email capture of some sorts.
I have the feeling that it's very hard to get Facebook traffic to spend any money on the spot.
There is a reason why those other niches are "done to death".. because they work for the market you are advertising to.. IE FB users.. nothing is "done to death" if you find a new unique angle.. that could put the fun into this for you
Originally Posted by dynamicsoul
I will give this a fair go, but you're right in that I shouldn't bet on a horse with three legs.At least this whole exercise is teaching me valuable lessons around ad optimisation and what to promote where, so the money won't be completely wasted if it doesn't work
Subscribed, thanks, what you've already noted is helpful.
I'm starting from scratch with paid models, so it's great to follow along,.
Okay, it's time to follow Caurmen's advice from this thread.
The actual offer link doesnt work because of CPVlab's click detection.I have a few ideas in the pipeline with different offers, one of which I could potentially push live over the weekend. Stay tuned!
Ps. can an admin please change the title of this thread to 'a newbie on Facebook'? I'm afraid my golfing plans got cut short by disappointing results
Done! Looking forward to your next steps.
Originally Posted by caurmen
When you made your recommendation, did you mean sending 4x worth of clicks to the landing pages or having these clicks actually land on the offer page?
Alright! Here we go! I've set up a second offer, or rather a split-test of two very similar offers whereby the user gets promised a free iPhone5 after signup up with their email and doing a few things. Here's my approach:
- I created a fully functional polling landing page related to the superbowl. 5 seconds after voting, the user gets taken to the offer
- I set up my Facebook demographic to target people liking a specific team (should be easy to guess which one)
- I'm bidding fairly low, since the offer payout is only $1.60, but my lander is pretty nice, so I hope I'll do well
- I created 5 campaigns, each with one image
I mean having spent 4x the offer payout on traffic, period. You'll want to adjust that if your landing page is really terrible, but honestly I think it's unlikely your LP is that bad. (You did check that you can click through to the offer from your LP and there are no broken links, I assume!)
New campaign looks interesting - keep us updated!
My LP was decent. It did a 40% CTR, although I wonder if it maybe did a poor job at qualifying the traffic. To check out my landers: lander 1 and lander 2. Note that the offer link will not work because of CPVlabs click detection.
My new campaign will indeed be interesting, because it all hinges on cheap traffic. The payout is only 1.60, so will have to see how that goes. I had to waste a bit of money on poor CPM traffic just to get my ads going (FBs algorithm is absolute voodoo!), but I have now altered my bids to optimized CPM to focus more on clicks and conversions. That should see the CTR pick up soon, hopefully.
Is the rule of thumb 4x the payout for the entire campaign....or 4x the payout for each ad within the campaign? $40 to test the entire campaign seems really low to me.
Originally Posted by jdenhaan
Originally Posted by cmdeal
Thanks for your feedback and moral guidance cmdeal; I do like to stay on the good side of karma if possible. AM is an industry where you'll find yourself balancing on the fine line of conscience quite often.
I have slightly tweaked my ad copy and lander copy to be a bit more honest without ruining the intent piece of it all. My ads have been practically invisible yet because my bid was too low. I've jacked up the bids a air bit. I'll probably lose money on it this way since the offer payout is so low, but I don't mind. I want to get my stuff showing and start collecting data.
I'm quite excited about this idea, because I can scale it horizontally in so many ways. There's so many teams, celebrities, interests, etc. out there that I can engage people on... I just hope the low offer payout won't burn me; if it does, it might be wise to trial this on PPV instead.
Email submit on facebook ? no no
Yeah, I've been looking into prepop offers to capture e-mail addresses AND forward the user. Double whammy
First stats are in. My ad performance looks promising, I'm getting some really cheap clicks on my top performers. I think I sort of understand how to get people to click Facebook ads; now to get them to convert!!
No conversions yet, but I'm adamant to run this for at least a couple of days before I decide on anything. I've put a fair bit of work in the lander and I believe this can work.
Action: I'm switching off the bottom three ads and redistributing budget to my better performing ones.
Action: add a Google remarketing tag to my lander to capture these people into an audience I can bombard later through the GDN
Originally Posted by hd2010
If anyone has managed to get a paid offer on Facebook profitable, I'd love to hear about it.
Originally Posted by jdenhaan
First stats are in. My ad performance looks promising, I'm getting some really cheap clicks on my top performers. I think I sort of understand how to get people to click Facebook ads; now to get them to convert!!
No conversions yet, but I'm adamant to run this for at least a couple of days before I decide on anything. I've put a fair bit of work in the lander and I believe this can work.
Action: I'm switching off the bottom three ads and redistributing budget to my better performing ones.
Action: add a Google remarketing tag to my lander to capture these people into an audience I can bombard later through the GDN
Okay, Internet marketing is a matter of solving each problem one at a time. You are able to get people to at least click on your ad at a decent price. Now you need to solve the next problem of getting people to convert. You will need about a 10% CVR for you to be comfortably in the money. So this will be all about landing page optimisation.
Hmm.. People are definitely engaging on my LP. I'm getting quite a few poll results come through, so that's good. Looking at my CPVlab stats, the lander sends roughly 50% of the people through. Not bad. It's just that people are not filling in the email submit!
Here's my lander; happy to take on board any criticism.
http://seahawksurvey.com-jdh.com/
Note that the redirect through to the offer won't work because of CPVlab click detection (ie it will only work when you use the campaign URL vs the direct lander URL). Note that the offer allows pre-pop. Maybe I should capture the mail addy on-page, store it in a database and then proceed to pre-pop the offer?
design it to not allow the vote unless they fill in email.
Great idea for a split-test.
Speaking of split testing, I just added a second and third offer to rotation. Will try to split test networks as well.
I'd make sure that all of the landing page content is above the fold. You can probably increase the width of that main body div too.
Originally Posted by jdenhaan
Conversions!!
It looks like something's wrong on my CPVlab tracking setup. My aff network is actually showing four conversions today, but because my tracking is somehow botched I can't really track my ROI and I have no idea which ads triggered it. I might have to manually import the subids.
Obviously this is a major priority and I'm looking into this right now. Maybe the tracking pixel I put in Peerfly is wrong.
UPDATE: looks I used the wrong pixel URL indeed. My setup was to use http://clk.com-jdh.com/adclick.php?c...%&sid=%subid1%
Turns out the subid parameter is actually subid and not sid.
Correct tracking URL: http://clk.com-jdh.com/adclick.php?c...subid=%subid1%
UDPATE 2: turns out I cant retrieve my converting subIDs because I forgot to append the "&s1=" parameter to the the offer URL. ARGH! Lesson learned: test tracking setup.
Overnight stats
Clicks: 49 ($-4.90)
Landing page CTR: 50%
Total funnel CR: 10%
Conversions: 5 (+$7)
Profit: $2.10
ROI: +30%
Boom!
I've wittled down my ads to two images. These are giving me a CPC of only 10 cents, at which I seem to be comfortably in the money on this setup.
I've done some tweaking and dialing on my lander in response to the review thread I opened here. The result is that I now have a funnel that converted at 10% last night and is operating at a profit margin. I realise the day is far from over though, so I'll let this play out before I make any scaling decisions. I might ask for a bump on this offer as well at some time; it's converting well, but the payout is rather low at 1.40.
These are my funnel stats:
Action list
- Create a second variation of my lander to split-test
- Shop around on networks to find better paying offers
- Double campaign budgets pending today's results
- Try to sort out facebook pixel firing in CPVlab to help facebook with better OCPM bidding
Hmmm day parting might be playing a big role here. My profit pretty much evaporated as the (AU) day passed by. This is the second day in a row where I'm seeing this pattern. If it happens again tomorrow, I'm cranking up my budgets at night.
When is FB going to allow time of day scheduling? I'm nervous about pausing ads because I've heard that messes up CTR.
In other news, I've created a minor variation of my lander with a little urgency message ("offer only valid while stocks last. 7/200 playstations remaining in stock") on it.
Originally Posted by jdenhaan
Stats time
Ad clicks: 225
Conversions: 9
Revenue: $12.60
Ad CPC: ~$0.10
Spend: $25
Landing page CTR: 50%
ROI: -50%
Not great, but not a disaster either!
My only question is: how long will this campaign be viable? At what display frequency do you consider an FB campaign burnt out?
Next steps
- Create a second conversion path using a different offer (my landers are tailored to the offer, hence the necessity to set up a multi path campaign in CPVlab to fairly test the offers)
- Spin my winning ad images using inversion, sharpness, borders, etc.
Nice! At least you're getting close to being on the positive ROI side. Gonna keep following along
Have you tried collecting emails? I know in niches like diet opt in is a big no, does facebook allow opt ins in general? Anybody know?
I'm not sure Sergeru, but it's something I might look into in the future. I believe it's a bit tricky to prepop offers with CPVlab tracking links though, so I'm holding off on this for now until I figure that out.
Okay, time to come clean here and face reality: this isn't going to work. Not because of my lander (it's actually pretty good), not because of my ads (CPC is pretty low), but mainly because the offer payout is simply way too low. I will keep this campaign idea on hand for PPV at some later stage, but for now it's time to regroup and rethink my approach on Facebook!
So, after some thinking I've come to the conclusion that I may have abandoned my earlier golf offer too quickly. I need more patience and persistence in my efforts. It's a solid product with a clearly defined target market that's not too expensive at $29. The payout of $9.95 is also lots better than the 1.40 I was getting for the email submits.
I think my problem earlier was that my landers were just abysmal. I've therefore resorted to creating two new landers:
- A simple video lander: http://perfectpendulum.com-jdh.com/1/
- A sales letter (which is still in progress and therefore not in rotation yet): http://perfectpendulum.com-jdh.com/2/
I will also refine my targeting to target one specific US state and split-test age demographics and gender. I will initially target:
- Location: Florida
- Gender: Male
- Age: 30-40 vs. 40-50
- Like: various golf brands and stores
Happy to receive any feedback on my landers or campaign setup so far
Thanks, seeing actual live landers etc is really helpful.
Just remember that I'm a beginner too! Don't fly blind on my stuff, as it's untested and probably has its share of rookie mistakes.
For sure, I've zero intention of even looking at the same niche, much less same landers etc - just brings a follow-along to live seeing the nitty - gritty.
@jdenhaan - be aware, higher payouts aren't always your friend!
Hi Caurmen,
I understand that, but is 9.95 really a high payout?
It's not that high - I'd call it the high end of medium - but it will make a difference to the testing budget you'll need. It very much depends how tight a budget you're on as to whether that is an issue or not.
Home > Facebook & Instagram > Follow-along Campaigns