Baby Steps to Facebook Mastery (18)
Hi everyone! I’ve been learning a lot from this forum, reading anything I can find about Facebook. I’ve tried running campaigns on my own after learning from the other posts here. Unfortunately, I have yet to have a profitable campaign on Facebook after launching more than 10 campaigns. I thought that perhaps it would be good to do a follow along and to have people critic the way I have been running a campaign.
This is my first follow along so please do guide me as to how to better present data so that it is easily understood. I’m trying to be as open as possible and provide as much data as possible so that everyone can benefit from this sharing in one way or another. Cheers!
Traffic source: Facebook
Vertical: Mobile Gaming App
Offer(s): Last Empire – War Z
Geos: Global
Objective: Clicks to Website
Placement: Mobile News Feed
Budget: $5/day per ad set
Bid: Auto CPC
Day 1
Angles:
Angle 1: Zombie Mobile Games (18-24)
Angle 2: The Walking Dead (18-24)
Angle 3: Movies (18-24)
Campaign structure:
• 3 ad sets per Angle
• 1 ad per ad set
• 3 duplicates per ad
Ads:
Landing pages: Direct Linking
Updates:
Spend for the day: $28.61
Revenue: $0
Net profit/loss: -$28.61
Overall results:
Angle 1: 0.86% CTR $0.99 CPC $13.88 spent -100% ROI
Angle 2: 0.25% CTR $1.14 CPC $6.82 Spent -100% ROI
Angle 3: 0.32% CTR $1.32 CPC $7.91 Spent -100% ROI
Actions I took:
• Created 3 duplicate ads per ad set and keep the duplicate with the best results ASAP.
• Cut ads that fall below 1% CTR & no conv.
• Each image went through tweaking of Levels, Vibrance and Sharpenning.
• Angle 1 has the best results so split test that with some other testing.
Current tasks:
• Split test winning angles with more things.
Angle 1: Zombie Mobile Games (18-24)
Angle 4: Zombie Video Games (18-24)
Angle 5: Zombie Mobile Games (25-34)
Questions:
• May I know how I can improve my CTR and thus lower my CPC?
• Are my ads decent enough or do they suck?
• Is there anything I can do to improve my ads?
• What is a rough benchmark that my ads should hit to be considered decent?
Day 2
So here are my stats for day 2. I got 2 conversions! Seems like a possible doable campaign.
Angles:
Angle 1: Zombie Mobile Games (18-24)
Angle 4: Zombie Video Games (18-24)
Angle 5: Zombie Mobile Games (25-34)
Updates:
Spend for the day: $40.20
Revenue: $4.80
Net profit/loss: -$35.40
Overall results:
Angle 1: 0.51% CTR $1.07 CPC $13.87 spent -100% ROI
Angle 4: 0.36% CTR $1.26 CPC $12.63 Spent -100% ROI
Angle 5: 0.63% CTR $1.37 CPC $13.70 Spent -65% ROI
Actions I took:
• Created 3 duplicate ads per ad set and keep the duplicate with the best results ASAP.
• Cut ads that fall below 1% CTR & no conv.
Actually, all the ads fall below 1% CTR
Current tasks:
Add 3 more images to split test with 2 ads that had conversions for a total of 5 images:
Image 2, Image 3, Image 4, Image 5, Image 6
Will be using Angle 5 to do the testing. I think that having better images would probably boost my CTR and lower my CPC.
are you cloaking? FB wont let you to add DL to app install I guess
Good job on the 2 conversions.
Make sure that when a certain angle/ad doesnt work, you hypothesize WHY that angle isnt appealing (ie. The ad colors are too washed out, the image is ambiguous, it doesnt make the person feel powerful, etc).
Also, start testing some landing pages against directlinking. Landers are very powerful and good ones will increase your cr. I would build an entire content site around your landers though, because fb is very prickly about what domains can advertise on their site nowadays.
Originally Posted by triangle
Originally Posted by johnaff
I'll create a lander to split test tonight and see how it goes.
However, I'm wondering how do you find out why a angle doesn't work? Do you spend money to test your hypothesis or is it just an educated guess?
You create a hypothesis by running more tests.
Lets say you put up an ad image of gameplay and it doesnt work.
Lets create 2 hypothesis...
1.the ad didnt work bc the gameplay image was too zoomed out. You would test this by testing some ads where the gameplay pic is more zoomed in
2.the ad didnt work bc people dont like gameplay images. You can test this by testing images of a character leveling up, or a unique powerful item, or an icon, etc.
You could create MANY more hypothesis, but thats basically the jist of it.
Makes sense?
Day 3
Angles/Tests:
Images 2 - 6
I realized that I numbered the ads wrongly. The day 2 ads should be images 4, 5 & 6. On the bright side, I got another 5 conversions!
Updates:
Spend for the day: -$23.86
Revenue: +$12.00
Net profit/loss: -$11.86
Overall results:
Image 2: 0.84% CTR $0.83 CPC $5.00 spent -100% ROI
Image 3: 1.13% CTR $0.81 CPC $4.84 Spent -100% ROI
Image 4: 0.70% CTR $0.55 CPC $4.95 Spent -51% ROI
Image 5: 0.75% CTR $0.61 CPC $4.86 Spent -51% ROI
Image 6: 2.19% CTR $0.70 CPC $4.21 Spent +71% ROI
Actions I took:
• Created 3 duplicate ads per ad set and keep the duplicate with the best results ASAP.
• Cut ads that fall below 1% CTR & no conv. (images 2 & 4) EDIT* there was a delay in reporting and image 4 didn't have any conversions then
• Cut Image 3 due to High CPC
Current tasks:
1. Add a lander to split test.
2. Run Image 5 and 6 for one more day to confirm results.
3. Split test additional 3 Images.
Lander:
After reading John's advice (Thank you John! 
), I tried to think WHY did the other angles & images not do as well as the others.
I think most of the angles probably failed because it is not as targeted as the zombie games. They might like zombies but might not want to play them. As for the zombie games, if they are already playing one game, it is more likely that they will play another similar game. I think it is the same reason for why zombie video games did worse than zombie mobile games, because a video gamer might not be interested in a mobile game.
However, I can’t seem to think of any logical reason why 25-34 would have better results than 18-24. Perhaps the data has yet to reach significance and I jumped the gun and said one is better. They both may very well have the same results.
Is it wise to test this to get a definite answer? If it is clearly different, I won’t have a clue as to why it’s so. If both have the same results, then there seems to have no conclusion other than both of them are the same.
Also, I have no idea why any image did better than the others. There seems to be too many possible reasons why one image can outperform another.
Day 4
Spend for the day: -$23.46
Revenue: +$2.40
Net profit/loss: -$21.06
Overall results:
Image 5: 1.00% CTR $0.71 CPC $5.00 spent -100% ROI
Image 6: 1.62% CTR $0.71 CPC $4.84 Spent -100% ROI
Image 7: 2.95% CTR $0.71 CPC $4.95 Spent -52% ROI
Image 8: 0.45% CTR $0.77 CPC $4.86 Spent -100% ROI
Image 9: 2.19% CTR $0.66 CPC $4.21 Spent -100% ROI
Actions I took:
• Created 3 duplicate ads per ad set and keep the duplicate with the best results ASAP.
• Cut ads that fall below 1% CTR & no conv. (images 8)
• Cut Image 5 as it seems unlikely that it will ever become profitable.
Current tasks:
1. Run remaining images for 1 more day
2. Create new ads to prove hypothesis
Hypothesis: The previous angles might not work due to it not being relevant enough to the audience. Test this by changing the ad copy and images to fit the angle.
Angle 2:
Angle 3:
Day 5
Spend for the day: -$43.53
Revenue: $0.00
Net profit/loss: -$43.53
Overall results:
Image 6: 2.75% CTR $0.71 CPC $5.00 spent -100% ROI
Image 7: 1.38% CTR $0.71 CPC $5.00 Spent -100% ROI
Image 9: 1.41% CTR $0.71 CPC $5.00 Spent -100% ROI
Angle 2: 2.03% CTR $0.72 CPC $13.73 Spent -100% ROI
Angle 3: 0.61% CTR $0.65 CPC $14.30 Spent -100% ROI
Actions I took:
• Cut angle 3 as CTR below 1% and no conversions
• Cut image 9 as spend more than 4x payout and no conversions
• Cut image 7 as overall spend more than 4x payout but has only 1 conversions
Hypothesis Results:
Both the new angle 2 & 3 has a huge improvement in the metrics when compared to initial test. This shows that angles 2 & 3 sucked previously, as the copy was irrelevant to the audience.
Hypothesis: No conversions today might be due to it being a Thursday In US time zone. So far, my other offers have always done poorly on Thursdays. Maybe Thursdays are a bad day for mobile app installs. See if Friday would do any better.
Current tasks:
1. Continue testing image 6 and angle 2
2. Test images from angle 2 with image 6 audience
What is the payout? You have high click prices so you will need a correspondingly higher payout to make this back out for you based on current stats.
Originally Posted by cmdeal
Originally Posted by gasorama
Originally Posted by cmdeal
Originally Posted by gasorama
Day 6
Spend for the day: -$34.57
Revenue: $0.00
Net profit/loss: -$34.57
Overall results:
Image 6: 1.40% CTR $0.56 CPC $5.00 spent -100% ROI
Image 10: 1.34% CTR $0.62 CPC $4.95 Spent -100% ROI
Image 11: 4.51% CTR $0.83 CPC $4.95 Spent -100% ROI
Image 12: 1.48% CTR $0.55 CPC $4.95 Spent -100% ROI
Angle 2: 4.07% CTR $0.53 CPC $14.72 Spent -100% ROI
Actions I took:
• Cut angle 2 as spend more than 4x payout and no conversions
The best of Image 11’s duplicates has a 7% CTR thus wait and see if the CPC could reduce low enough.
Hypothesis Results: Friday had no conversions despite the increase in metrics. Might be due to link issues. Checked the links and they are fine. Thus, no other conclusions other than it just didn’t work.
Hypothesis: Games angle + angle 2 should result in more specific targets and get cheaper CPC and Higher CTR.
Current tasks:
1. Continue testing images
2. Change ad copy
3. Test new audience
Day 7
Spend for the day: -$44.51
Revenue: $14.40
Net profit/loss: -$30.11
Overall results:
Image 6: 2.13% CTR $0.56 CPC $5.00 spent -4% ROI
Image 10: 0.91% CTR $0.77 CPC $4.95 Spent -100% ROI
Image 11: 2.62% CTR $0.42 CPC $4.95 Spent -4% ROI
Image 12: 2.22% CTR $0.63 CPC $4.95 Spent -100% ROI
Control: 1.81% CTR $0.55 CPC $13.25 Spent -64% ROI
New Target: 2.74% CTR $0.51 CPC $13.34 Spent -82% ROI
New Copy: 2.07% CTR $0.52 CPC $12.92 Spent -81% ROI
*control is the combined stats of images 10 - 12
Hypothesis Results: Data shows that indeed the results were much better the new target. The new copy also worked pretty well.
Hypothesis: Seeing the results of the new things are both better, the new copy and the new targeting be even better than everything else.
Actions I took:
• Cut image 10 as the CPC was higher than $.70 and still no conversions.
• Cut image 11 in new copy as the CPC was higher than $.70 and no conversions.
Current tasks:
1. Continue testing existing stuff
2. Test new audience + new copy together
Day 8
Spend for the day: -$46.38
Revenue: $16.80
Net profit/loss: -$29.58
Overall results:
Image 6: 0.87% CTR $0.57 CPC $4.54 spent -100% ROI
Image 11: 3.21% CTR $0.35 CPC $4.55 Spent -100% ROI
Image 10 (1.1): 1.04% CTR $0.47 CPC $4.73 Spent -49% ROI
Image 12 (1.1): 1.93% CTR $0.47 CPC $4.73 Spent -100% ROI
Control (2): 2.81% CTR $0.50 CPC $13.57 Spent -82% ROI
New Copy (2.1): 3.17% CTR $0.59 CPC $14.26 Spent -50% ROI
Hypothesis Results: Data shows that indeed the results better in ROI and insignificant in others. The new copy also worked pretty well.
Actions I took:
• Cut image 6 & 11 as so many days and yet have yet to breakeven. Thus results should be constant.
Current tasks:
1. Continue testing existing stuff
2. Test new images using new copy
3. Test new copy using Img 11
4. Test smaller age groups by 4 18-34 using Img 11
Day 9
Spend for the day: -$66.01
Revenue: $12.00
Net profit/loss: -$54.01
Overall results:
Image 10 (2): 0.87% CTR $0.57 CPC $4.68 spent -100% ROI
Image 12 (2): 1.21% CTR $0.45 CPC $4.83 Spent -100% ROI
Image 10 (2.1): 5.05% CTR $0.52 CPC $8.72 Spent -100% ROI
Image 12 (2.1): 4.76% CTR $0.26 CPC $8.92 Spent -100% ROI
Image 13 (2.1): 2.59% CTR $0.55 CPC $4.91 Spent -51% ROI
Image 14 (2.1): 6.19% CTR $0.35 CPC $4.55 Spent -100% ROI
Image 15 (2.1): 3.26% CTR $0.46 CPC $4.60 Spent -48% ROI
Image 12 (1.1): 3.25% CTR $0.45 CPC $4.99 Spent -4% ROI
Image 11 (2.2): 6.57% CTR $0.49 CPC $4.40 Spent -45% ROI
Image 11 (25-29): 5.30% CTR $0.56 CPC $5.00 Spent -100% ROI
Image 11 (30-35): 4.21% CTR $0.66 CPC $4.63 Spent -100% ROI
Hypothesis Results: Need to test more
Actions I took:
• Cut a few images as so many days and yet have yet to breakeven. Thus results should be constant.
Current tasks:
1. Continue testing existing stuff
2. Test new lander
Day 10
Spend for the day: -$44.06
Revenue: $9.60
Net profit/loss: -$34.46
Overall results:
Image 12 (2.1): 3.42% CTR $0.34 CPC $9.96 Spent -100% ROI
Image 13 (2.1): 3.33% CTR $0.56 CPC $5.00 Spent -52% ROI
Image 14 (2.1): 5.72% CTR $0.28 CPC $4.71 Spent -100% ROI
Image 15 (2.1): 1.78% CTR $0.57 CPC $4.57 Spent -100% ROI
Image 12 (1.1): 4.60% CTR $0.42 CPC $5.00 Spent -100% ROI
Image 11 (2.2): 8.12% CTR $0.23 CPC $4.98 Spent -100% ROI
Image 11 (25-29): 7.56% CTR $0.28 CPC $5.00 Spent -52% ROI
Image 11 (30-35): 8.81% CTR $0.28 CPC $4.84 Spent -1% ROI
Hypothesis Results: Need to test more
Actions I took:
• Cut a few images as so many days and yet have yet to breakeven. Thus results should be constant.
• Also, the EPC is way lower than CPC thus not profitable
Current tasks:
1. Continue testing existing stuff
2. Test best age group with best copy (30-34 + 2.2)
3. Test new audience with walking dead instead of resident evil
4. Expand gaming audience for zombie
Day 11
Spend for the day: -$44.06
Revenue: $9.60
Net profit/loss: -$34.46
Overall results:
Image 13 (2.1): 2.90% CTR $0.63 CPC $5.00 Spent -4% ROI
Image 15 (2.1): 0.99% CTR $0.54 CPC $4.33 Spent -45% ROI
Image 11 (2.2): 3.96% CTR $0.27 CPC $4.80 Spent -50% ROI
Image 11 (2.1x): 5.11% CTR $0.30 CPC $4.78 Spent -100% ROI
Image 11 (2.1xx): 7.03% CTR $0.27 CPC $4.89 Spent -51% ROI
Image 11 (2.2xx): 4.23% CTR $0.62 CPC $9.34 Spent -74% ROI
Image 11 (3.2): 7.02% CTR $0.21 CPC $8.95 Spent -73% ROI
Image 11 (4.2): 5.12% CTR $0.43 CPC $9.41 Spent -100% ROI
Hypothesis Results:
control (Resi evil + zombie game)
3.2 (walking dead + zombie games)
4.2 (resi evil + expanded zombie games)
Hypothesis:
4.2 sucks because there’s PvZ and it probably is not congruent with the game driving up cost and all. If take away PvZ, it should work better.
Actions I took:
• Cut a few images as so many days and yet have yet to breakeven. Thus results should be constant.
• Also, the EPC is way lower than CPC thus not profitable
Current tasks:
1. Continue testing existing stuff
2. Test new ad (img 13 + 2.2 + 30-34)
3. Test 4.2 again w/o PvZ audience
Day 12
Spend for the day: -$57.81
Revenue: $16.80
Net profit/loss: -$41.01
Overall results:
Image 13 (2.1): 2.36% CTR $0.55 CPC $4.43 Spent -46% ROI
Image 11 (2.1xx): 5.17% CTR $0.29 CPC $5.00 Spent -100% ROI
Image 11 (2.2xx): 4.83% CTR $0.34 CPC $9.82 Spent -27% ROI
Image 11 (3.2): 4.80% CTR $0.19 CPC $9.73 Spent -75% ROI
Image 11 (4.2): 3.70% CTR $0.29 CPC $9.58 Spent -100% ROI
Image 11 (5.2): 5.81% CTR $0.21 CPC $9.85 Spent -100% ROI
Image 13 (2.2xx): 2.77% CTR $0.59 CPC $9.40 Spent -49% ROI
Hypothesis Results:
PvZ is the cause of 4.2 to not do as well as it could.
11 is a better image.
Actions I took:
• Cut a few images as so many days and yet have yet to breakeven. Thus results should be constant.
• Also, the EPC is way lower than CPC thus not profitable
Current tasks:
1. Continue testing existing stuff
Day 13
Spend for the day: -$41.59
Revenue: $4.80
Net profit/loss: -$36.79
Overall results:
Image 13 (2.1): 2.55% CTR $0.40 CPC $4.44 Spent -46% ROI
Image 11 (2.1xx): 3.70% CTR $0.23 CPC $4.50 Spent -100% ROI
Image 11 (2.2xx): 4.16% CTR $0.31 CPC $8.87 Spent -100% ROI
Image 11 (3.2): 4.59% CTR $0.21 CPC $7.58 Spent -100% ROI
Image 11 (5.2): 4.67% CTR $0.18 CPC $7.90 Spent -100% ROI
Image 13 (2.2xx): 2.32% CTR $0.46 CPC $8.30 Spent -71% ROI
Hypothesis Results:
Indeed without PvZ, it’s so much better.
Actions I took:
• Cut a few images as so many days and yet have yet to breakeven. Thus results should be constant.
• Also, the EPC is way lower than CPC thus not profitable
Current tasks:
2. Continue testing existing stuff
3. Test 3.2 + 5.2 audience
Home > Facebook & Instagram > Follow-along Campaigns