Home > Facebook & Instagram > Follow-along Campaigns

---

MOBA game in Facebook - Starting now! (20)

---

Hey Guys,
First of all, this is my first report, hope i'll have much success to wrote much more

So basically, i really don't have a clue about any of the Online games, but because of the low payout, i will be able to getter some data with low budget.
Our offer is a Multi Online Battle Arena game, which i won't mention in this part.

Offer
Payout - 2$
GEO - US
Daily Budget - 60$

Creative
The number of ads which will make sense to run on one Adgroup will be 3, BUT, i choose to bulk upload 10 ads because there are 3 angles of creative inside (3-4 ads for each).
I decided to use only RHS first, because i think that if a person will be able to see the ad on the side (which is now bigger than the old 100X72), he will be more engaged. In addition, the clicks will be cheaper - means more data.
So we got those 10 ads (some are with button in the image, which looks very small on RHS. my first thought when i create them was to use Desktop Newsfeed)

Title: Play (name) Now! (hard CTA)
Description: Choose X - Be Y

Segments
The first thing i did was to run a quick report on Audience Insights and understand the demographic of the chosen game. In our case, both Man and Woman are pretty much the same. The age range is of younger people.
After i understand that, i created 6 different segments (10$ a day each).

1. Fan page of similar game which i've got admin access to - 10K
2. Affinity > 10 (as Zeno mentioned in one of his posts) - 200K
3. Affinity > 10 and played lately - 80K
4. Played lately - 300K
5. MOBA games - 200K

We're all set, Facebook approved all the ads.
Lets see how we're getting forward - will update tomorrow

---

Quick Update,
I decided to add another geo, in a different time zone (Greece) so the campaign will be effective 24h.

Some thing i forgot to mention it tracking:
Right now i'm not using any tracker, but i implemented and tested the FB Javascript pixel in the Affiliate network.
The geo-redirect is taking all the users to same page, just without the aff-id, so everything is cool with that.

---

First day update (the first half day. From now on i'll report after every full day that the campaign has run).

US
Spent: 27$
Conversions: 4
Revenue: 8$
Profit: -19$
ROI: -70%

GR
Spent: 10$
Conversions: 3
Revenue: 2.5$
Profit: -7.5$
ROI: -75%

First thing first - i see that there are some conversion, so i hope i'm in the right direction.
What should i do next? - optimiziation.
I will pause the different segments which didn't had a single conversion (unless they had less than 20 clicks).
Will update tomorrow.

---

4.18.15

US
Spent: 35$
Conversions: 0
Revenue: 0$
Profit: -35$
ROI: 0%

GR
Spent: 10$
Conversions: 1
Revenue: 0.83$
Profit: -9.17$
ROI: -91%

I saw in the first run that there is might a chance that the campaign will work.
I decided to move it from RHS to NF. I paused the US campaign and staid only with the GR. the Clicks are cheaper because of the greater CTR.
The reach in GR is much lower (around 30K) so i will split test it by increase the age range vs add some more interests to the 2 segment who gave me the conversions, i will consolidate both as well into 1 segment.

Second thing i'm going to do is check Direct VS LP.
So we will have 4 segments now (with 10 Ads each with 3 angles):
1. Seg 1 - Direct
2. Seg 2 - Direct
3. Seg 1 - LP
4. Seg 2 - LP

In addition to that, i changed a lil the text copy by asking my AM (which played the game) "what is unique with this game? what does the game have that is better than the others?"
The title will still be hard CTA (Play for free!) and the description will be social proof (more than 300K players..)

---

I think you're jumping around too much.

So far you've:

- Picked 6 segments and started with some RHS ads.
- You then added another country.
- You then moved to the news feed and paused US.
- You then decided to expand an age range and alter some of the segments, and consolidate into one segment.
- You then decided to throw in a landing page or two.
- You even changed the ad copy and did so by asking your AM about the game, rather than checking it out yourself (do your own research here).
All of this in the space of two days with about $60 and $20 spend.

This is a bad way of doing things as you have started with a set of conditions, proceeded to buy a small amount of data and then have made a lot of wild changes to important variables with no data behind those decisions.

Fact: Your initial RHS ads converted.
Fact: You likely spent less than $10 on any one segment before killing them.
Fact: The RHS is often harder to make work, so if it does OK straight away, it's a good testing zone for porting to the NF.

It would be much better to start with those 6 segments and spend more on them, testing them properly and learning what segments convert more (with a reasonable spend), and focusing on one country, rather than wildly jumping between approaches - this is a surefire way to lose money and learn very little unless you get very lucky.

For this follow-along approach I would go back to the beginning and make your tests more structured. Collect data on each segment and present that data here, then explain why you did what you did - make sure there is reasoning behind your decisions rather than emotion, excitement and irrationality.

And of course, good luck!

P.S. Definitely get a proper tracking system in there and with a custom domain if using Voluum. Direct linking to an affiliate network's URL is bad for business regardless of their geo-redirection.

---

Thanks a lot for your feedback Zeno. I really appreciate it!

This is part of my research! what will be more accurate than asking someone who's playing this game itself?
I don't know the difference between this game or another, because i'm not playing in any one of those. This is the reason i needed an opinion of someone who's already played this game.
In this case, it was my AM (which own the network, and i know him for a long time). It wasn't came from laziness!

That is true, BUT from my experience with other campaign behaviours, it will be more likely to perform better with NF ads. The proof to that is the CPC dropped down.
My reason to run RHS was to collect initial data, which i got. Now i have to scale it in terms of performance (not spend).

What do you think will be reasonable spend when my payout is 0.83$?

After a lot of thinking, was creating 4 new segments, which will run the same ad copy and image.
The segment i use now are consolidation of Affinity > 10 games and list of other MOBA games:
1. Group Age 1, Direct linking - 30K
2. Group Age 2, Direct linking - 15K
3. Group Age 1, LP - 30K
4. Group Age 2, LP - 15K

I'm willing to spend on each segment 10$ a day. If i will get results which well be worse than the previous one, i will pause this campaign and active the less one.
Like i asked before, how much data do you think that will be enough in order to decide what to optimize?
I will do so, that is something i already working on and i hope to get it done by today.
But what do you mean by bad for business?

---

4.18.15
Spent: 15$
Conversions: 1
Revenue: 0.83$
Profit: -14.13$
ROI: -94%

It seems like the click price start to increase, but i didn't get enough data.
Facebook spouse to deliver total 40$ per day, the reason for this 15$ are that i start those Adsets in the middle of the they.
We'll wait for tomorrow to understand better.

---

I suppose if you have no interest/insight about games then asking someone who has played it (and is a gamer) is the next best thing.

I would still however download/register/install and so on to have a look at it and you can drag yourself through Google searches and forums to learn a bit more about user opinions of the game, which can give you angles. Understanding your prospective customer is a big part of marketing.

Sure, but I don't think you really collected much data - you spent $60, but it was very fragmented, so only really said one thing: "conversions are possible". A relatively weak observation. Also note that CPC going down is not proof of anything... you pay for a lot more 'click' actions on NF ads and this will, all else being equal, result in a RHS ad almost certainly getting higher CTRs when ported to the NF. In the end your CPA matters. Not intermediate metrics.

Facebook is complicated. There are a lot of targeting options and segments you can test. With that payout I would probably spend at least $10 on each age/gender segment and would there could be half a dozen other segments to look into. This could lead to spending $200 on it in one day, so really you have to pick a few broader segments first and look at average performance across all of them, then use some intelligent thought process.

If you spend $10 on 5 segments and get 100 clicks to the offer, not a single one converts, then the decision is obvious. If you spend the same and 2/5 segments run at -40% ROI, the others at -80% ROI, then there is more testing to be done.

There is no specific set of rules - there is simply too much possible variation. I'd suggest some 'sensible guidelines' like spend at least 5x the offer payout on any one segment you test.[/QUOTE]

These are extremely small audiences. Greece? Chances are you will spend very little and hit 50% of the users in a day, giving no potential for scale. This isn't something I would invest much time into.

Do a quick sanity check. If CPM in the NF is $1, and on average there might be 75% of those users on in a day, then that's a max spend of less that $25 on segment 1 in a day. That audience will burn out very quickly. In a best case scenario, say you get 5% CTR and of those who click through, 10% convert. That's a potential end result of 4500 people clicking through, 450 converting, and about $375 revenue over the entire campaign duration - so maybe $100-200 profit and then it dies. That's not even enough to get a network payout!

Lastly, you should not be making different campaigns to test direct linking and using an LP. This is something that should be done on the back-end by a tracking system. Doing it this way (making separate traffic source campaigns) wastes time, money, and is bad split-testing practice.

Read the part of my FB guide about geo-redirection. Pasting an affiliate network link into Facebook is account suicide as they eventually will screw you over with their stupid geo-redirection, and their URLs will often be blacklisted (using their tracking domain = flags you). You should never leave things like this outside of your control unless the circumstances are appropriate (e.g. you have a fine-tuned campaign, direct with advertiser, going direct to their link saves redirects > increases profit).


Having a small number of impressions will make your data less reliable, but has nothing to do with the CTR being low or high. It will be that your ads don't appeal to the audience, so don't get clicked much.

Re: your other observations - yes, this kind of thing will happen. It's up to you to just collect enough data for your comparisons to be statistically significant, so that you make sound decisions.

One issue here is that the DL and LP clicks come from different ads, which compromises your test - its possible that the ad performance and placement is confounding these results, hence why you use a tracking system to rotate DL/LP options.

Now, as for moving forward, the bottom line is your ads suck.

The CTRs are very low leading to prohibitively high CPCs. Some of your segments + using an LP (or split-tesitng LPs + DL to get to the best combo) might get to break even, maybe even profit if you can make much better ads.

I think the imagery is letting you down - it's very weak. Make things pop, stand out and be vibrant and exciting. Right now they look bland, and bland doesn't do so well in the NF where there are dozens of images from news articles/ads vying for a users attention.

---

Thank you Zeno, for every single minute you're writing here.

Reach
It seems like i will pause the campaign on GR. If i was seeing better results on the beginning of the campaign, it would make sense to stay with it. But unfortunately it looks like almost anyone inside the target saw the ads and didn't enthused

Tracking
Already set Voluum, will now on work only with that

Ads

You're abso-f***ing-lutely right.
I don't know what i was thinking to myself.
I read your post with the images, and been trying to figure out what process did you used in order to get along with the right materials and put them together to 1 piece of great ad. Can you continue spread some light here with that? it is really the difference between loosing campaign and winning campaign!

My last thought with this campaign are:
1. Should i use the same game, in different GEO? (US/CA for example, when i'll get 2$ payout), or should i try other game in different GEO, when i'll get reach above 150K?
2. The first segment i used in the campaign was MOBA, Afinnity > 10, and people who played games lately (when i tired US). It look like the MOBA games provide me stable results in GR. What should i learn from this case? should i use the whole Affinity+MOBA segment and spread it to different demographic in the next campaign? or should i use different interests and start collecting data one them from the beginning?
3. I want to start again from scratch, the process i've been taking really wasn't stable. Any recommendations on starting this new process from beginning?

---

Bump

---

OK, images (specifically for games):

1) Find the "media pack" for the game you're promoting. That'll be put together by the game's creators or publishers, and will contain what they think are the best images to promote the game with. An excellent start!
2) Learn a bit about shot composition. Googling that term and reading some digital photography blogs will get you 90% of the way there - see http://digital-photography-school.com/rule-of-thirds/ for example.
3) Make sure all your images "read" well even when they're not full-size on your monitor. This is super-important. If you can't see what the image is when you've reduced it to the size it'll appear in Facebook, then it's not good at all.
4) If in doubt, show your images to someone else - quickly, don't give them too much time. Then ask them what the image is. If they can't clearly explain what they were looking at ("A tank crashing through a wall", "an elf girl with implausible cleavage"), the image doesn't work.
5) Finally, learn a bit about image processing in Photoshop, particularly contrast-related stuff. Levels are your friend, as is the Sharpen tool (in moderation). Here's one tutorial, but there are tons more: http://digital-photography-school.co...ur-images-pop/ . Search "make images pop" - that's what you're looking for as a basis.

Hope that helps!

---

Just wanted to chime in and stress this point. For better or worse, facebook makes decisions extremely early in an ads lifetime. You can create completely identical campaigns/ad sets/ads and they will perform very differently depending on what happens in the first 500 impressions.

As far as everything else said in this thread, I agree 100% with zeno. Very good advice.

---

The image is important but only part of the puzzle. I'm not sure what additional info you're after, I've talked about it quite a lot in the other threads. The main thing is to source high quality material, not be every other lazy affiliate and just use stuff from Google Image searches, and get familiar with Photoshop. Treat it like your dojo.

1. Do you mean switch targeting segments and use the same game or purely switch geos? Why not both? You should be testing things so don't funnel yourself into an A or B dilemma when doing A and B just requires more effort and a little bit more funding for testing. Ideally do aim for audiences that are bigger, i.e. 100,000+ so you have some room to test.

2. Always test a variety of segments. In this case you mentioned affinity > 10 (I presume you mean the interests you got from audience insights related to a MOBA target) AND who played recently - are you intersecting those? Or doing them separately? I would not mix the interests list AND people who had played recently as this will take the affinity > 10 audience and cripple it to a much smaller audience. In this case, you'd test these separately to look at the affinity list only + how that list performs when further qualified down to only those who ALSO fell into the 'played games recently' list.

3. Make a list, follow it. It's easy to get distracted. Focus on the offer first and picking one, getting feedback from AM and any insider info, and only then committing to time on creatives and campaign setup. Try to get a list of multiple offers then quickly research them. Sometimes one stands out as easier to start with due to their being better creative material available.

---

Nice work! Yes, those images are much better. Image 3 and 4 are particularly good to my eye, although obviously the data will tell.

FYI, the first and second images would probably benefit from a bit of additional Photoshop work just to make them pop more. Here's the second image before and after I did some massaging of the levels and ran an Unsharp Mask on it, for example. The effect is subtle but noticable - shapes are sharper and colours more distinct and easier to pick out on an FB page:

---

Looks great!
Would love to know which exactly tools did you used to highlight the image?
Lucky me, i sort of know how to work with Photoshop

---

Three simple tools to use:

1. Vibrance (alongside brightness/contrast, hue, etc.)
2. Colour dodge tool
3. Blur tool to add depth of field (blur out background clutter).

I wouldn't have thought this offer would convert too well - well done!

The landers are clearly the way to go. One advantage you have here is that Star Wars/Star Trek fans are often very passionate about them, almost cult-like.

---

Fairly simple, really:

1) I used the Levels tool to bring up the brightness of the highlights (notice that the lens flare is blown out, ie solid white, in my version) as well as the mid tones.
2) I pushed the Vibrance up until it wouldn't go up any more.
3) I applied an Unsharp Mask to sharpen all the details.

If I had more time I'd probably use the Camera Blur tool as Zeno suggests to bring out some elements of the picture - although a long shot like that doesn't need it as much. I'd also reframe a bit using the Content-Aware Fill tool to frame the main picture elements a bit better, with the space station on the right-hand-side line of the Rule Of Threes quadrant lines.

---

Just something I noticed:

It says StarT Trek in your ad headline. Is that on purpose? Might hurt credibility a bit.

---

4.28.25
Spent: 60$
Conversions: 5
Revenue: 11.9$
ROI: -80%

Time For Optimization
With a view of the whole campaign duration, i can see that 1 segment have the best results (in therms of CPC, CTR, CR). As well as the first LP.
The Multi Product ads didn't performed well unfortunately but there are plenty more ads types.

What i would do next?
I'm going to eliminate the other segments that didn't performed well, and stay in this time with the first LP.
I Going to test 4 different images (2 that already performed well and 2 new, similar). In addition to that, i'm going to split each ad into different ad set so i will have full delivery for each creative.
Will try Website Conversions objective and Clicks To Website as well.

---

Just stumbled upon this and following. Are you still doing it? The german text could also need some improvements.

The CTA "Jetzt Spiele" is definitely killing some conversions, sounds super weird. Let me know if you need help with german texting.

Cheers

---

Home > Facebook & Instagram > Follow-along Campaigns