Home > Paid Traffic Sources > Facebook & Instagram

---

Insane number crunching for optimization process, please advise! (13)

---

Hi guys,

I have two sets data now using Voluum, across last 2 days and the previous 2, all within this week. But i think i'm getting lost on which ones I should split-test further - or if I should even bother continuing at all.

Payout: $ 0.27
Total FB Spend: $ 383.00
Total Revenue: $22.37

Yeah crazy bad numbers. But I managed to dig out some ads recently which looks promising. Please shoot and advise which ones I should continue to split test

*Note, I've culled alot of ads which are not converting, so figures here will not tally with the total spend and total revenue stated above.

*Those ads which does not have a corresponding value in the previous batch are new ads.

Currently, the variables I have tested are AGE, IMAGE, CARRIER, WIFI, GEO.

Constant variables are GENDER, AD COPY, INTERESTS, PHONE OS. Thought I test the above variables first before going in deeper.

Ad budgets are set to $2-$2.50 daily.



(1) At this stage, is it worth spending more to test further on those constant variables?

(2) Images is the most important variable for Facebook, and so far I have tested about 14 images for 3 angles. Have I tested enough images/angles?

@Zeno, @franco...I apologise in advance for tagging you here. Hope you can spare some time & expertise.

---

OK, first things first - I'd definitely advise testing more images than that. 8-10 per angle is probably a better number. How many different copy texts did you use with those images, and how did the angle play into that?

I'm slightly confused by the spreadsheet: can you explain a bit more what each of those lines consists of, in terms of geo, image, ad copy, and angle? Are these from a mixture of angles or just one? And the "change from previous batch" lines - what's the "previous batch" in that context?

Also, were these campaigns direct-linked or using a lander?

Finally - if you feel comfortable showing some of these images and explaining the angle a bit, we can probably give you more advice.

---

Solid post by Caurmen.

Please feel free to show some images as well - and a little bit of insider info into your thought process in terms of how you approach angles, johndeng. You don't have to give your campaign or creatives away, but its easier for us to give you better and more concrete optimization advice if you share some more vital details

---

It's a bit hard to figure out what's going on here.

Firstly, you should probably break the data down by geo, gender, etc. - whatever you are testing and splitting by, so that we are looking at the segments individually. (rather than having geo1-m-18-img1 actually break it down into different tables).

Secondly, it would be helpful to see impressions - clicks - CTR - conversions - spend - revenue rather than the current data. No one wants to take your spend - profit to get revenue, divide that by payout to get #conversions and then divide that by conversion rate to get clicks, then figure out CTR from there. Put all the useful metrics in plain sight. Note CTR is useful to state for Facebook because it does have an important effect on ad performance, depending on the bidding mode.

E.g.

---

That is a really low payout, especially if you are throwing facebook traffic at it ... just sayin'

---

Hi again,

Sorry, too abit long to reply...busy weekend and made some mistakes in my re-stitching of the figures...

All these are direct. I'm on Mobile Appetizer essentially. Plus the payout was just $0.27, and I read somewhere here that payout this low should just be a direct link until I can get a +ve ROI before moving resources into testing landers.

Yeah I know...I was going along the saying of low payout = more data at lower costs, plus "sticking to a traffic source till I master it". I tried promoting a $6 health offer on Facebook last year, spent $5k with $500 revenue.

I ended up taking a break LOL.

I'm posting my stats here to see if there was anything wrong with my mindset / process in my FB testing. Or if I should just ditch FB as a traffic source and move to DSPs.

Apologies...I thought those figures will kind of provide enough insight. Wanted to post more figures but thought that will just be spammy and confusing. Like to add that the ads here do not represent all the ads that I have ran. I culled away quite alot, and presenting their numbers here will just make a mess.

Geo1:

I ran these ads for 2 days, and thought the number of impressions wasn't significant enough, so proceeded to ran a further 2 days on it. No attempts at optimization, all settings remained.



Geo2:

This set was silly. I identified 3 promising ads but missed them when setting up my second batch of ads for Day3-4. So no optimization here too. Have to confess I actually got lost sometimes analyzing the numbers and forgetting which ones to dig in.



Geo3:

Spent quite alot on this, but my optimizations did not yield much. Hope the screenshots explain what I tried to do.



*EDIT all using the same ad copy for Days 3-4

Geo4:

This is the most frustrating one. I had ran campaigns late last month when my CPVLab was still on shared hosting, and again when I tried moved my CPVLabs to Digital Ocean earlier this month thinking that the improved redirect speeds should yield some results. Nothing too.

I ran campaigns again this week on Voluum to give it another shot, again nothing.

*Note: There maybe just 5 ads shown here, but I have ran about 30 ads via CPVLabs and spent about $100 bucks here. Why? Because I have a gut feel this geo will be huge, plus I had CTRs hitting 10%-12% with eCPCs around $0.05-$0.10. But just little to no conversion. In total it's about 6-10 conversions.

I'm suspecting that maybe it's because the telecoms infrastructure is just bad, so I ran another campaign focusing just on the biggest city which I checked should be better. Again nothing.

So is this a dead horse? There's no point in hitting further?



Images:

Well, my images are pathetic. Tried looking for people to do it, but it's too costly at $50 per banner.




Angles

Using those images, I tried 3 angles out of 8.

1) Speed up your phone
2) Slower battery drain
3) Testimonial style, ripped from what people are saying about the app in Google Play store reviews (i.e. "I thought this app is just another clone. But my phone just got so much faster! This is no junk app!")

Please shoot

*Just to add...I think those volumn visit figures are always higher than what is presented by FB due to those review visits / bots. Still kinda disappointed Voluum can't allow those stats to be modified. At least there should be an option to disallow those reviewer/bot stats from making it into the reports.

This just simply makes Voluum useless as a platform to do any numbers. It's currently a $99/mth redirect and conversion counter to me.

---

Yeowww! Now that's nicely presented :P

Bookmarking to read through and respond tomorrow!

---

THANKS ZENO!

Looking forward to it

---

Ok, so the overall impression I'm getting is that you're spreading yourself way too thin over multiple geos and targeting options to the point that you have no useful data.

The fact you think Voluum is just a redirect/conversion counter aligns with that.

E.g. some miscellaneous points:

1. What is with the Behaviours - Samsung or All Android devices <- I'm not sure why this is involved at all? If you are targeting Android, that's it, that's all the targeting you need. Don't overcomplicate things at the risk of reducing your audience.

2. Motorola smartphones. WiFi and carrier in capital city only. <- these are not things you target initially, you are wasting money. This is what your tracker is for. Voluum can tell you about the cities, countries and ISPs of your users. Use that data to make decisions and get insight rather than needlessly spreading your testing out.

3. "I'm suspecting that maybe it's because the telecoms infrastructure is just bad" <- well, sure, but again you should have data that will help you validate this. I don't think people often go targeting a major city only because of speculation like this, it's a big time waster when you have GeoIP data.

4. in Geo 1, image 1b was kicking ass. Why not kill off everything other than that and iterate from there?

5. Your CTRs are just too low given the costs in your current demographic and the payout of your offer. You're dealing with 5-10 c clicks and your offer barely pays out $0.30. It's a stretch to rely on 30+% conversion rates to break even. You should always take a step back and assess whether the economics are against you. Facebook is expensive and competitive. The fact is there is a limit to how cheap your clicks will get and what you will get in more realistic circumstances.

6. Your images suck. The CTRs reflect this but then it depends on the demographic and what you're doing. In any case, you really want to shoot for much higher CTR in the news feed just to get traction. Your images don't really say anything to me. What am I even looking at? What does this image mean? Your images should capture people's attention and right now they look too generic.

You mention $50 per banner - where on earth did you get this from? I don't know of anyone charging $50 per Facebook banner! That's insane! For that I could hire a professional designer on oDesk for 2 hours and just ask them to make 20 awesome looking images, and a good designer would smash them out. Hell, I could do it!

Anyway, you should get better versed with Photoshop. It's freely available so there is no excuse for using terrible images with no pazazz or effort. Get your creativity hat on.

7. Lastly, you mentioned angles. Were these on landers? I'm assuming so since ad copy 1 was used throughout. Think more niche and targeted when starting out. Improve battery life. Uhuh. Heard it before a gazillion times. Everyone wants it, everyone has offered it, everyone has seen it.

What about battery jedi - newly discovered passive app throttling techniques reduce how much energy your phone wastes every day, giving you more time on the go.

Battery dying while watching that damn Uber driver? Reduce how much map browsing zaps your battery!

All that swiping draining those precious minutes? Swipe right on this ad and we'll give you your perfect app match (AND we'll hook you up with better battery life, how's that for putting out ;-)

You get my drift. There's a lot more you can do than the most generic things that come to mind.

8. Ages matter. I'm not sure how big your audiences are, but focus on a geo, target all Android phones, and split it into Male 18-21, 22-25, 26-35 for example.

---

THANK YOU @ZENO

for making the time to respond with such elaboration. Can't tell you how much I appreciate it.

I was targeting Samsung as I was testing copy calling out to Samsung users only. But point taken, about overcomplicating things.

*facepalm.

Thanks for pointing this out. Ok, Voluum is more than just a redirect / conversion tracker.

Yeah. My first run using my CPVLabs kind of indicate that, but I remained hopeful the new redirection speeds with Voluum and the slightly higher payout on a new offer will make it better.

I have spent over $5k now on Facebook with no profitable campaigns. Do you think I should move out and go into mobile DSPs like I have read in so many success stories here?

I just feel stuck about the notion of mastering one traffic source vs moving on to the next patch of green grass. Spent 1 or 2 thousands with SiteScout and Adwords and Google Display Network after being banned by FB twice, no success too.

So...I don't really trust what I'm thinking sometimes. What do you think? Please advise...

Yes $50 per banner. I got the quote from oDesk too! I posted a request for quotation for "8 banners for Facebook advertising", and received those quotes. Thought it was for the whole batch, but no, it was $30-$50 per banner. The best I got was $60 for 8 banners. For the following quality. No way I can spend on that.

How did you post your request to get those prices?



I'm ok with using photoshop as I used it to edit my own landscape photography shots every now and then, but I'm working on churning out banners from nothing. Took me a whole afternoon to churn out 8 banners. So yeah, working on it and trying to get some better designers to do it for me to spend more time on the angles.

Ok...will get to work more on my angles.

---

Facebook is a great traffic source to master.

We spend $XX,XXX a day on Facebook, but to be honest, I am not sure we could get anything on FB to back out at $0.27 a conversion ...

---

A few more points on top of Zeno's here:

- I'd strongly recommend choosing images specifically for your angle. The image and the copy needs to work together.
- Your images aren't actually terrible - however, I think you're being a bit too "clever" with them. Common problem! Bear in mind that people won't take any time to think about them, and to think, for example, that the cobweb is meant to imply that the phone's slow. You need an image that grabs right out of the gate.
- As Zeno says, Image 1b in Geo 1 is definitely the most promising from the data you've given - try iterating from that. You're nearly at breakeven there!
- As both Zeno and cmdeal have said, your payout's pretty challenging on FB. It's possible to get single-cent / two-cent clicks, but far from common, and looking at your EPC that's what you'd need to be profitable currently. Try having a chat with your AM and seeing if there's any way you could get a higher payout. Probably not - app installs don't usually have much room for that - but it's worth asking.
- Failing that, I'd recommend trying a new offer, or seeing if this offer is available from any other networks with a higher payout.

You're actually doing pretty well here! Some of these ads are looking very promising, and you're seeing plenty of conversions, which is the first hurdle.

Looking forward to seeing how this develops!

---

Hey CMdeal, hope you may answer some question that I'm curious about

1) What verticals that you spend $xx,xxx a day on FB?
2) Did u build a list? Or direct them to a LP then to CPA offer?
3) How much payout you think is suitable on FB?

---

Home > Paid Traffic Sources > Facebook & Instagram