Mobile follow along (nothing held back), looking for advice from seasoned AM's! (11)
You haven't spent anywhere near enough yet to be making changes, I'm afraid.
Optimising for CPC is all very well, but it doesn't actually make you money. You need to optimise primarily for conversion rate - and in order to do that, you'll need to run each campaign for a good while longer.
Additionally, the payout for this campaign is kind of high for a mobile campaign - is there a lower-payout campaign you could try running? That'll help you get more data faster, in all likelihood.
Originally Posted by caurmen
I could choose a different offer and will probably start testing one with a lower payout. But my AM recommended this offer and from what I've read the main issue with testing higher payout offers is budget.
What kind of budget would I need for testing this offer and also, at what point should I start making changes for a payout at $2.64?
For a $2.64 offer, I'd probably run to about $25 spend on each of the 4 app/site/carrier/wifi combinations. I'd kill non-converting placements at $7.50 spend per placement, but otherwise just leave everything alone and see what the data looked like after that initial spend.
Total budget to optimise to profit is impossible to say, but you'd want to budget at least $300 at a very, very rough estimate. Obviously, you could end up in profit after the first $100, but you could also end up spending a looooot more...
Originally Posted by caurmen
Got it. Thanks for the reply.
I took your advice and am running 2 other offers at lower price points and testing on decisive now. I will post stats once the spend is where it needs to be.
By looking at your images I see banners matching with the following offers:
Battery booster
ICQ
Speed booster
Dolphin browser
Antivirus
Download
Sure you can put a lot of angles on Dolphin browser but it is not going to work IMO when your banner mentions "Install Battery saver" and has the logo of DU battery saver. Did you find any good placements (Apps and Site tabs in Decisive) ? This is the first thing I would look at after running an initial test.
I'm not really sure what to make out of the data you have. I would either move on to a new offer or retest this offer. This is what I would do differently:
- Properly prepare the creatives: in the target language only, make sure they match the offer (if its not a generic banner like 'Download now'), don't test too many banners during the first test (to avoid scattered results). Your initial goal with the banners is to find a good angle. When you find a good angle you can make matching LPs and more banners. I like mixing in the 'Download now' banner but these usually only work with a LP
- Test 1 country at a time
- No need to split your campaign in Wifi/Site, 3G/App etc. If you put everything in one campaign it will be easier to analyze the placements. Once the campaign is optimized on placements and creatives you can analyze other metrics like connection type, device, OS version etc.
Aside from what keepthecar mentioned, I think there are several critical issues here:
1. The large % of placements being undefined really cripples the optimisation process.
2. You apparently went into this without a clear optimisation plan - or at least didn't execute it in a timely manner.
From your data I can tell your need to be much more aggressive about optimisation - or at least more diligent about it. If you don't have a plan, process, system or even a checklist written down, you are just going to burn money on mobile.
You have a tonne of placements that have spent 5-20x the offer payout without a single conversion. Why are they still active?
You also have plenty of placements that have spent 2-5x the offer payout and have terrible CTRs == very high effective CPCs compared to the offer payout. You are unlikely to make money on a placement that costs $0.10 per click when your $0.30 payout offer is converting at 5%.
Based on the overall performance of the campaign I would ditch it, but the issue is you've kind of crippled your stats by polluting them with far too much data that you should have never bought!
If you want to test more I would break out a new campaign with the better banners, focus on the quartet(s) that looked most promising, and immediately blacklist bad placements and things like Grindr that will eat your budget.
Oh, that explains that - never mind!
Originally Posted by zeno
Thanks for the feedback. I didn't touch the campaign because I was under the impression that until the spend was completed that blacklisting shouldn't be done...What you're saying makes more sense, however.
Regardless, I am starting a couple new campaigns and will update everyone with the stats as they occur.
A couple questions before the next run:
1. keepthecar recommended I not seperate everything into 4 campaign,s which also makes sense. However, in the original mobile cookbook guide it recommended spending a minimum of $15 across all 4 campaigns, totaling $60. If I am bundling everything on 1 campaign, is $15 enough or should I collect $60 worth of data?
2. Zeno, you mention having a clear optimization plan prior to executing. I am happy to take this advice on the next round, but I assume you mean actively blacklisting as campaign spends exceed certain percentages vs. conversions. Is there anything else you'd recommend doing prior to executing a campaign to properly prepare to optimize besides blacklisting after the fact?
3. In regards to blacklisting undefined Apps prior to them eating up the spend, is there a way to immediately identify and blacklist the "undefined" category within decisive or do you have to wait until a conversion pops up to do so? I am just wondering because there are something like 40 pages of apps to sift through. Just wondering if there is a more efficient way to immediately stop that from happening.
Thanks again for your feedback. Back to the drawing board!
1. There are different approaches. You could lump it all together > optimise placements (app and site!) then split it. I would normally recommend splitting carrier and WiFi as they can perform quite differently on the same placement. It's probably better to follow the cookbook for now just to keep things consistent.
2. Yep, actively blacklisting placements as they breach certain conditions, e.g. over x spend with no clicks, over y spend with clicks but no conversions, etc. Stuff like 3x offer payout with no conversions > kill, spent 1/4 offer payout with no clicks > kill, etc.
3. Keepthecar mentioned that the undefined part may come from the app / site crossover, e.g. if you are targeting apps and sites, then in the apps tab the site traffic appears as undefined and vice versa. You should check that this is not the case for your campaign. If the undefined is definitely from unidentified placements, ask Decisive support about it. I have a feeling a lot of people get unnecessarily confused by this when they target apps and sites together.
One other thing - are you targeting Brazil with the English and Spanish creatives, or just the Portuguese ones?
Brazil is 90%+ Portuguese speaking. I'd make sure that the translations that you're running in Brazil are ALL accurately-translated Portugese - otherwise people won't even be able to understand your ads!
Mexico, likewise, won't speak Portuguese as a rule (Portuguese and Spanish are similar-ish, but not THAT similar). Running Portuguese ads there is pretty much throwing money away. Likewise, only 5% of the population speaks English.
(Also, make sure you're using the right version of Spanish. Mexican Spanish and European Spanish are not identical. )
Home > Mobile > Follow-along Campaigns