When to cull a campaign? (4)
Hey guys,
I just wanted to get some feedback on my understanding of how to find a winning campaign (banner traffic)and optimize it, ive read heaps of the forums and just want to see if ive understood the gist of it correctly.
So my understanding is that the rule of thumb is 9 out of 10 campaigns will fail and have no hope, so our job is to root these out and make way for the 1 in 10 that will be able to make profit. To do this we launch and collect data. If, after a decent adspend (for an app install of $0.25-$0.60 i'll give it a good run for at least 4 to 5 days with an ad spend between $60-$100 all up) the campaign doesn't have a -50% or better ROI then it's probably not gonna be profitable and should be culled. Do this enough times (whilst improving and learning along the way) and you'll start to find a few campaigns that, right off the bat, are getting ROI's of -30% or -40% (or even better), and these are the one's that you put your time and effort into and have a very good chance of becoming profitable?
I've had my fair share of -90% or -80% ROI campaigns straight off the bat, i kill a few placements, kill a few banners, and then wait....and it improves a bit but i just kind of hit a wall there, and then hit the reset button and go after a new campaign.
Also, if you're culling a campaign is it worth trying the same offer again or are most people better off starting from scratch with a clean slate and a new offer?
Thanks!
IMO, deciding whether to cut an offer and when, will largely depend on the particular type of offer you're promoting, as well as how it fits into your overall test strategy.
The first questions I'd ask is, when the offer is pulled will I be able to easily find a replacement offer to take its place?
If the answer is YES - e.g. if the offer is in an "evergreen" niche such as adult dating, weightloss, sweeps, antivirus etc. - then I wouldn't focus on ROI at all in the beginning. I would still test a few offers in the beginning to find one that converts relatively well, but once I do have a decent offer, I would mass-test banners and landers to optimize them. The reason I wouldn't care about ROI is because I know that with a lot of testing, my creatives WILL improve and eventually reach green. And once I have good creatives I could mass-test different offers in the same geo and even for other geos (if I keep the angles generic and not geo-specific). Once I "crack" this niche I could bank for a long time (knowing that even if the offer gets pulled that I could find others to take its place), so all the investment in time and money would be worth it. Therefore I wouldn't care how bad the initial ROI was (as long as the offer is decent compared to similar offers), or how long it would take me to get the camp green.
On the other hand, if the answer is NO - e.g. if the offer gets pulled I probably won't be able to continue using the same creatives on a replacement offer (many mobile app install offers belong in this category) - then I know I'd need to test and scale quickly, in order to make enough profits to justify my efforts before the offer gets pulled. In this case, I would keep testing offer after offer until I find one I deem promising enough to continue. I would try my reasonable best to brainstorm many angles, pick at least 3-5 angles I feel would work best, make at least 3 banners per angle - basically give it my best shot. If it's not even close to breaking even then I'd just move onto the next offer. This is where rules of thumb like cutting an offer if it doesn't do at least -50% ROI right off the bat can come in handy.
The reason why I wouldn't spend a lot of time and effort to optimize my creatives in this case, is because I wouldn't have the time to do so, as most mobile app install offers are short-lived. I wouldn't want to be in a position where the offer gets pulled after I've wasted time and effort on optimizing creatives, but before I've had a chance to make back my investment.
Also, if you're culling a campaign is it worth trying the same offer again or are most people better off starting from scratch with a clean slate and a new offer?
It's always easier to take something that works out of the gate and make it work even better, than to take something that's really far from breaking even and try to make it profitable. This is also why it's important to give every offer your best shot when coming up with angles and ad text and making banners - so that when the initial stats don't look promising you can say to yourself, "I've already tried my reasonable best - let's move onto another offer". It's important to not rush through the research and brainstorming stage - otherwise, when you see bad results you'd be left thinking, "would I have made this offer work if I'd tried harder?"Hope that answers your question!
Amy
while most of the campaigns are going to fail, i think with experience, u could pick one more campaign out of 10 that could possibly work, maybe even better than the one that worked right off the bat. the standard banner and placement cuts definitely have to be done, but wont make your campaign any better than anyone else who already knows this, which is everyone.
1 thing i always do with my campaigns as long as there is some kind of -70% ROI, is to do some POTENTIAL ANALYSIS. chances are you hit many things wrong, wrong bid, wrong offer, wrong targeting, but somehow something went right. so we want to investigate if we could do some important things better, how much better could it have been? these are all guesses, and they get better with experience. and you wont really know if there is potential to improve, but chances are if you picked the right variables there is always room for improvement, and also by how much?
lets start with a few assumptions - we want a campaign that after optimization, will retain at least 30-40% or more of the traffic we started with. it should be reasonable for a single person to maintain within 2-3 hours in a day. data should be coming in fast enough, significant results within 1-2 days usually.
lets start with a case here of a -70% roi..
1. offer - how many more offers can you test from the same affiliate network? they may not be exactly the same, but somewhat similar, the more offers u can test, the higher the potential.
2. how many affiliate networks have this same offer? the more the number of affiliate networks the higher the potential
3. how developed are your ads in this country/offer? how much research and testing have you done here? the lower your experience, the higher the potential
4. how much have you tested the bids? how is the competition like? how much traffic is there in this geo? the larger the traffic in this geo, the less the bid testing, the less the competition, the stronger the potential
5. are there portions of your traffic that are performing significantly different? there could be an offer, banner, lander that is outperforming the rest or sucked alot of traffic and totally died.
and then roughly estimate how it could work out. if u work with a simple campaign that is almost breakeven with normal ads, chances are others are doing the same. we will definitely do this and take the immediate $$$ on the plate. but usually these campaigns die out alot faster. but if you dig deeper into campaigns that seems a little confusing and performing differently across many variables, your competition significantly drops, your campaigns last longer, you build up your skills, your negotiating power increases for the offers you are promoting because less people are promoting them, and when it shows up in the top charts of the affiliate networks, the competition comes in a little slower than the standard du speed du battery 360 campaigns
@Vortex - Thanks so much Amy, lots of value right there! That's a higher level way of looking at it that i hadn't really thought about, i've always been offer focused and never really considered a more broad, niche wide approach, i'll take this all into consideration for my next campaigns
Too true, i've always generally had the -80%to-90% campaigns and working them up from there is a real uphill battle, i'll put more focus on finding those one's that work right from the get-go.
@xavierfok - Wow, thanks for that Xavier! I've never considered potential analysis, but yeah there's obviously heaps of different variables out there to test and many small things that can go wrong and sink a campaign. So what you're saying is that the campaigns that have strange combinations (eg work really well on with users who are all Android, T-mobile and site placement but bombs with everything else) have greater longevity and less competition since they're a lot harder to find?
Home > Paid Traffic Sources > Mobile