Spend 5-10x of the payout or spend more? (6)
Hi
I’ve been testing different things to see what can work for me or not. Maybe I have taking many wrong steps not having much experience, but you learn with everything you do and mistakes are a part of that. I’m a bit confused about some follow along and the way many affiliates approach their campaigns. They basically take one of these steps:
- Test campaigns until they spend a 5x or 10x of the payout of an offer, and if they don’t see conversions, stop the offer and go for a new one.
- Try to make the campaign profitable whatever it takes. That means running the campaign (offers) for at least a week and spending much more than the payout. Even if the offers have a low payment, some affiliate spends 80 – 100 dollars daily to see if they can make them profitable.
What would be the best approach? I’m saying this because I’m seeing that its usual to spend more than the 5x – 10x payout on mobile offers. Maybe it would be better to focus more on one campaign with x number of offers and spend a bigger amount on money to gather data and make decisions. What do you think? Thank you.
I remember vortex mentioning that if you have landers (rather than direct linking), then you should go above the 5-10x payout, which makes sense since there are more variables to optimise. But if direct linking, then stick with 5-10x payout.
Not sure where that's posted, so maybe she can chime in?
It's an important question to ask, and every time I explain this I'm happy for yet another chance to clarify.
It's exactly as Roland said: The decision would depend on how much room there is for optimization.
If you're doing direct-linking and targeting carrier traffic, you really won't want to spend too much money testing each one, due to these reasons:
-There's a lot less room to optimize by cutting bot-infested placements, because there's next to no bot traffic (compared with wifi).
-There's no room to improve your funnel through testing additional landers or tweaking the best lander.
-Because it's so easy to direct-link, there are probably a lot of people doing this, so basically you need to test a ton of offers to find profits. So if you spend more than a little bit on testing each offer, you'll lose your shirt - your next profitable campaign may not even make back the money you spent on testing so many offers.
Basically, when it comes to direct-linking carrier offers, you only have 2 main variables to optimize aside from testing the offers themselves: 1)cut placements (and maybe OSs and browsers), and 2)test bids. And in many cases, these will not lead to the potentially massive increases in ROI that you can achieve by cutting bot placements for wifi traffic, or testing landers.
As such, if an offer doesn't perform well from the start, then chances are slim that it will become profitable with the limited optimization. This is why you can use the 5x-payout rule.
Whereas, if you were running with landers and targeting wifi traffic, the additional room for optimization will justify and require additional spend.
-If you're targeting wifi traffic, that in itself would necessitate cutting more placements than if you were running carrier traffic, due to bots being more rampant. And it takes spending money to identify and weed out bad placements.
-If you're using landers, then you'll need to spend money on cutting landers down to a winner, which will require money. Then, when you use that winning lander to mass-test offers after that, that will require additional test budget. And when cutting offers, you can't even go by the 5x payout rule - because the camp may be profitable after you cut placements!
This is why you can't use an x-times payout rule to cut offers when using landers and running wifi traffic. An offer that doesn't seem to perform well from the start, may very well become a winner after 1)you've cut down to a good lander, and 2)you've cut some junk placements.
In this case, it would be better to cut inferior offers when each reaches statistical significance. Then see if the best offer can make enough of the total traffic profitable (if so, cut unprofitable traffic; if not, test more offers/landers to improve the funnel further before trying again).
Hope that makes it more clear! In the end it's all about assessing the likelihood of profitability. Only continue to spend money for as long as you feel the campaign still has a good chance of ending up profitable.
Amy
Thank you both for the great answers. I have a more clear idea where I should spend more money on testing. I can see the variables are huge and that’s why having a good strategy that works is a must. Does this apply with other types of traffic? For example, in search or social traffic (Adwords, Bing, FB, etc) there are no bots I suppose. Is it justified to spend more on testing using other types of offers/traffic sources?
Originally Posted by bachroll
Basically, the more variables there are, the more room for optimization. The more room for optimization, the bigger the test budget you need to spend.
Then, there are variables you can test ONCE, find the optimal setting, then apply that to future camps. e.g. Advertising angles that work best for your particular offer/niche. e.g. Which keywords (on Adwords) or demographic (on FB) convert the best for your offer/niche.
Of course, if you change your niche, you'll need to test new angles and keywords / demographics all over again.
And then of course there are variables that you need to test with every single camp, such as new ad text/banners/images, or bids.
The bird's eye view answer to your question though, can be pretty much summarized as follows:
Only continue to spend money for as long as you feel the campaign still has a good chance of ending up profitable. (I'm repeating myself here. Example here is on pop, when you've tested most of the popular landers and many offers for the vertical+geo, and most of the placements are still in loss for the best lander+offer combo, then there would be little point in continuing.)
The more profits potential the campaign has, the more money you can justify investing in it. (For example, say you were running a somewhat broad-appeal offer on FB, where you can target a big demographic, to potentially make 5-6 figures/day, then investing a lot of money on testing upfront would make sense.)
Hopefully I've provided more food for thought. It's always tricky trying to provide a one-size-fits-all answer to general questions.
Amy
Originally Posted by vortex
Thank you Amy. I know the question was complicated to answer being the possible variables so broad. I think I get the general idea. I'll take it in mind for my next campaigns. Have a great weekend!
Home > The Newbie Zone > Questions and Answers