How Much Should I Spend Testing (9)
Is there a certain formula or recommended amount?
There is no formula but typically people will spend about 3 - 5 times the offer payout per placement if you're doing PPV to get a good read.
Thanks for the help, would this recommendation be the same for facebook, pof, etc?
If you're just testing the offer to see if it converts and if you're ads aren't "overly clever" and direct then yes.
Seems to me that most affiliates do very little split testing while running a test campaign, either it works for them or it doesn't. I look at it more as can we continue to make progress, optimizing to a certain CPA? There's great arguments on both sides so there's no right or wrong but for me a test is a lot more extensive.
Just as an example, I have a self serve network we use where a $5 lead gen off a presell is paying me $2.10 for every $1 I spend, but a $40 bizop that does great for me in a ton of buys couldn't get below $52 per acquisition. Shit like that happens all the time depending on the traffic. Far as I know most guys haven't even tested low payout offers off an advertorial or any kind of presell, but in certain cases it will save a campaign.
With CPV, did you add audio? video? if you already had them did you take them off? Sometimes audio decreases my conversion. Have you tried adding social proof like facebook likes? Exit pops? Sitescout adserver has a great exit pop block you can create with multiple offers, you could even do something like that on your own with regular HTML.
There are so many elements that can save a campaign by lowering your CPA, it can be very frustrating but if a campaign has signs of life, I personally won't trash it unless the numbers are absurdly off, I'll test some other variations and see if I can make quick progress.
Other things like going direct play a role for me as well. I ran an ecig offer for awhile that sucked on a lot of traffic sources, couldn't make it work. Ended up going direct to the advertiser and making $12 more per conversion off the bat which changed everything.
There's just a lot to think about in this game, and I see most taking a lazy throw and stick approach without actually learning anything.
I spend up to ~$500 per offer, testing on Facebook - I'm bound to see profitability.
Another thing to consider is the popularity of the offer you are testing. If you are running on Facebook, and you are doing dating offers for example, you might want to spend more in testing because you know for a fact that people are banking with them. You see them everywhere. On the other hand if you are trying something completely out of left field and it's a toss up whether it will even work at all on a given traffic source, then you may want to spend less. Really it's up to your tolerance for losing money. As long as you are learning something, like what doesn't work, then you can keep optimizing in hopes of profitability. I've spent $20 testing a campaign and called it quits before, and I've also dumped in over $1,000 on a campaign before it finally broke open and took off... So the rule is there are no rules.
@peraction - Great description of the process, thanks.
There are so many variables you can tweak if something is borderline. I had a campaign on Adwords that took me a couple months to make profitable (I could only afford to spend like $150 a day so that's why it took so long). I tried a bunch of landing pages, advertisers/offers, went direct, split-test networks, etc. and then spent a lot of money getting the necessary data to find the pockets of profit. I tracked everything, and waited. When I felt like I was nearing my tolerance threshold for losing money, I let the data do the talking. Once optimized, my EPCs were more than double and the campaign became profitable. But if I hadn't negotiated better rates with an advertiser, gone direct with another advertiser (there were multiple offers involved in my site), and split-tested significantly different landing pages, I could easily have given it up as lost.
There have been other times where I've spent $500 testing something on Facebook that I KNOW people are making work, and I just couldn't get it to work. In retrospect I think that my targeting was too broad and my angles were too lame. My split-testing wasn't wide enough. I was testing apple 1 vs. apple 2, not apple vs. orange, etc.
I think one of the main competitive advantages of the bigger/more successful affiliates is that they can afford to lose more money testing, and stick with a so-so campaign longer trying to make it work. At the same time, thorough testing is probably what made them successful in the first place...
Originally Posted by brianb
Home > Questions and Answers > General Questions