In your experience, which method is better and why? (6)
In your experience, which method is better and why?
Scenario: One banner has 10k clicks, making 0.026 EPC, cpc is 0.021 - other banners for the exact same angle, are all under that based on significant data... (for the sake of this post, banner1 has 0.016 epc, banner02 has 0.025 epc, banner03 has 0.026 epc, banner 0.024 has 0.020 epc)
Is it best to optimize a campaign and then and leave it as is to continue making money -- and use its elements to create brand new campaigns?
-Or- is it best to isolate the best banner, upload new variations of the best banner (diff borders, font color, images, etc) into the existing campaign, and work it that way?
I would always advice to keep optimizing your campaign.
When you find combo's that are turning a profit you should be banging out more of those combo's to
A: Make more ROI B: To keep the campaign fresh with new materials.
And once you are comfortable enough with the ROI and it has been stable for a few days you will want to scale to the max, but keep optimizing to keep it fresh and to keep those ROIs up.
You can isolate the best banner and still do what I explained above but if you have more banners/combo's that are close to profits or little profits I would also optimize those to broader the campaign as you might end up with a couple of winning angles/combo's instead of just one.
At the end of the day; the more angles that work you can find the better it is. Could even rotate those to avoid it running out to fast or run completely separate campaigns.
What's the offer payout? My advice would be very different depending on the payout ( although I guess it much be pretty small for the epcs to be that granular ).
What I'd probably do in that situation is keep running everything unless the minimum likely epc is at or below break even (using a stats calculator), but also add new banners.
Some of the new banners would be based on the best old ones, some would be totally new. I'd keep optimising and trying to drive epc up from there.
Of course, this is very traffic source dependent. How I'd treat the campaign if it was a low-login pof camp vs, say, a mobile camp where I hadn't optimised carriers and handsets yet would be very different.
Originally Posted by caurmen
In the event where majority of the banners will be <cpc after significant data has been collected, I will make variations of the only banner that is positive, and add them in to the campaign itself.
When I have lets say 4 banners, which are equal, but one jumps out huge. Lets say banner1,2,3 all have epc's 0.02x, and banner 4 has epc 0.03x I will duplicate campaign, taking with it the excluded placements/carriers/devices/etc and bid higher CPC, and load in variations of banner4 -- to test to see which one renders the highest EPC out of them and work it from there.
I am about to test another idea, taking banner4, making a brand spanking new campaign - but I am not going to use the excluded devices/carriers/placements, etc -- i will let it run from scratch -- my theory is that, this is redundant, the shitty placements that were -30% below cvr needed to break even at the CPC being paid, will remain shitty even if I start from scratch. That is the theory, will test and see what the data says though.
I don't like messing around too much with my campaign if it's profitable. I might just pause banners and upload a few new ones with the same style to keep it going. And also rotate landing pages or just test landing pages. Then I make a new campaign to test a series of new banners and such.
Always optimize otherwise 1 or more parts of your campaign will become stale (seen too often) and your conversions will hurt!
Home > General > Affiliate Marketing Forum