Trying to nail down sweepstakes (10)
Let me give you a few tips and observations 
1. the data is agregated by using all the landers and angles, I mean, is this a mix? In case it is, do you see some better working combinations in there or is everything perforing about the same? In case there are some better combos, try running them again isolated.
2. you say you are bidding low. I have seriously seen EPVs double as a result of a higher bid, definitely test this. Dont jump to the highest bidder position straight away tho, do it one step at a time. This might also solve your volume problems. In case the higher bid helps, I recomend taking a second shot with the blacklisted sources, some of them might start working now.
3. The CTR is hard to judge, it seems low, but sometimes lower CTR means way higher CVR. Only you can see if the lower % is a problem or not. It can definitely go up a bit, it always can. Play around with the imagery, it almost always helps. Use a better/larger photo of the price they can win, or sharpen it ... these are easy tweaks that can help.
4. Dont worry about the EPV expectations that someone has, the only thing that is important to you, is that its higher than your CPC
5. Dont worry about the scalability of your first campaigns too much either, you're learning now, as soon as you hit something, you can scale to different GEOS, similar offers etc ... for now focus on the optimization process and the whole workflow. There is no point in thinking about $500 per day campaigns when you dont have the $50 per day running yet
Good luck
Originally Posted by matuloo
1. the data is agregated by using all the landers and angles, I mean, is this a mix? In case it is, do you see some better working combinations in there or is everything perforing about the same? In case there are some better combos, try running them again isolated.
2. you say you are bidding low. I have seriously seen EPVs double as a result of a higher bid, definitely test this. Dont jump to the highest bidder position straight away tho, do it one step at a time. This might also solve your volume problems. In case the higher bid helps, I recomend taking a second shot with the blacklisted sources, some of them might start working now.
3. The CTR is hard to judge, it seems low, but sometimes lower CTR means way higher CVR. Only you can see if the lower % is a problem or not. It can definitely go up a bit, it always can. Play around with the imagery, it almost always helps. Use a better/larger photo of the price they can win, or sharpen it ... these are easy tweaks that can help.
4. Dont worry about the EPV expectations that someone has, the only thing that is important to you, is that its higher than your CPC
5. Dont worry about the scalability of your first campaigns too much either, you're learning now, as soon as you hit something, you can scale to different GEOS, similar offers etc ... for now focus on the optimization process and the whole workflow. There is no point in thinking about $500 per day campaigns when you dont have the $50 per day running yet
Good luck
As a follow on question, then how significant are minor changes in landers like larger images, different font colors etc. because obviously there are many iterations like this for each angle I go for, but I'm not sure i have the "winning" angle yet so am hesitant to sink too much time into these iterations lf landing page.
The smallest changes can sometimes have the most ridiculous performance increases.
Assuming you're doing mobile:
Font colours and images affect readability - a colour that looks good on your computer screen isn't necessary readable on a cellphone, and a smaller image may make your text fit better on the user's phone.
One or two words added or removed can double your CV (from experience!)
To answer your iteration question, I usually try to test a bunch of wildly different stuff at first. After a few days of data you usually know where to narrow down to for more granular testing.
Discarding the clear losers before going deep makes your life easier (and cheaper): it's probably not wise to test 3 variations of each of 5 landers all at the same time, but 2-3 variations each of the best 2 landers is pretty manageable.
Here is a example...I recently ran a campaign promoting sweeps(gift card). On 1 lander I had a image of th gift card and on another lander I had a cartoon version of Santa holding the gift card. That's all it took to boost the cv almost additional 1%.
STM community so helpful, thanks everyone so far.
To further along the discussion, another issue I seem to have is balancing the number of iteration versus the the traffic I get. Often I simply don't have enough traffic to be testing that many landers hence I am hesitant to add too many.
Anyone have a good rule of thumb or a solution to this? I know there are some methods that are used in machine learning to balance out the number of trials to reach optimization but would not go down that path yet since I like to adhere to KISS (keep it simple stupid)
I'm not a big expert in this yet, but I had the same dilemma as you not too long ago, so I thought I'll chip in.
Initially I wanted to test as much as possible, as many offers, as many landers. But at the end I just got confused the hell out, and more landers = more budget.
I would advise you testing out 4 landers initially. Make sure they are all drastically different. Not just a simple word change or a font change. If you've got WAY more budget, feel free to add in more. Or once you get an idea, you can test more.
But keeping it at a relatively small number initially helped me stay way more focused. Paying way more attention to each lander's stats, rather than getting overwhelmed with figures from so many landers.
As others have suggested already, you should start by testing VERY different landers and only once you are sure one or two approaches seem to be the right ones, then move to minor changes.
How big difference the small changes can do? Webdev gave you a nice answer, it can do wonders. But to be fair, these small chages usually dont change -50% campaigns into +50% ones.
Its a matter of testing tho, sometimes the results are really unbelieveable. I have seen banner CTR improve by 30% just by changing the CTA button color. These are really hit or miss things, but ones not to be overlooked.
First of all, dont worry much about CR, CV, CTR etc ... ROI is the only number that really matters and in case the numbers are correct, I see very healthy ROI with some of those campaigns.
But I understand the temptation to improve the CR a bit and earn more, so lets take a look at some simple steps you could do to achieve that.
1. Assuming that you are still running sweeps - try to play with the wording - the closer you can get to the "you won" wording instead of "you might", the bigger the CVR will be. This is skating on the edge tho, so be careful
2. CTA - Im sure you have some CTA on the LP, try to make it bigger, different color, move it more to the top - these are things that usually can help with CTR and more traffic to the offer CAN mean more revenue.
3. I've recently ran a few tests with domain names and looks like GEO related domain has a small positive effect on the campaigns. Try to register a domain with some word related to the country and put the LP on it.
4. Can you have entry JS alert on the LP according to the traffic networks rules? This can help for sure.
5. More scriptage - backbutton redirect, popunder on exit etc ... if your traffic sources allow them, they should add a few % to your ROI.
These are just a few I could come up with really quickly, Im sure you will get more tips from others
Good luck and congrats to your progress !!!
Thanks Matuloo, helpful as usual
Anyways been experimenting around with some of these (as I was implementing some of these prior to my post but haven't gotten the results in). Just some thoughts thus far on your pointers that I thought I'd share:
1. Am a bit lazy with copy to be honest since I believe in the 80/20 rule and want to make it so that I can use a 1 size fit all offers LP and with different geos being varying levels of strictness I thought the difficulty to manage the copy and multiple LPs didn't seem worth the time. But if you or anyone else disagrees and that the copy can bring about a significant improvement would love to hear it. (though we explore this once I reach a higher threshold where small improvements have noticeable improvement to bottom line).
2. Mixed results here, nothing too conclusive still testing as the larger and flashy CTA generate more clicks but not necessarily more conversions.
3. This is interesting, same as point 1 would look into once I generate more traffic.
4. I have all the standard JS alerts (entry, exit) or whatever I can run. Definitely improvements on CTR.
5. Agree with this the current ones I've tested are: backbutton redirect (though with popups not sure if they work), entry/exit pop, audio (though on mobile this doesn't work). All bring about a positive result. If I've missed anything out let me know always happy to try.
Understand the general mantra is the more aggressive you can be the better the results. Though having done a lot more testing this week I've noticed that although my stats seem low but it is also due to the traffic I am getting i.e. if I whitelist and filter I get significantly better results which has put me at the conclusion that my LPs are okay and hence for now I've put a halt again on focusing on LPs. But admittedly my results have been declining, not sure if its due to the competition or the deterioration of the offers themselves, guess time will tell.
TL DR: Read Matuloo's post above if you want suggestions on improving LPs for sweeps though quality of traffic can significantly skew your results thus if your LPs looks weak it could also be due to traffic and not the LP itself.
Originally Posted by sushiparlour
I understand the desire to have a one fits all solution, managing multiple LPs across a ton of campaigns can be extremely stresful and frustrating, but thats how it is in AM. You will have a hard time finding one thing that will work in the whole world. There are situations where it can work - the battery saver or speed booster apps for android were a perfect example of this, the same angles and LPs smashed it across all GEOs pretty much. But more often than not, it will probably work the other way around - you probably noticed I talk about adult dating a lot on this forum, and thats exactly a market where you cannot target more GEOs with the same angle and LP style. Same with many sweeps offers like the win a voucher for TESCO - you cant really use it in a GEO where Tesco isnt popular or its not there a lot - in this case you could most likely replace just the shop name and still use the same adcopy so it would be close to what you are trying to achieve tho.
To put it in a simple way - there are GEO specific aspects that you need to consider when planning your campaigns and they can have huge effect on your success or failure. Searching for shortcuts and the easy solutions can sometimes work, but you shouldnt always count on those. Dont be lazy
Home > Paid Traffic Sources > POP / PPV / Redirect