Optimizing Your Campaigns

Personally, I’m a huge fan of taking one campaign and targeting a single demographic to minimize my testing costs. For example, if I was testing SingleParentsMeet I would start by testing Single Women in the US, with Kids, aged 25-29, Login Count ≤ 50, Session Depth ≤ 20, Ads Clicked ≠ 20. I would be split testing and optimizing this campaign until I got it profitable. I would split test direct linking versus using a lander (depending on your angle), including CTA’s, background color, header, etc… on your lander, split test offers, split test images and of course, split test ad copy. Only test one thing at a time. Plus, you need to get enough data to be statistically significant, which I discuss further down this page. Once you’ve gotten your campaign profitable and, you can scale your campaign, which we’ll discuss in the next section.

Some affiliates preach that you should start by setting up a handful of campaigns and testing every demographics. For example, you would target Single Women in the US, with Kids, aged 18-24, a campaign for those aged 25-29, another campaign for women aged 30-34, yet another campaign for women who are 35-39, and finally one that targets Single Women in the US, with Kids, aged 40-49, but I don’t think that’s necessary when testing. Sure every demographic will be a little different, but you could save yourself hundreds or even thousands in testing costs doing it my way. I like doing it my way so that I can minimize my testing costs, but everybody has their own method to their madness.

The best way to learn affiliate marketing is to jump in, hands on, head first. Once you’re getting data, you have to consider it knowledge of what works and what doesn’t. This is the only true way to learn affiliate marketing.

Optimizing Targeting

The first thing you’ll want to do is take advantage of the many reporting features that POF gives you, especially the Conversion Reports. The Conversion Reports will tell you exactly which criteria are and are not converting. This data will only be available if you place the POF pixel on your offer.

In my opinion, once you have a few hundred clicks on your campaign, you can start getting rid of non-performing criteria. I suggest a few hundred clicks because it’s hard to optimize off criteria that may only have 10 or 15 clicks. It’s hard to judge anything with that amount of data. On the other end of the spectrum, if you have any criteria that are just converting like crazy, try creating a separate campaign. Making your targeting very specific, for that criteria only, tailor your ads appropriately, bid higher and get all the volume available! I’ve done a couple case studies where I do exactly that and I’ve included links to them below.

POF Conversion Tracking for Profitability – Case Studies – Research – POFPrimer.com
POF Conversion Tracking to Weed out Non-Converters – Case Studies – Research – POFPrimer.com

Optimizing Offers

If you’re split testing offers, most people agree that you can judge how well an offer converts by sending 100 clicks to an offer, but I prefer to send more like 200 clicks to an offer before deciding which offer converts best. Unless one converts at 2% versus 10%, that’s obviously a no-brainer. It’s easier to judge a high converting offer with 100 clicks than it is to judge a low converting offer with 100 clicks.

Optimizing Images

One of the most crucial parts in optimizing doesn’t happen with tweaking and optimizing your targeting. You want to be sure that you’re tracking as much data as possible, especially the image. An image could have super high CTR, but horrible conversion rates. Think images (hint: girls with huge tits) that draw your attention and is only clicked out of curiosity. These usually don’t have the best conversion rates, but if the CTR is high enough, it can still be profitable.

If you’re using POF’s conversion tracking pixel, you can check your Ad & Image Report, to find out which images are converting best. You should be tracking this with your tracking solution anyways. I go over the Ad & Image Report more in-depth on the reports page.

I’ve included a few commonly accepted practices you can put to use below. Again, everybody has their own specific way of doing things. I’ll include what I do in parenthesis.

  • Let the image run 2000 – 2500 impressions before pausing it for low CTR (I only do 1000-1500)
  • CTR Goals: .15+ is good on POF (I shoot for .2+ targeting males and .1+ targeting females)
  • If the ad is getting a good ctr, but no conversions, let it run 2-3x the offer’s payout before you kill if it’s not converting
  • Whatever you do, don’t kill an ad after 5 clicks. Give the ad an opportunity to perform.

Optimizing Ad Copy

This is another area I prefer to do overkill on. From what I can tell, most people like to optimize off ad copy if its not performing well anywhere from 20-50 clicks per ad copy. I like to give it at least 100 clicks per ad copy to figure out works. I usually start by testing only 2-3 ad copies.

Again, If you’re using POF’s conversion tracking pixel, you can check your Ad & Image Report, to find out which ad copy is converting best. And again, hopefully you’re already tracking ad copy with your tracking solution.

Dayparting

If you’re running a pretty dcent amount of volume through an offer and have received a large number of clicks, something you will definitely be interested in is Dayparting your campaigns. If you’re using Prosper202 to track your campaigns they have a feature specifically for this. It’s under the Overview tab, the third option on that submenu. You’re able to figure out what times of the day convert best for your campaigns. This way you can pause your ads when they aren’t making you money and only run them when they are profitable. Unfortunately, I don’t know of any type of script that will automate this for you.

Weekparting

Same ordeal as dayparting, except this is just figuring out which day(s) of the week are profitable for you. After advertising daily in the dating niche for the last three years, I’ve gotten a pretty good feel for how well each day converts. I’ve covered every day below. These judgments are for dating campaigns only and this will vary by vertical.

  • Sunday – Far and away the best converting day for dating campaigns.
  • Monday – Hit or miss when it comes to converting well for dating.
  • Tuesday – One of the worst days for dating campaigns.
  • Wednesday – Another one of the worst days for dating campaigns.
  • Thursday – Better than Tuesday/Wednesday, but Thursday, too, is hit or miss when it comes to conversions.
  • Friday – Conversions will pick up and Friday is a great day to run dating campaigns.
  • Saturday – Another great day to run dating campaigns, usually the 2nd best day of the week.

Comments

  1. Hey Riley,

    You mention the way you test is basically you pick a targeted demo and go at it until it's profitable. At what point or what amount of money do you stop testing and tweaking and make the decision that a campaign just isn't gonna work?

    Thanks

    • rileypool says:

      I don't really have a set amount. I'll continuously be getting rid of demos or criteria that are not profitable for me. Once I get it narrowed down to what has been converting best for me (in that campaign) and I still don't have it profitable then I'll give up. Additionally, if I narrow the targeting so much that I don't get enough volume or the return isn't worth my time then I pause it.

      Sorry the answer is so vague, but that's a very tough question to answer with a very specific and exact methodology.

  2. Do you start with Direct Linking and then move on to Landing Pages?

    Do you have any example landing pages showing the connection from ad to offer?

    • rileypool says:

      I've always done direct linking myself. Some people preach by direct linking while others only use landers. Due to my lack of using landers, I don't have any examples to provide you. Sorry.

  3. Do you have to raise bids when your targeting becomes tighter or is it the other way around?

    • rileypool says:

      Yes, when your targeting becomes tighter you should raise your bids because your target demographic gets smaller.

  4. I have a question about the image optimization.

    For one of my ads, the CTR was good but didn't have any conversions so I let it run before it met 2x the payout offer. The ad converted but now the ctr is below good and it has passed the 2x the payout offer. What would you suggest?

    Here's the stat:
    Clicks: 17
    Ad Impressions: 23945
    CTR: 0.07%
    Conversions: 2
    Payout: $7.50
    Spent: $18.75
    Net: $-3.75

    When you give an ad the opportunity to perform, how far is too far until you generally kill the ad?

    • People put a little too much emphasis on CTR on POF. What matters more than CTR is ROI. You can have an image with a .3 CTR, but it might not convert. The only thing that matters is your ROI.

      Personally, I would ditch that ad.

      Your final question is hard to answer. When I'm testing a new campaign I give ads about 2k imps to get a decent CTR. Then, I usually give the campaign as a whole a spend of about 5x the offer to get conversions to see if it has potential instead of focusing on each specific ad.

      • What would be the negative roi threshold until the ad is ditched?

        • That\’s almost impossible to answer without given some numbers to go on. It\’s a very case-by-case basis.

          If the campaign is profitable but that particular ad isn\’t, I pause it almost every time. If the campaign isn\’t profitable, I just compare the ad to the rest of the ads and pause the worst ones.

  5. would you ever direct link an offer lie true to a niche demographic or would you always use a lander for a case like that and hope the ad sells itself enough for them to sign up through the offer?

    • I\’ve done it countless times. Sometimes profitable, sometimes not. It\’s tough to direct link a niche demographic to a general dating offer. If you don\’t sell the offer really hard in your ad copy, you\’ll probably be better off creating a lander, but that creates a whole other set of variables.

      It\’s advisable to test both approaches for best results.

  6. Weekparting
    Hi,
    Having just joined your site yesterday I was very interested to read that really apart from fri sat and sun there is not much point in buying traffic for dating offers.
    I have just tried to get my first dating campaign going with plenty of fish but can’t seem to get a high enough CTR and no conversions at all possibly the days of the week I was testing were hurting me a bit.
    Is there also good and bad days for other type of offers?
    Thanks
    brian

    • Hey Brian,

      I’ve never weekparted or dayparted my campaigns. I would always let them run constantly. I would have too many going concurrently to micromanage every campaign. When you’re starting out I can definitely see how advertising on just the best ROI days would be in your best interest. Also, don’t forget Mondays and Thursdays tend to good days too. Not the best, but good.

      Your CTR will be crucial. Be sure to test a handful of images.

      Days of the week always matter, but it depends on the vertical. Every vertical is different when it comes to best performing days.

  7. Hey Riley,
    I’m a bit confused about this…
    You mention that you like to go for one campaign and targeting a single demographic. But when you’re split testing the head lines and ad copies, you have to create different campaigns for each. Am I correct?

    And when you found out which ad copies / head lines work best, then you’re going to optimize the targeting in one single campaign. Right?

    Thanks.
    Sascha

    • Yes, you’re exactly right.

      When I’m split testing I’ll create separate campaigns for each ad copy. I’ll name the campaign something related to the headline, so I know which one it is. For example, what I usually do is use a keyword from the headline. Like if i was split testing the ad copy Want a Girlfriend? and Need a Girlfriend? I would name each campaign US-TrueM3135-Ads-Want and US-TrueM3135-Ads-Need, respectively.

      And yes, when I find out which combination of headline and ad copy work best, then I create a separate campaign using the best ad copy and scale it to every demographic I can.

  8. Okay, got it…but by creating separate campaigns ( and let them run all at the same time ) for each adcopy or head line, I have this feeling that I’m competing to myself. What are your thoughts on this?

    And how many impressions do you suggest to give to the whole campaign or each ad before making any decisions on?

    • That’s exactly the purpose in my opinion. You want your ads to compete amongst themselves. As long as you’re bidding the same and have your frequency set the same, your ads will be competing against each other and you’ll be able to see which ad(s) perform best.

      That’s very difficult to answer. I try to maintain at least a .10% CTR for ads targeted to females and .20% CTR for ads targeted to males. If the ad isn’t getting any clicks by 3000 impressions, I pause it.

  9. Very cool! I totally understand it now. Thanks for taking the time answering my questions Riley.

  10. Well I definitely liked studying it. This information procured by you is very effective for good
    planning.

Speak Your Mind

*