From red to green numbers - Optimization Process

[ironbull]

Hey AffiliateFixers!

I would like to know your thoughts about the optimization process of campaigns. I have my optimization process mapped out in an Evernote document so each time I optimize campaigns, I follow that sheet.

I mainly run Mobile Pop Up traffic.

Here is what it looks like:

1. Launch campaign with at least 3 landers (one angle per lander) and 3 offers (same/similar offer, different networks)

2. Spend 1x lander, 5-10x type of traffic (pop) and 1x offer per day (if there are banners involved, will add 1x banner to the equation)

So for example if the payout of the offer is 0.20$, having 3 landers and 3 offers rotating in Mobile Pop Up traffic, I would spend 0.20*3*5*3 =9$/daily

3. Wait 3 days to gather data without touching anything

4. Start taking decisions on landers, banners and offers with a 95% of significance

So for example, if I got 545 clicks and 11 conversions in one lander and 769 clicks and 21 conversions in the other landing, I don't take a decision yet until there one lander is 95% more likely to be better than the other one.

Here is what I am talking about:

5. Start blacklisting variables (targets/apps, carriers, OS, devices, browsers etc.) - If anything has spent more than 3x offer payout without conversions, I will blacklisted or redirect the traffic to another network so I can monetize it better.

6. Add more landers/banners if I came up with new ideas about angles and start split testing them with the winners of the first round. Also, implement winner angles in a different design of landing page.

7. Rinse and repeat until numbers turn from red to green.

8. Start split testing more in detail being on the greens (header, text, call to action, images, scripts...)

OK, that is mainly what I am doing to optimize a campaign but I would like to know your thoughts about it.

Here there are some questions I have:

- What is a good strategy for bidding? - What I am using right now is bidding high in the first 3-5 days to get fast data and then lower the bids when I have a profitable campaign with positive ROI.

- What do you think about day parting? Is it really worth optimizing that deep?

- If some lander/offer/target/device...made some conversions but ROI is under -50% by the third day, what should I do?

- When do you find out that your campaign is not gonna work because the offer is already burnt/declining in the network?

- Should I increase my daily budget per campaign after the third day if I see potential in order to gather data faster and not lose the "potential" of the offer in the right moment (even if I am in negative ROI)?

Wow, that was long for me used to write 200 words max landing pages

 

[Acura]

Hey AffiliateFixers!

I would like to know your thoughts about the optimization process of campaigns. I have my optimization process mapped out in an Evernote document so each time I optimize campaigns, I follow that sheet.

I mainly run Mobile Pop Up traffic.

Here is what it looks like:

1. Launch campaign with at least 3 landers (one angle per lander) and 3 offers (same/similar offer, different networks)

2. Spend 1x lander, 5-10x type of traffic (pop) and 1x offer per day (if there are banners involved, will add 1x banner to the equation)

So for example if the payout of the offer is 0.20$, having 3 landers and 3 offers rotating in Mobile Pop Up traffic, I would spend 0.20*3*5*3 =9$/daily

3. Wait 3 days to gather data without touching anything

4. Start taking decisions on landers, banners and offers with a 95% of significance

So for example, if I got 545 clicks and 11 conversions in one lander and 769 clicks and 21 conversions in the other landing, I don't take a decision yet until there one lander is 95% more likely to be better than the other one.

Here is what I am talking about:

5. Start blacklisting variables (targets/apps, carriers, OS, devices, browsers etc.) - If anything has spent more than 3x offer payout without conversions, I will blacklisted or redirect the traffic to another network so I can monetize it better.

6. Add more landers/banners if I came up with new ideas about angles and start split testing them with the winners of the first round. Also, implement winner angles in a different design of landing page.

7. Rinse and repeat until numbers turn from red to green.

8. Start split testing more in detail being on the greens (header, text, call to action, images, scripts...)

OK, that is mainly what I am doing to optimize a campaign but I would like to know your thoughts about it.

Here there are some questions I have:

- What is a good strategy for bidding? - What I am using right now is bidding high in the first 3-5 days to get fast data and then lower the bids when I have a profitable campaign with positive ROI.

- What do you think about day parting? Is it really worth optimizing that deep?

- If some lander/offer/target/device...made some conversions but ROI is under -50% by the third day, what should I do?

- When do you find out that your campaign is not gonna work because the offer is already burnt/declining in the network?

- Should I increase my daily budget per campaign after the third day if I see potential in order to gather data faster and not lose the "potential" of the offer in the right moment (even if I am in negative ROI)?

Wow, that was long for me used to write 200 words max landing pages

Hello
Looks like your on the right track I would take the best converting lander and scale up the traffic to it. I have found that sending alot more traffic made a big difference in the ROI. If that does not work I would do some more tweaking to that lander. Best of luck

 

[ironbull]

Ok, just now I have a Campaign running with a lot of data and in terms of angles and offers, I think it is optimized. It will soon turn green.

It is well known that if a target, carrier, OS version etc. has spent more than 3x offer payout, we should block it or redirect it to another offer to not lose the traffic we are buying.

What about targets, OS version, browsers, devices that had conversions but they are -90% of ROI?

When should I make the decision to block/redirect them? Is there a number of clicks should I wait for?

Thanks in advance

 

[Mobidea]

Hi @ironbull , you developed a great optimization strategy!!

I'll try to answer below to all your doubts.

2. Spend 1x lander, 5-10x type of traffic (pop) and 1x offer per day (if there are banners involved, will add 1x banner to the equation)

I agree with you, but you need to take also in consideration that the performance is not the same along the week, for some days you can get more volumes than for others.

3. Wait 3 days to gather data without touching anything

It's better to focus on the data amount than on the day amount. It really depends on the campaign, spot, country, etc. For example, if you're promoting a campaign on a country where the conversion rate is usually higher, you can take a decision with less data, so you don't need to wait for a lot of clicks; for other countries you should wait more. If you have already some performance history to take into account, you can use it as a reference for next optimizations.

- What is a good strategy for bidding? - What I am using right now is bidding high in the first 3-5 days to get fast data and then lower the bids when I have a profitable campaign with positive ROI.

The bid strategy you're using is the right one. You should bid higher at first, to gain the first positions in the market and then optimize it taking into account your ecpm, so you should reach the bid that is profitable for you. Are you able to use smart bidding?

- What do you think about day parting? Is it really worth optimizing that deep?

It is a great tool only if your campaign has good volumes, otherwise it's not worth it. We are actually using it. We don't cut the traffic for some periods of the day, because we can lose potential users.What we normally do is choosing one or more campaigns that have a lot of volumes but for which we have not the first positions, create a campaign separately for some periods of the day for which you get a good volume of impressions (let's say your peak period of the day) and you're actually profitable and bid higher. You can maintain the other campaigns, since they have different bids, they will not competing each other.

- If some lander/offer/target/device...made some conversions but ROI is under -50% by the third day, what should I do?

My advice is to stop targeting it, but only one thing at a time. If you cut every target you are not able to figure out what element impacted the negative ROI. So my suggestion is to make a small change, analyze the result and then optimize it again. Other thing you can do is to split the campaign and put the "bad targets" with a lower bid, just to continue receiving traffic in a profitable way.

- When do you find out that your campaign is not gonna work because the offer is already burnt/declining in the network?

Only after you tried everything.

- Should I increase my daily budget per campaign after the third day if I see potential in order to gather data faster and not lose the "potential" of the offer in the right moment (even if I am in negative ROI)?

My suggestion is to never put a maximum daily budget on a test, because the results will not be trustworthy. The performance of a campaign is not the same along the day, so if you put a small daily budget on a campaign for which the peak period is at night, maybe you'll not have enough budget to test the campaign when it gets more volume and to take any conclusion based on it.

I hope this helped you

 

[ironbull]

Great reply @Mobidea

Thank you for solving me these doubts.

What about recycling traffic that is not converting? Right now I am using YTZ to monetize it but it is not performing well at all. Do you recommend any network similar to YTZ?

I am using ZeroPark to test my campaigns so they just give you the ability to pause targets, change frequency or set up day parting, not much to touch really.

I am thinking about changing to PopAds which I think has more options in terms of optimizing variables.

 

[Mobidea]

Thank you

Well... you can try Mobidea We offer a lot of solutions for you to monetize your inventory. A great tool that I think it can be adjusted to your needs is the performance fallback. You can set your target and/or a mininum eCPM you should have to define if your traffic goes to YTZ or stays in Mobidea. For example, you set a minimum eCPM of $10, if you get more than $10 of ecpm your traffic goes to Mobidea, if it is less than $10, your traffic goes to YTZ. It means less risk for you and the chance to test two networks at the same time.

We are actually using PopAds to buy pop ups and pop unders, among other adnetworks. You can exclude or include websites ID, taking in consideration their performance and it works based on the smart bid, so you always pay the minimum bid to maintain your position. And you have more targeting options as well. The inventory is really huge and also you can target by ISP (Internet Service Provider). I think it is worth a try! Are you buying only Pop's?

 

[ironbull]

I am going to try PopAds definitely. ZeroPark should develop their targeting better in order for me to use their network again.

I am focusing on Pop traffic first because I am starting out. Once I find this profitable, I will expand to banners also. When you include banners in your equation, you add another layer of optimization and although it could be as profitable as pops or more, I find out that it is best to focus on one type of traffic first, make profit and then move to a next one.