Home > Paid Traffic Sources > Mobile

---

Cutting placements, Where to start? (8)

---

When comes to placement cutting, which category you guys usually start from?

For popads, we can cut OS versions, Browser Version, Web ID etc
For propellers, we have OS, Browser and Zone ID.

For me, if cutting from Zone ID, the cutting process will be faster but this will also cut a lot of "maybe good" placements from OS and Browser. But if cut from OS and Browser, it will take more budget and time to cut as we need to collect enough data for each of them. The worst thing is, most of them are inter-related. So how?

Any ideas for me?

Thanks

Harrison

---

Hello

Exactly as you said, cutting zones/placements too soon is not good, because those "maybe good" as you named them, will be out before yuo have a chance to optimize the funnel.

My method is like this :

Note: I run 95% banner traffic, but the system is the same as with pops, except for banner CTR.

1. First I need to make sure I'm getting decent quality traffic and not all some BOT crap, so I cut zones/placements with atrocious performance - 0% or very low banner ctr, very low LP ctr and placements with a very high % of bots.
2. Next I get rid of banners with lousy CTR, because when the clicks are too expensive, it can't work no matter what - you can leave out this step on POP sources.

At this stage, I know I'm buying mostly legit traffic and that I'm not paying unrealistically high rates per click

3. Now come the LP's and offers. At this stage, I need to get a reasonable amount of conversions so I can narrow my funnel down to the best offer and LP. This is where we ended in your follow along so to speak.
4. When I have the best LP and best offer, or multiple ones, I need to collect some more data and only then can I start cutting based on the other factors.
5. Whether I start with OS, carrier etc ... depends on what the data tells me. I'm looking for the WORST parameters again. If I see I made no leads from wifi traffic, I cut it and leave everything else in tact and run for some more. Then comes another parameter ... I like to cut placements as the last form of optimization.

The basic principle of my method is to start by cutting the WORST things and getting down to the minor parameters ones step at a time.

---

Hey Matuloo, thanks for the info!

Cut the worst. I will remember this.

So after cutting a placements, how long you will run the campaign for before another batch of cutting? I usually let it run 24 hours which I think you will say it is too fast.

---

This will depend on the volume again I'm afraid. You need to get enough data to see some targets falling clearly behind, compared to the rest. Translated to AM language, you need to get multiple leads and see that some targets are not converting, while the others are.

---

Hi Matej,

1) Do you initiate your bot test when you start your campaign or when you've found the winning lander/offer? The reason why I ask this is because in the beginning phases of testing your campaign you don't want anything to interfere with you finding your winning offer/lander. I see the bot test as a barrier that doesn't allow you to get the best lander results because if you're doing a bot detection test with a script(redirecting users after 300ms) then it will slow down your landing page. Which in this case it doesn't help you determine a good lander right? Unless you implement bot scripts in all your landers that you're testing. Hope this makes sense to you, let me know if it doesn't.

2) When finding the winning lander/offer are you already greening on those landers/offers? Or are those landers/offers still in the red and you're just picking the best lander/offer from your initial test in hopes of now getting it profitable?

3) Looking at your WORST parameters approach it seems to be something like this. And in the end, it all seems to be a "drilldown" method in terms of what to cut first

- Connection Type: Wifi or Cellular
- ISPName
- OSName
- BrowserName
- Placements

So you're starting from the top to the bottom in terms of cutting. When you're testing in the initial phase it seems that you might either separate campaigns in a few different ways:

1) Separate between OS and Wifi or
2) Create Multiple ISPName campaigns or
3) Create campaign between iOS or Blackberry or Android

One of the biggest issues I'm facing is that when I look at my campaign data, sometimes I'm not sure if it's worth working on the campaign when a Browser like Android Stock Browser is generating 1000 conversions while Google Chrome 51 is generating 20. The Android Stock Browser is in red with a profit of (-$13.50) and Google Chrome 51 is profiting on $2.00.

---

This is a nightmare I would say. But as a rule of thumb, we run campaign to make money. So keep those with profit even $0.10 while cut those with a loss even it is losing penny daily.

---

Guys, let me write a separate thread detailing this method a bit more, then we can continue the discussion there. I will post the link here once I'm done with that.

---

OK, thread posted here : http://stmforum.com/forum/showthread...146#post279146 Check it out and ask questions I already tried to answer those you asked Simon.

---

Home > Paid Traffic Sources > Mobile