Home > Display / Media Buying >

---

Media buying Ask Me Anything, Episode 3: Revenge of the Ad (28)

---

You guys know the drill, ask me anything about media buying. I reserve the right not to answer a question if it is too personal or off topic.

Fire Away!

We'll also give away one free month long license to WhatRunsWhere for the best question as a little bonus

---

I'm doing great volume of media buys on Facebook, I'd like to port my skills obtained on Facebook to standard media buys. Where do you suggest I start?

---

Do you prefer text ads over banner? There is some banner blindness going on for the past couple of months.

---

What is your experience on Lean.com?

---

sitescout vs first impression?

---

I'd start with your core competency, you already know your demographics, you know who you're going after. Now just leverage data against that and you're pretty much ready to go. Start buying through RTB platforms and scale from there

I don't mind either or. Both have their merits, banner blindness is easy to fix, just test out new concepts or change small components of the ads regularily

---

Haven't used it to buy, but Cyrus is a good guy, would say test it with a small budget and see

Both have their merits, SiteScout is older and crammed full of direct response affiliates, on the other hand First Impression is newer, more raw but shows some promise. It's a personal choice about which you personally have success with. For example, some people kill it on Adsonar but can't make Pulse360 work, it's personal preference about which system fits YOUR needs

---

Let me make it a little more complicated then My core competency is Facebook apps, and I wan't to promote something new, what do you suggest?

---

Come on numerouno! Go get you feet wet and go spend some money figuring things out

Personally, (with all the due respect to mattaw because he helped me) I don't think he has the magic answer for you. Think about it -- on one hand he wants to help you, one the other hand he can't really point you to the right direction because the right direction can only be found by testing.

If you already have a feel for your demographics you could start with Google Ad Planner + Quantcast + Compete.com + Alexa. That could help you identify sites where your desired demographics goes for a daily dose of internet.
If you can drop $10K-$20K without breaking too much sweat go to large ad networks, launch a campaign and optimize a living light out of it.
If you think apps are the future -- fool around with mobile ad networks doing only in-app advertising, create a lead gen and achieve greatness with mobile app advertising...

---

Pretty much answered

---

from what i have been reading, there are two options to limit the risk when we start off with a new buy + new creatives

1) wrw -> see what's showing up in the last 10 days for a given niche... find similarities and improvise
2) wrw -> keep track of what's been running the longest (or sort by ad strength)... find similarities and improvise

--

i am having a hard time balancing between 1) and 2)

for 1) just because they are showing up, we can't tell if they are profitable or not because they haven't been running for a long time
for 2) since these ads have been running for a long time, we know they are profitable if they are still seen up until current day... but wouldn't these ads be more prone to banner blindness if we hit the same audience?

i have only tested banners innovated from 2) ... which has given mixed results in different buys

--

my question is... why does 1) work? ads seen in the last 10 days could be just test ads someone throw up and recorded in wrw... how do we know if we are innovating from profitable ads?

---

Whats your view on Automated Buying / RTB and how this will / is this changing media buying?

---

bump?

---

This would be useful. Also another thing with the spy tools, even FB ones, is guessing what ad is profitable or not, you cant just look at live duration as it may be a brand advertiser. I like to review ones with potential to see if i can see any direct response elements. I wonder if there are any other tried and true ways of confirming an epic find so to speak.

---

Ok, I am a bit confused. I have gone thru all my optimization notes and what not and when it comes to optimizing I hear so many differnt stories. Though my common sense says Optimize based on CPA cause CTR doesnt mean S**** if there are no conversions.
however when it comes to the proccess. So say Im going direct with 6m impressions prepaid, I have 12 banners/4 differnt styles/2 LPs and 2 offers.

So I test 12 banners 2xcpa each (some say 3xcpa?) then I pick up the ones that performed the best (most conversions + .2ctr) then i tweak them around. and next round testing I test these again for another 2xcpa. But what if half way thru testing say 1xcpa the banner has bad cr/ctr. Is it fair to cut these sort of banners off or it's better to keep testing till the end?

Cause I realized a lot of $ can be saved if bad creatives could be detected earlier. I might be wrong but any feed back would be appreciated.

Cheers.

---

few questions unanswered... last bump :P

---

They used to help out once in awhile but no one is really looking after these threads in media buying like before? Polar? maynzie?? can you guys answer some of these questions?

---

When we say test 3-5x CPA, its a rule of thumb. If you havent spent a ton of impressions, but you see all variations of banners with a "Get Help >" button are outperforming the other half of banners with a "Submit" button, then you should pause the latter variations. When you get a feel for advertising, you know not to even use c2a buttons that say "Submit" or "Send" in the first place.

Unfortunately, testing so many elements like you are typically will help you end up spending a lot. Over time you get a feel for what appeals to different audiences, but you have to be smart about testing only enough elements to get feedback QUICKLY, and make your next ad or LP variations on the fly based on the feedback loop.

Right now your testing 192 variations. I would recommend dropping the styles and extra offer. I always test real simple, headline, image, call-to-action. Get feedback on whats working the most, then alter your approach on the fly. It works a lot better than throwing up 192 variations in the beginning.

I often test to some degree using CTR if I enter new niches initially. Keep in mind there are two major elements of CTR:
- INTENT: This is super important. I can say "Free Money by Clicking Here!" and will get lots of clicks but few buyers of my bizopp. If I say "Earn $3,650 with "Robbins-like" Mind Program" I will get MUCH more qualified clickers
- CLICKBAIT: Throwing up an image of some nice underboob in a bizopp ad will get you clicks but not conversions.

---

This picture.

I like it.

---

They're all the same girl....only diff are numbers on chest and hair style....

---

I got a question...

i was ready to do a media buy using sitescout rtb along with adshuffle (adserver), having done research on the top 100 publishers for weightloss.

And to start with i was going to go on just 10 Publishers and of those pubs, to go on 10 specific url's (so 100 specific url's).

I was having traffic, even though the 10 pubs are all USA sites, come from both the UK and the USA.

It was all set up and live but no traffic came? NOTHING! It did not show that the Pubs were blocking us and we even changed the cpm from $2 to $5
in the case that the bids were not high enough and still nothing, not 1 impression?

Can you tell me why that would be?


I spoke with sitescout about this and they were looking into this and they did not know why either.

Then 1-2 weeks later, nothing from sitescout. So i speak with them again and they are looking into it again!


So no traffic at the mo which is frustrating.


Another question that has just come to me...

What RTB/Adserver etc (tools) are the best?

Or which ones allow you to go url specific and not site specific (meaning...a website has 500 web pages and i have identified that advertisers are on placements for 25 urls of that publisher (website), so that i do not blow money on the
other 475 urls)?


Is it best to go direct with each of the publishers that i have researched (yes it takes a lot longer and do not want to go down this route to start with).

Cheers

Rick

---

What do you see as the essential elements of a successful banner? Are they comparable to a PPV landing page?

Also, I saw a few articles that showed 300x250 banners are the most clicked of all traditional sizes by a pretty good margin. I decided I would only use 300x250 when I started. Then when I had a profitable campaign one way of scaling would be to try different sizes.

This is all theoretical since I haven't had a profitable campaign yet but do you think it's valid?

---

I am also having same issue with Sitescout. My 2 profitable targets suddenly stopped sending traffic and they're still active (not blocked).
What would you like to suggest Mattaw?

---

Rick,

Try posting in the Sitescout's 'official' thread. http://stmforum.com/forum/showthread...-Thread/page16

Maybe Dr. S will see it.

---

I'm currently trying to tackle dating on SS. How do u tell if a campaign shows promise or not? Will it be almost profitable?

Currently i'm running a general dating offer using FB look-a-like ads. My CTR is high, but my CVRs are horribly low. I've tested only 2 sites, and both didn't do well. I made about 4 CVRs and spent ~$300. These are losers right? I mean is there any possible way in hell that by running a different angle/message that i could turn a $145 spend and $12rev into a profitable campaign? Logically i think "No" but to hear from someone thats done it for a while would be cool.

---

test a few more offers on those sites, if that doesnt work take your best offer and try a landing page.

---

Bump

---

Chances are that inventory has already been sold since SS traffic is remnant traffic. When the space is not sold impressions should flow through.

---

Home > Display / Media Buying >