Banner Ad CTR Thread (24)
ok, I wanted to make a thread on banner ad CTR to share a bit of knowledge and gain some as well. As you are all aware banner ads are all about CTR, obviously your cost per click is cheaper the higher the CTR.
So the aim of the game is getting the highest CTR possible and the only way to do this is to split test creatives. I'm showing you guys a $95k ad spend I did on google content (display) network last year. The ads in this example were USA only, I had them running in other countries as well - but for now I'll show you my USA stats.
Couple of things to note:
- I started the campaign off with around 5 variations of 300x250, 728x90, 120x600, 160x600 and 468x60 sizes
- once the clicks and data started coming in it was easy to can the losers and keep the winners
- creatives that worked in the USA tanked in other countries. In some cases banners that were getting 3% CTR's on one website in the USA were getting 0.005% on websites outside the USA
- the USA easily had the worst CTR's. Internationally I was getting CTR's as high as 9.0% on some websites
- this offer was direct linked, I split tested using a lander and never got it as profitable as direct linking
- i got google slapped doing this campaign, which is why i've done some writing and stuff all over it - so if you want to copy it you can't as google have closed this offer and will ban your account. it wasn't a rebill it was a clickbank ebook/course. this account is now fried/toast/totally fucked. All up I spent around $300k, total ROI for the campaign was around 50% when I factored in refunds. Refunds were nearly 4 times greater in the USA than any other country. In fact USA clicks were expensive and the fuckers kept hitting the refund button so quickly it made me wonder what the fuck goes on inside American consumers heads.
Anyway, so after gathering all this data I had a goldmine of information to play around with. I knew which creatives were getting clicks on which websites and what I should be paying. So I went to the top of the list and looked at foodnetwork.com
Now being in Australia I don't know much about fodnetwork, but with a bit of research I discovered that it's the portal website for a lot of cable/pay tv shows like rachaelray and others. Heaps of recipes etc, demographic was mainly female aged 35-55.
Now fodnetwork is owned/run by a company called scripps media. So I rang them up and spoke to the advertising department about buying some inventory on their websites. I played the part of an advertising agency and had my own website set up so it looked professional.
I knew with google I'd spent around $16k with a CTR of nearly 2% with a target CPM of $2.37. So when they quoted me $15 CPM I nearly fell off my chair. I told them that was nowhere near what I was prepared to pay, I offered them $1.00 CPM for $20k worth of traffic.
We went back and forward for a bit, I sent emails and kept in contact, finally they agreed to a $10k buy at $3.00 CPM. A bit higher than I wanted to pay but it was well worth the test. I was using adshuffle as my adserver, as I wanted to be billed off my numbers - but their technical department said I needed to use their adserver or the buy couldn't go ahead.
They provided full login access to their platform so I could monitor my numbers. I wasn't happy about this, but agreed for the first buy I'd go off their numbers. Next was the payment terms, being from Australia they couldn't credit check my company so I had to put up the $10k before they ran any traffic. So I wired the cash, sent them my creatives and we got rolling.
I had a $10k spend over 10 days, even delivery, freq cap 3 per day per user, USA and Canadian traffic only, out clause of 24 hours at $3 CPM.
I won't post the exact figures of the buy, simply because it would mean taking internal screen shots of their platform, but in the end my $10k returned me $14,500 - but after fucking refunds I came out with $11,000 for a profit of $1k at only 10% ROI.
Couple of reasons for this could be;
- I was billed off their numbers, I paid for more clicks than my tracking showed, but it was off around 5%
- CTR of my ads tanked. On google I was getting 1.97%. Buying direct I was getting 0.80%. Hmmmm, were they putting my ads below the fold... not sure why this would be. Perhaps google has far better retargeting options and maybe their platform is more relevant as adgroups are bundled by keywords
- banner blindness? perhaps, but I was buying a very small amount of their available inventory
My scripps media rep asked me if I wanted to buy some more traffic and I balked as the CTR and ROI was down. He offered me the same buy but with a $2.00 CPM for $10k more.
I agreed, and the results on the $2 CPM buy were nearly identical to the $3 CPM buy with my CTR now only around 0.50%.
I did similar tests on other websites as well and in every case I found the same results. CTR on google was nearly 2-3 times as good as buying inventory direct. I'm not sure why this is, if anyone has any suggestions I would be happy to hear your thoughts.
Also, one service I am looking for is finding a good photoshop guy who can make banners both static and animated/flash for reasonable rates. I've posted on warrior forum and only got answers off guys in India or wherever wanting $15 per banner, which gets expensive when I want 50 or more of them made. So if anyone has a resource they know of where I can get bulk banner ads made that would be great.
I'm particularly after anyone with skills in making flash banners. So hit me up or reply if you can help.
Hey Harry,
Thank you for sharing your data. I had the same experience when I was using adwords a few years ago. I ran a diet too (rebills). I noticed a site that was converting very well and set up a direct buy with them. My CTR went to shit and I was paying more than $1 per click so I cancelled it in 24h.
The way adsense works is awesome. It targets visitors over the content network that have searched the terms you put as keywords in your campaign (search retargeting + site retargeting). IMHO thats why we see this difference.
I also have a site with SEO traffic and my adsense gives me around $10 CPM in the finance niche. Theres no way I can make that with CPA offers and shit.
Google is just too good in microtargeting all the visitors.
Them some juicy juicy CTRs, not to mention tasty CPCs.
You had the right idea contacting the owner of foodnetwork, good work negotiating down from $15 CPM. Out of curiosity how did you twist their arm into dividing their intial offer by 5?
I just offered them $1 and their first counteroffer was $5. I firmly believe companies with no idea about ROI will happily pay the $15 without negotiating. I also spent some time building a relationship with my rep so finally we were able to work it out.
Generally when doing buys I'll counteroffer very low, making out that I'm buying unlimited volumes off their competitors at the low price. Doesn't always work, but I won't pay stupid CPM's. Everything is negotiable.
Sweet stuff, the first thing that came to mind for shitter CTR was lower CPM = traffic from certain pages/below the fold?
Proper negotiating! I need to learn from you. Any negotiating tips?
google is a contextual targeting engine.....
so with your banner/txt ads you ads are shown ONLY on relevant content.....not site wide.....
so to get the same results ctr wise you would need to be placing them on the exact same page...that google does...
google has a very complex engine that handles all its placements.....best advice I can give is talk to an adsense guy who runs auto blogs they can tell you excatly what they do content wise to draw out the good keyword clicks....
also
google has a pretty hefty click fraud engine built in that def cuts out or severely limits placements that have a high bounce rate on the click thru as they determine that to be a "potential fraud click"
all this ads up to some very high CTRs ...un natural even....when comparing with other sources.....
awesome post! thanks for sharing
its a bummer that your ROI tanked after going direct. What polarbacon mentioned above makes perfect sense though
p-bacon it's like you just rained down knowledge, relevant ads on the keyword specific pages didn't even cross my mind lol. I almost clapped when i read that!
Yeah, I heard from a google employee that their algorithm takes into account every google search you do as well. So someone who wants to lose weight and types in "deit, weight loss, jenny craig" etc. will get shown more ads based on their previous searching history. Very nifty. Almost "big brother" knows what you are searching for online and will show you what it wants.
....didn't John Connor fight something similar in the Terminator movies...
anyway, still looking for some banner ad options. I've been sent a few and I'm more than happy to share with you guys any gems I find - so if you know of a good photoshop guy/gal who can do multiple banners hit me up.
Awesome tutorial this is literally media buying at its finest.
I'm also surprised going direct didn't back out better. When you try to negotiate your buy next time create different tags for difference sections of the site. For example one buy I did the homepage just didn't perform but all the internal pages did great.
Also, get a proxy and always make sure your ads are being served properly. Sometimes these sites will tell you that you'll get first impression order and instead they serve you way down the que.
Great thread - will be following.
PM sent for your banner needs.
Here is what I know about media buying from doing a ton of it myself.
You did everything correct except explicitly ask for "above the fold" on your IO. Also, by the 2nd or 3rd day I would have already had them pause the non performing ads ( you mentioned you had 5 of each ) so I would have found the lower performing versions and cut them off and then found the lower performing sizes and cut them too so by the 4th day I would be going into it knowing my control versions and making new ones from those.
Also, the most expensive CPM I ever paid for is $4 and that was b/c I really wanted to try a certain network. Now days I will not dare pay more then $2 CPM.
Also, you mentioned only doing a test of $20k. This might have put you into the "small" group of advertisers they have and the reason they didnt give you better pricing. I always go into an IO with the idea of spending 1-2mill over 6-12 months so they know I'm serious going into it, even though I might execute my "out" in the first 10 days.... You will normally get better pricing that way.
So you are negotiating an out-close of 10 days +?
Very nice info by the way, I really like this thread.
harry you can use bannersnack until you find a good designer...its pretty good and does the job
@eliquid - would I have been better off using the $10k as a test buy, meaning I load in multiple banners to multiple landers promoting multiple offers that suit the demographic of foodnetwork.com?
It was capped at $1k a day only anyways, so maybe I'd have been better off running diet, skin and biz opp to see which ones converted the best.
I also wanted to post that for me over many buys in many verticals my 300x250 banners above the fold always outperformed anything else. In terms of ROI I ranked my banners like this:
1. 300x250
2. 160x600
3. 120x600 - not too many websites have this size
4. 728x90 - easily the banners size that can burn through the most inventory, be careful if you're waaaaaay below the fold with this one
5. 468x60 - not much traffic and not great ROI
I didn't include the 336x280 spot as it seems to be exclusive to google and performs around about the same as the 300x250.
on your direct buy, do you think it might have turned around had you had your rep shut off bad performing pages?
Originally Posted by harry1970
Spreading yourself thin creates a headache, unless your just trying to master that 1 and only URL/Media Buy
Originally Posted by harry1970
I almost always end up sticking to the 300x250 ones.
Harry,
I know this is taking away from the point of your post -- but 12 cent clicks? In the diet niche? Is that just a question of keeping CTR so high? What did your banners look like to yield such awesome clickthroughs?
Any chance your could drop some knowledge on how you tamed the Google Content Network beast? Would love to see a case study or something along those lines! 
it all comes down to ctr. just like facebook, for ages i just couldn't work facebook out, then finally i got an ad 1.5% CTR in Canada and they flooded me with 0.03c clicks to a dating offer. google content network was exactly the same. i tried multiple banners and went backwards in terms of ROI at the start.
as far as the banners the main ones I ran were the shrinking bikini ads. i'll have a look around my hard drive and see if i still have them floating around.
Originally Posted by eliquid
Originally Posted by caleb
When you spend $2700-$3500 a day on a new campaign with Google and you don't get sales in the first 48 hours ( so now your up to $5400 - $7000 ), do you keep running it? Now replace "Google" with Bing/TrafficVance/Facebook/Pulse360/AdMob/7Search, etc.
Too many affiliates are doing media buys wrong. These reps at RightMedia, Synacor, TribalFusion, etc know what is working for who. The biggest scam in media buying is when you take your weightloss/biz op/skin/teeth lander to them and ask them to run it with your creatives and they say they need to but you on their "RON" so they can find the profitable placements for you. ( this is true maybe for an offer they have never ran though )
These guys have internal campaigns running and other affiliates running the same landers and niches and they know where its profitable already. You need to make sure you being but on targets that are already proven on their network for the same landers/offers your running.
By doing this, the only "what if" is your creatives and offers, since you might be running a diff acai or teeth offer then another affiliate. By the time you spend $2500 you should have a solid idea of which creatives are going to work the best by looking at CTR. If you asked for above the fold placements, then all you are judging is which size/design is working the best on the placements.
If I do the work for you, its mostly going to be the 300x250 placement above the fold. With this knowledge I just gave you for free, all your looking at will be pausing all other ad sizes, only running 300x250 and looking at which 300x250 design is the best. You are also going to know which of the premium placements are doing the best for you too since you asked the rep to only place you on placement other advertisers are already doing well on.
So yeah, by day 2 or 3 you should have already paused at least 50% of your targets and almost all your creatives except the best preforming ones. On day 3-4 your going into it with a honed in campaign and your only trying to improve on your creatives at that time ( or your LP ).
Home > General > Affiliate Marketing Forum