Often times the affiliate game is played like this:
1) The affiliate decides the niche or offer they want to promote
2) The affiliate finds relevant converting traffic for that offer
3) The affiliate profits
Although there are more steps involved like building a landing page, tweaking the campaign, getting rid of non converting traffic etc, that is how affiliate marketing works in its most simple form. What if you turn the traditional way of thinking about setting up a campaign and turn it on it’s head. I call this the “Offer Matching Method”.
Most affiliates have a favorite source of traffic. Often times this source of traffic is easy to get ads approved to and converts well (otherwise it probably wouldn’t be their favorite). Rather then picking an offer and finding a traffic source to match, PICK A TRAFFIC SOURCE AND FIND AN OFFER TO MATCH.
Let’s use this as an example. You have a friend who owns a popular forum about music. The site gets tons of traffic but he isn’t good at monetizing so he sells his banner spots off to you. Rather then picking an offer that you think matches the demographic, like ringtones, you make 20 different banners for different offers and let the stats tell the story. Which got the most clicks and then more importantly the best EPC/conversion rate. It could very well be an offer that you don’t expect. You then know the best niche for that traffic source. You also may have found several niches that work well that you can rotate so you don’t burn out your audience.
Analyze the situation above. You don’t need a friend with a music forum. This can work with most traffic sources. You can even pick a demographic on facebook and make like 100 different ads from various niches/offers and see which performs the best. Yea, there is an initial outlay of money that you will probably lose BUT, winners can make up for that by many multiples over time.
Have you tried this? Will you now?
This is to Media Buying, what Ian’s Keyword Blasting is to PPC. Kinda. Good stuff, Brandon.
@Kiley – Yea it kind of is…
great post. have not tried this… will try now
This is a great post but you can’t forget one thing… there’s always the chance that nothing in your test will convert well enough to make a profit on that source, haha. It’s happened to us before a lot.
While this is true, don’t fotget that you ad, if it’s a text ad or image ad, needs a reasonable amount of traffic and clicks before you really decide which is the better one.
If you test 2 or 3 banners ad the time you will see quicker which of those is the better and then you can replace the less performing ads with new ones.
Do you agree on this approach or do you launch more then for example 3 ads at the time. (it probably depends on the amount of reach and traffic you would expect to get out of the campaign.)
It’s a given that you need significant traffic and statistically relevant data. When I test like this its always on high volume traffic.
That’s what thinking outside the box is all about, thanks for sharing your ideas. 🙂 This method would require a lot of testing to find a champion.. Making your demographic choose which offer they want to see/buy, smart move.
How do you define “significant traffic”? How many uniques per day?
Also, what is a good software to automate the banner rotation and picking a winner?
Thank you.
@Ilya
There are a lot of ad managements systems out their which you can use to manage you ads. Bigger advertisers/publishers & media buying companies use systems like:
DART – http://www.doubleclick.com/products/dfp/index.aspx
Atlas – http://www.atlassolutions.com/
Admeta – http://www.admeta.com
Smart AdServer – http://smartadserver.com/smart/uk/
But if you look for a system that’s free, check out:
Google Ad manager – https://www.google.com/admanager
OpenX – http://www.openx.org/
@Ilya: “significant” can have different meanings for different people, but in terms of traffic 1mil impressions per day is pretty significant. OpenX is great for tracking impressions, clicks, CTR and gives you the ability to deactivate losers very quickly. Unfortunately, this is not always an option, depending on the site you’re advertising on.
Thank you for the information you provide through this article, interesting and good content makes me want to visit your other articles.
@Madsurfer: Thank you for the links. I tried only OpenX hosted so far 🙂 Will check them out.
@Mike: Wow, that’s definitely significant. And costly, too 🙂 Thanks for the determination.
yeah, it works pretty well sometimes. 😉
@madsurfer thanks for providing the list of ad serving platform.
Good post.
Few thing I would add to the post is that…
1. Only test this method with one website or one traffic source. I would stay away from self-serve platforms.
2. More banner variations you have, the more money it will take to build S/S data. Any site or traffic source that can give you above 300k impressions a day should be good.
3. Most websites have 300×250 and 728×90 ad placements. Focus on testing any of these three ad units. 728×90, 160×600 or 120×600 and 300×250.
3. I think if you pick atleast 5 offers from different niches while matching the demographics, you may have good chances of finding a winner.
4. Always have your ads above the fold. They may be costly but CTR will be most be high too.
-G
With media buying, you should always be using your own ad tags. The free tools like OpenX are good but majority of publishers will not accept it. For PUBs to accept you ad tags, so you can optimize your own campaign instead having to rely on them for everything (which also can kill your cmapaign), you should be using paid servers like DART, Atlas, etc.
-G