Using affiliate banners on your site is a great way to make money. One of the big problems is choosing the best affiliate banners to use. Related to that, is the biggest mistake affiliate marketers make. Rotating too many banners.
This all boils down to average human response to web advertising. Your site may do better or worse because there are things you can do to make your site perform better (or worse) with pay per performance advertising. So for the most part, marketing with affiliate banners is a numbers game.
You need to expose a lot of traffic to your banners and the likelihood of having someone click on a banner and then make a purchase. Let's use the numbers below as an example.
| Expected | |
|---|---|
| Click Through Rate (CTR) | 2% |
| Conversion Rate (sale/lead) | 2% |
| Commission | $10 |
So, using those assumption. To get one commission of $10, you would need one sale. To get one sale, you would need 50 clicks. To get 50 clicks, you would need 2500 impressions. In this case, the expected value of 1,000 impressions is $4. Applying the law of large numbers, the more real impressions you have the closer you will get to achieve that expected value. Bots, click exchanges, traffic generators, traffic exchanges don't count.
Now, lets look at how to use that to our advantage and avoid making a common mistake with banner advertising.
How much traffic do you need?
This is a tough question. Not all sites perform the same, but given normal circumstances the theorem of the law of large numbers (LLN) generally holds true. So we need to determine what would be considered a large number.
If you visited the Wikipedia page on the Law of Large Numbers theorem, you might have been confused by some of the equations but don't let that confuse you. Just know the more times you do something honestly, the more your actual results are going to converge with theoretical results. In an ideal world, somewhere around 1,000-2,000 would be a large enough number to start seeing convergence between your actual results and theoretical results.
Click Through Rate Expected Value
CTR is pretty small in most cases. So you're going to need a lot of impressions to reach the expected value of 2%. Below are the results of a simulation of 5,000 banner impressions and clicks with a CTR of 2%.
You'll notice that at around 1,000 impressions, the CTR starts to stabilize a bit. If you know that the advertiser has an average CTR of 2%, you start getting a decent impression of how your site's CTR will perform after about 1,000 impressions. 2,000 would be a better goal. I ran different simulations using different expected CTR and there wasn't much difference in the number of impressions it takes to converge.
Conversion Rate Expected Value
Clicks alone don't yield any money with affiliate advertising. You need the people to click to actually buy something once they click on a link. This is known as a conversion.
So let's take a look at the following chart showing conversions and conversion rate from the same simulation as above.
Just one conversion is not that impressive. So what happened? Did the law of large numbers not work? This is a problem with the definition of conversion rate. Conversion rate is the rate at which each click converts to a sale. So we need a large number of clicks, not a large number of impressions until we can see how we're performing. In this simulation, out of 5,000 impressions, we only had 93 clicks. 93 clicks is not a sufficiently large number. So lets extend our simulation to 100,000 ad impressions and see what that looks like below.
The highlighted range is between 1,000 and 2,000 clicks. That's where the actual data is beginning to converge with the expected value. At 1,000 to 2,000 clicks you should have a good idea of what you're conversion rate is going to be. To get that many clicks in our scenario, you're going to need 50,000 to 100,000 impressions.
Law of Large Numbers to Boost Ad Revenue
So how do you use this information to help you improve your affiliate revenue? The secret to getting better affiliate revenue is to put ads that perform the best on your site. This involves testing different advertisers and banners to see what's effective.
As you can see, you need a lot of impressions for the results of your testing to be meaningful. If you don't have a lot of traffic, you need to make sure you have the most effective banners and the best placement. This can take some time to figure out.
You can increase click through rate by playing around with ad placement. Make sure your ads are someplace where most of your visitors will see them. This means above the fold, but that's not the only place to put ads. Put affiliate banners near search boxes, comment forms, menus or any place where the visitor is going to focus their attention. If your ads don't get seen, they won't get clicked. The next thing is to pick ads that your visitors would be interested in. Don't just go for ads that have the highest payouts. Just because they work on other sites, doesn't mean they will work for you. An ad for cameras may not perform that well on a site devoted to writing for example.
There's only so much you can do about conversions. A lot of that is up to the advertiser. If they don't have a good landing page and make it easy to complete a purchase, you might have a hard time getting conversions.
A couple of things you can do to increase your conversion rate, is make sure the banners on your site would be of interest to the visitors you are getting and try to link as close to the page where the visitor will actually make a purchasing decision as possible. Don't just link to the home page, link directly to a product where they can click to make a purchase.
Don't Rotate Too Many Banners
Commission Junction is one of the most popular affiliate programs out there. They offer a feature called a Smart Zone. It allows you to place a bit of code on your site, select the banners to display, and it will choose one to display based on the weights you assign to them.
The problem with this, is people get excited and join a lot of advertising programs no Commission Junction because they don't know what would be the best ad to display or they get excited by how much some of them seem to pay. On a site that doesn't get much traffic, that can really cause problems. Lets say your sites gets 100 impressions a day and for many people, 100 a day is a goal not the norm. It would take you 20 days to decide if you're getting a good click-through rate or not. If you added a Smart Zone with 2 ads, it would take you 40 days until you get a good understanding of how effective the ad will be in terms of CTR. Ten ads and it will take 3 months. That's just to decide if you're getting a good click-through rate. It would take longer until you see what your conversion rate is going to look like. If you don't have a lot of traffic, you should not rotate banners.
Instead of rotating banners, consider putting a number of smaller banners in the same space. For example, instead of a 728x90 leaderboard, you can put 5 120x90 ads with 12px space between them. This way a big area is filled up so it attracts attention, it gives the viewer more than one thing to look at and different options that might make them find something interesting. You're also testing the performance of 5 ads instead of just one. This will help you figure out which ads do better. If you find one that just does exceptionally well, you can then replace the 5 with a leaderboard.
Posts
0 comments: