I got a scorecard, but can I give one?
Today I got a scorecard report from Overstock. Actually, this is the second one. A few days back they first sent a test email and I was expecting them to start sending the real ones. Finally it arrived today. I am not doing as good as just a few months back. I know things are not that good with the economy and all but the real problem for my account not doing well is because I use Overstock affiliate data feeds and for more than 2 months now, the feeds are not getting generated by them. I even sent an email a few weeks back and one person promptly replied that they are looking into it. It’s been a while now but still no fix to the problem. So, my website is filled with old data. As a result, many of the products are no longer available, or the price is different and so the bounce rate has been high. What’s worst, Google sees stale data and my site gets penalty.
Now I am all fine with receiving a scorecard from Overstock, but can I also give them a scorecard on how they are doing with their affiliate program? lol. Actually they are good except now and then. First they stopped giving the list price in their feeds to which another person replied that they are looking into it and that has not been fixed for almost a year now. Second, the feeds completely stopped. I hope they fix these two issues soon so that I can increase my score.
How are affiliates using Twitter?
Twitter has reached a milestone of processing 50 million tweets a day. And the volume of searches on twitter apparently surpassed that of Yahoo! and Bing combined. Twitter now has to be part of any marketing strategy. So, what are affiliates doing about it? At present I don’t have any experience with this. I have a few ideas though, but I am not sure how effective they are going to be. For example, when processing a product feed, detecting good deals and tweeting them using automation can result in several tweets, but could be considered a spam. Also, if you are tweeting hundreds of such tweets a day, who would want to follow you? I think, the strategy should be in such a way that the affiliates post only the best deals and offers and sometimes even attract the audience to their own websites rather than directly to the retailers. In process, may be it would be possible to gather some following on twitter and build some brand and loyalty. But that’s not something that’s easy for most affiliates.
I would love to hear what others think of Twitter as a marketing platform for affiliates.
I wish the commission rates are a little bit more
Having been both Amazon and Overstock affiliate, I can tell that they both have completely different commission structure. Amazon’s is both category based and performance based. That is, for Electronics it’s 4% (and computers only up to $25) but for most others, it’s up to 8.5% based on how many items are sold in that month. A few other categories get as high as 10 and in some cases even 15% commission, but don’t be fooled by those (either the item prices are very low or it’s hard to sell those or Amazon want to really push those hard when heavily competing with competitors).
Unlike Amazon which has it’s own affiliate infrastructure, Overstock relies on Commission Junction’s tracking system. I am not sure if CJ has performance based tracking and calculations, but Overstock only provides category based. The reason that prompted me to write this post is, for the last few days I didn’t have any commission from Overstock and suddenly I saw today that I got $5.38. When I looked at the order amount, it’s $537 and few cents. That’s when I really paid attention to the rates and realized that Overstock only pays 1% for electronics, 3% below what Amazon does.
I know Amazon’s cookie lasts for 24 hrs. Now if only someone can tell me what’s the cookie time for Overstock.
Upselling & Cross Selling For Affiliates
Big retailers do what’s called upselling and cross selling. Upselling is the concept of trying to sell more expensive alternatives. Cross selling is a concept of selling accessories and other related items. In both cases, the essential goal is to make the customer spend more.
In the pre-internet age, the sales people mostly used their knowledge and experience to do this. But in the online shopping, retailers use sophisticated algorithms to do this. In the simplest form, retailers could just have some employees defining these upsell and cross sell relationships. In a little advanced form, retailers could pick up products within the same category but higher price for upselling and products from accessories categories for cross selling. An even more advanced form actually analyzes the click log and orders. By looking at the click logs, it would be possible to identify what the user initially wanted to buy but instead ended up buying a higher product (or lower for that matter and just flip the algorithm). Similarly, in an order, users typically buy the main product and it’s accessories. So, it’s possible to come up with cross selling relationships through this.
But for affiliates, much of this information is not available. And employing manual methods may not be easier. Here are some techniques I use.
1) Hot products: If the price of a product increases from the previous day, I assume that the demand for the product has increased and hence the price increase. This is not always true, but as we don’t have much info, we have to come up with some heuristics and this seems to be reasonable.
2) New products: If I find a new product in the product feed not tracked before, I consider it a new product. Over a period of time, this tracks very accurately.
3) Products on Sale: If the price of a product decreases from the previous day, I consider that it’s on sale. Or discounted. Sometimes the price changes are due to error in the first place. But again, for large set of data, this usually tracks well.
So, in addition to displaying the search results, I also have these promotional units of items. At present, I don’t have a mechanism to track the efficiency of this. I am in the process of modifying my scripts to track more accurately. I will report on that in a few days once I have more data.
Best Single Transaction Commission
As an Amazon affiliate, it would be possible to know what items are ordered but not which items are ordered together in a single transaction. While this info may not be useful much to small affiliates, in Commission Junction it’s possible to see this info. Today I woke up and pleasantly surprised to see a commission of $154.94 and when I clicked on the link, I saw that there was only one transaction. However, it had 8 items ranging in price from $261.99 to $779.99. The total order amount is $2,437.95. This translates to 6.35%.
My Overstock affiliate fee has been between $30 to $40 a month. This new transaction is going to push the numbers up for this year.
One interesting thing is, in my previous post comments I mentioned how my website traffic is low and I don’t bother to do much SEO and instead focus on CRO (conversion rate optimization). This above transaction proves the point. Here are the stats for yesterday.
Number of clicks: 5
Number of sales: 1
But this is not the interesting detail. Of the 5 clicks, only 2 are from the US. 2 are from Russia and another one is from Philippines. So, it’s 1 out of 2 unique users that resulted in a sale. I will be honest, it’s not a norm but an exception. However, it’s not an exception that happened by chance. Rather it’s due to how I try to increase the conversion rate.
First a bit of digression. What is the safe distance to maintain from the front vehicle when you are driving at 65 mph on the freeway? One of the component that goes into coming up with this safety guidance is the reaction time. As soon as a lot of information is displayed, it takes time to process it. How long is it going to take to browse through 7 or 8 products each displayed with an image and 4 to 5 lines of detail?
Now, one thing I mentioned in the comments section of the previous post is that I look at the keyword with which the user visited my website and then pull information most relevant to that. I don’t advice this idea for most websites. But for a website that is showcasing a product feed from a retailer, this is very desirable. Initially when I setup my affiliate website, I used to get some traffic through Google but realized some of them never translated into clicks. Up on further research, I found out that when the users visited my website, they didn’t find what they were looking for. How can this be possible, that too with Google, you might ask. The reason is, the time lag between crawling and searching. If the search is against a few days old snapshot, but Overstock data feeds are changing on a daily basis, then the following can cause the users to not find what they are looking for
- New products are added and some products are removed from each category. So, a product showing up on 2nd page might move to first or third page. But if the search engine send the user to the 2nd page, they simply eyeball and leave. Even if you provide a search on your own web page, most users don’t have that much patient. They simply go back to Google or other search engine.
- Sometimes the price of the product changes (it may increase or decrease). So, if you have a directory that allows you the users to browse by price, again, the user might be ending up on a page where the product no longer exists because of the change in price.
These are some of the reasons why the products keep moving from page to page. So, it’s not your fault, it’s not the search engines fault and certainly not your retailers :). That’s why it’s important to take that extra step and provide the most relevant information to the user. So, what I do is, if I identify a search keyword from the referrer url, then I use that to do the search and present the results to the user. Of course, I also prominently display the internal search box with the keyword I used and the user has the option to change the search options.
Because of this, the users never miss out the information they are looking for. Now the interesting thing is, in the two US clicks I mentioned above, both times the product they are looking for are actually in the same page they landed up. However, the products were in the 7th position on one page and 8th position on the other page. But the time it took for the users to click the link and go to Overstock was 1 second and 3 seconds. Now, I agree that there might be some difference in time recording between my website statistics counter and my click tracking software as they are two different servers. Even then, I can safely assume that the time spent on my website is less than 10 seconds!
As a naive affiliate, I used to think that the longer time the user spends on my website, the more revenue I can make. But that’s not the case. You may be able to make a few extra cents from any advertisements that you place, but you can potentially lose out big on the affiliate revenue.
Most visitors to affiliate income focused websites are there not out of loyalty. They are there to get some more info on the products they are interested in buying (like may be better price, comparison of multiple items, reviews, features etc). And chances are they probably made up their mind on what to buy as well, just want some extra confirmation or information. So, these people are not interested in spending too much time on other websites. They just want what they are looking for. So, had I not done the search based on the keyword they are looking for, they would have to scroll through at least 6 products to find what they are looking for. My search results are organized such that the most relevant products for the search keywords are displayed first. So, if the original page had the product in the 7th position, based on the search, it might show up first or second along with a few more relevant products.
All this required a lot of observation, planning and programming. But as you can see, it is paying me off reasonably. Now I need to find out more ways to increase the conversion rate. If you have any ideas, please feel free to share them in the comments.
Life As An Overstock.com Affiliate
From my blog name, you know I started off as an Amazon affiliate. I had seen good 2008. 2009 was not as good, partly due to economy and partly due to Amazon affiliate web services (which now they call product advertising) api changes like signing the requests. I never had the time so far to go back and fix my scripts. My revenue for 2010 is horrible.
The only good thing that happened this year is a decent revenue from my other website that uses Overstock.com as the main source of revenue. It’s actually a website built just to showcase Overstock products using their data feeds. Here are some key statistics that I have been reviewing today, mid-way through the year, and wanted to share them with fellow affiliates.
First, Overstock.com uses Commission Junction (they changed from Linkshare). So, the numbers below are coming from Commission Junction reports.
Total number of clicks: 990
Sale Amount: $4427.79
Items Sold: 38
Sales : 21 (One order can have multiple items, hence the difference)
CTR: 3.06%
CR: 2.12%
EPC: $28.70
and the most important metric (or show me the money :)), commissions: $284.12
From EPC, each click is generating $0.287 which is not bad. Actually, I had some stupid mistake in my scripts that probably cost a lot of money. Realized it half way through this period and ever since things have actually been much better. I will probably post these details in the future where they are going to be more accurate.
One thing is, I don’t directly link to Overstock. Instead, I redirect the click and hence internally track each click. So, some of the additional details that I have are
No of clicks: 969
No of unique visitors: 712
No of clicks (US): 712
No of unique visitors (US): 540
So, the EPC is actually even higher if you consider only US visitors. My internal record of 969 tracks very closely to what CJ reports as 990. Usually my internal count should be more than CJ if at all due to network issues, but the reason you see otherwise is because I have cleaned up some clicks in my internal database which I think are related to some search robots which didn’t follow the robots.txt instructions.
Both in 2009 and I am sure in 2010, I would be making more money through Overstock program than through Amazon’s. I am not hinting that one is better than the other. It’s just that my current websites designed for Overstock feeds are doing better than those designed for Amazon’s web services. I need to spend some time updating my Amazon scripts and see if things are going to improve. Back in 2008, I used to have people buying LCD/Plasma TVs fetching as high as $80 per transaction. Gone are those days, along with the economy, Google’s ever changing algorithms, changes to Amazon web service APIs and my laziness to fix them.
What Percent is 3rd Party Sales On Amazon?
Amazon affiliates immediately recognize that they are paid for orders of items from Amazon as well as 3rd parties. Also, when a 3rd party item is ordered, Amazon immediately gives the commission while for items sold by Amazon, the commission is given once the item is shipped.
So what percent of orders are from 3rd party merchants anyway? If you want to know the answer, this job post by Amazon on craigslist has it.
“Have you ever noticed the “See more buying choices” or “New & Used offers” while you’re shopping on Amazon? Maybe you’ve sold an old textbook or extra movie yourself. Selling on Amazon is one of the fastest growing businesses at Amazon.com with about one-third of all items currently sold originating from 3rd party merchants. The Merchant Technology team drives this multi-billion dollar business and develops solutions that enable millions of sellers around the world to sell on Amazon’s Marketplaces.
We are looking for senior engineers with a strong sense of ownership and a passion for delivering creative solutions for complex problems on an unprecedented scale. As part of the team, you will be given the chance to have a significant impact on our systems, our business and most importantly, our customers as we take on significant challenges that can reshape the ecommerce industry.
Successful candidates must also be innovative, flexible, self-directed, and able to design and write high-performance, reliable, maintainable code. The ability to function at a very high level in a fast paced environment along with a team of very talented engineers is essential. If you enjoy working in a dynamic environment to deliver world class mission critical systems, this may be the career opportunity for you!
Requirements:
• 5+ years of industry experience developing large scale distributed systems
• Fluency in C/C++ or Java
• Familiarity with Ruby, Perl, JavaScript, AJAX, XML/XSLT, SOAP, SQL, Oracle/Berkeley databases, caching technologies, web protocols, Web services technologies a strong plus.
• Solid understanding of Object-Oriented design and concepts
• Experience developing software in a Unix/Linux environment
• Excellent communication and analytical skills
• BS or MS in Computer Science
Please send your resume to career@amazon.com. Position located in Seattle, WA. “
Sales Tax & Affiliates
Sales Taxes are local and collected by each state from the companies that sell products to residents of that state. With sales tax in some states as high as 9% or 10%, this is a significant income to the states. However, the internet has changed all that. Unlike their brick and mortar counterparts who have to physically reside in a state to sell products, the pure internet players like Amazon can just operate from a single state and sell products to anyone in any state. As these pure internet players don’t register with the individual states, there is no need to pay the sales tax to those states. Of course, a decade back, internet retailing was in it’s infancy and the revenue loss wasn’t much. But as the ecommerce increased leaps and bounds, the amount of sales tax that can be lost is significant. States like California have introduced what’s called a Use Tax and shifted the burden of paying the sales tax from the companies to the tax payers.
Use Tax essentially tells a tax payer to figure out all the products they bought without paying sales tax and pay it at the end of the year, typically along with the income tax filings. It is unclear on how effective the use tax is. There is no easy way to enforce it and I don’t think it the collections would be as good as taxing at the time of the sale.
So, what can be done to make internet retailers to collect and pay the sales tax? One way is to change the definition of who affiliates are. Affiliates are actually not employed by the retailers. Anyone with a website can participate as an affiliate and earn a small amount of side income. Internet retailers rely on affiliates to increase traffic to their websites and improve their sales. Obviously, a person who is an expert in a specific subject area can act as a better faceless sales person to sell a line of products than the retailer itself could ever do. Realizing the fact that internet retailers rely on affiliates and that the affiliates are from all around the country, some states came up with the idea that because the internet retailer has affiliates residing in their states, they have operations within the state and hence should pay the sales tax. New York and North Carolina are a few examples of these states.
Recently Amazon has sent emails to their North Carolina affiliates that they would be terminating it’s relationship with them because North Carolina is going to enforce sales tax on retailers with affiliates in their state. What the people coming up with this type of tax collection schemes don’t realize is, on the internet, anyone can promote a product or a website. In fact, many of the 2 million+ affiliates of Amazon are from outside the US. People from India, Australia, Thailand, Canada and else where are participating in Amazon’s US affiliate program and getting referral fee. If every state in the US starts enforcing sales tax due to affiliates, the end result would be that there will be no affiliates left in the country. We already see so many jobs being outsourced to other countries. Now, affiliates are getting outsourced due to the stupidity of states trying to collect sales tax in a wrong way.
An Amazon Affiliate Blog 