the blog of
Shawn Plep

How to Use an Affiliate Datafeed (Effectively)

In the world of affiliate marketing, there are a whole lot of tools available. Merchants want you to sell their stuff for them. And even if you (as an affiliate) are unsuccessful, they still benefit due to the inbound links and citations you provide to them.

So, if you’re an affiliate, you need to understand how to properly use the tools provided to you so you can benefit from them as well (and actually rank, and sell stuff). The most important (and common) tool you can find is the datafeed.

A datafeed is usually a CSV file containing the products from a particular merchant. The fields consist of the product name, description, price, link, and maybe a few others. Merchants on affiliate networks such as Shareasale will typically have datafeeds.

Here’s how you should use a datafeed, with SEO in mind…

• Change the content. As you pull in a datafeed, you should always make some alterations. For one thing, I like to “clean up” the data – because I often find weird mistakes, strange abbreviations, or undesirable formatting. Secondly, I like to replace words and/or phrases with different ones. I do this across the board, just because it will make my data slightly different than what everyone else has.
• Add to the content. You can get as fancy as you want with this, but the fancier the better. At the very basic level, you can make what’s already there more “wordy” – just another variation on replacing words (as noted above). But the better option is to add totally new additional content. If you can scrape content and add a snippet of info to your product entry, your site will have the datafeed content plus the scraped additional content. One example of this (that I do) is a t-shirt site I own. The products all come from the merchant’s datafeed, but in addition to that data I write an additional few sentences or paragraphs about the shirt. If applicable, I embed a related YouTube video. Get creative and you can probably think of tons of ideas like this, for your datafeed.
• Re-format the content. Again this is basic, but can only help – especially when compounded with the other changes. If the datafeed uses bullet points, take out that formatting and use CSS or just dashes and linebreaks. If the description comes before the price, put the price after the title but before the description. If there is no “sale” price…add one! Put a strikethrough on it and put it in red. Pull out the first hundred characters of the description and make it into a short description, and re-order the sentences – and use that on your site’s landing page to link to the product. Get the idea?
• Re-do the images. For images, you should download them (either by actually downloading, or if not possible by caching). You should re-name them with the product name. You should re-create the thumbnails in a different size (even if only slightly different.
• Do not link to the merchant. (If you can help it.) Mod_redirect is your friend. I have been saying for a while now that learning htaccess is one of the most important things a web developer can learn, and this applies to aff. marketing as well. Maybe more. You can avoid linking to the merchant through various means, and if you can’t do it super-clean via htaccess directives you can do it through other means as well (with your favorite scripting language). The end result needs to simply be that all links and all assets either are being served from your site or they look like they are.

I probably don’t have to say it, but if you have scripting skills and can at least find your way around a server, you’re ahead of the game. To do any of the above manually will take a long time and become tedious quite quickly. I recommend learning PHP, becoming familiar with mod_rewrite, and learning how relational databases work. (A little regex can’t hurt either!)

There are other aspects to using datafeeds, but this list covers the big issues and will get you around most of the SEO issues that plague an affiliate. And surprisingly, most affiliates don’t even implement one of the changes on my list! So if you do what I suggest, you may end up in the top 5% of affiliates – thus giving you an advantage in your site’s optimization and avoiding  having exactly the same content as everyone else (including the merchant).

Datafeeds are a starting point, best used as a basis to build upon. To use them effectively, you must enact the fundamental rule of SEO which is to provide value (by adding it, in this case).

Thanks For Clearing Up that Metatags Thing, Google

I’ve thought this for some time (based on writings by some SEO guys I respect) but until now it’s been an educated guess that meta tags (specifically, the keyword tag) exhibit no influence upon Google’s ranking of a page. So I must say thanks, Google, for making it clear to all of us who care. Meta tags are not an all-important piece of SEO.

The headline of a recent Webmaster Central blog entry should say it all: “Google does not use the keywords meta tag in web ranking“. But if you want to know more, you can read or watch Matt Cutts elaborate further.

Now, another minor point: it might not be a total waste of time to at least fill in a few keywords. I say this because of the Bing webmaster blog taking a somewhat neutral position on meta keywords. It makes me think that perhaps Bing “pays attention” to the meta keywords, at best. But I don’t see it as any sort of actual ranking factor whatsoever.

Update: I found this excellent and balanced perspective on the meta tags issue by John Andrews. I recommend it highly.

My Thoughts on Google Caffeine

Edit: When this post was written, Caffeine wasn’t live (for my datacenter area, at least). As of this morning (9/21/09) I noticed that I am seeing Caffeine’s results for all my searches.

I was recently asked what my thoughts were on the latest Google update, known as “Caffeine”. More than a simple algorithm update, it seems Caffeine is actually an upgrade to what powers Google’s search. It would be analogous to placing a newer, better engine inside a car.

You can try Google, powered by Caffeine at http://www2.sandbox.google.com.

When I brew a cup of Community Dark Roast to get my caffeine fix, the caffeine usually gets me going in the morning and speeds things up. I don’t know if that’s why Google picked their name, but searches do seem to be faster. (I tested a few searches between Caffeine and “regular” Google and found searches were often twice as fast. Of course, we’re talking times like 0.52 seconds vs 0.21 seconds. But still.)

The number of results found is unpredictable – sometimes “old” Google has more than Caffeine, sometimes less. Why? I would expect there’s a difference in indexing what is supposed to be more relevant to the query, and I suppose that’s why the reported number of the results in the index differs. This shouldn’t actually affect my search experience too much, since I usually don’t delve into the SERPS much more than page 2. But it shows that there’s a significant change in how things work behind the scenes.

Another difference I noticed was the snippet was different on many results. I like the new ones more. For instance, for the search “square foot gardening” I noticed a few of the results had a date preceding them, while the Caffeine version only contains the text I care about. However, this could be a negative – specifically in the case of blog posts: including the post’s date in the snippet could actually be helpful to a searcher. So I have a mixed opinion on this aspect of Caffeine.

I wanted to see how much better (or worse) Caffeine handled current news. So I picked a few events and headlines from Drudge Report and tried various searches. Sometimes Caffeine was  better, sometimes it was about the same, and sometimes it was worse (showing older or less-relevant items). Overall, I don’t see enough of a difference to cite any real advantages or disadvantages between the two.

You wouldn’t believe how many random/weird/varied searches I attempted, just to see what Caffeine would spit out. One was “baptist church”. Both versions give Landover Baptist Church as one of the top results – and I don’t think that’s really the best result to serve someone performing that search (as much as I love the site and think it’s clever). Also (on that same search) “old” Google delivers some image and video results – while Caffeine does not. In my opinion, the video and image results are a good thing and add to the searcher’s experience.

With some other searches, I definitely saw the Caffeine results as being less-relevant and inferior in quality. But for the most part, this is a subjective judgment.

So, what do I think of Caffeine? My first impression is simply that it’s faster, but no more relevant. I’m sure we’ll have to wait until it’s officially put into use (assuming the engineers are still adjusting and tweaking it) before we can really see what it’s all about. But for now? I’m going to stick with searching via google.com, and will occasionally try Caffeine out of curiosity and to see if the results differ vastly when I’m concerned about the effect my SEO efforts may be having on a particular site.

If you want to easily compare the current incarnation of Google’s search engine with Caffeine, Compare Caffeine is your best bet

SEO Tool Under Development: SaturationChecker

I constantly have ideas, but I rarely have time. (Not an issue – if you don’t need to sleep. But I need my beauty rest each night.) Sometimes, I get a bit of bonus time when I can actually work on one of my ideas.

That recently happened with my idea for SaturationChecker. It’s a simple script, but what it accomplishes can save the SEO-minded site owner a great deal of trouble and help them improve their site.

Here’s how it works:

1. User submits the URL for the site they wish to check
2. SaturationChecker (SC) then attempts to find either  an XML sitemap, or an HTML sitemap file, or an RSS feed. (In that order.)
3. If a source for URLs is located, SC then checks each URL successively and writes it to the database. Each URL is recorded along with its status, and whether or not the URL was found in Google’s index.
4. The unindexed URLs are then available to the user, so they can be targeted for promotion or perhaps simply rotated to the top of the XML sitemap file.

It’s up to the user to find a way to get the unindexed URLs “found” by Google; what SC does is simply shows where the deficiencies lie in a site’s index saturation. (So, if a site had 100 pages, yet only 50 of them were in Google, that site would have only a 50% index saturation.)

This tool was developed for my own use, and as such it’s a bit rough around the edges. (It works for me.) I do wish to polish it and then make it available for others’ use at some point in the future.

Some additions I will be adding will be:

• The capability to track if a URL is added to the index, then dropped again. (This may indicate the URL was found, and added, but then dropped due to an on-page problem or simply a better page coming along from a competing site.)
• A percentage report, as given in the example above. Very simple to code, but I haven’t added it as of yet.
• Prettiness. The thing has no style, and if I release this for public consumption I want it to look nice!

Since I haven’t released this yet, consider this post as “pseudo-code” for you coders out there to make your own. It’s a fairly simple script – I began it one night in my hotel during this year’s SEOmoz seminar and had the basic functions, and finished the rest of it today. (And I happen to be a coding perfectionist.)

Additionally, I’m open to new ideas – so feel free to contact me if you have any suggestions and I’ll see what I can do.