What is a microniche? Imagine, if you will, a niche for “credit cards with low apr” (those words in that sequence, not low apr credit). According to google, there are 70 pages in the whole world that come up for that page. Getting to the top of that page would be, to coin a phrase, easy. According to the google external keyword tool, there’s about 36 searches per month. About 500 a year.

Most affiliate marketers would avoid such a small niche. What if you didn’t? What if, instead, you built a page with excellent SEO and excellent information on these credit cards with low apr ?

Honestly, what if you took wordpress, whacked together a few pages, and put in some advertising for, you got it, low apr credit cards. You put it on your site and forgot about it.

Suppose your domain purchase is $8. And you put it on your current hosting plan as a parked domain. It won’t get enough traffic to warrant its own hosting plan. It might not get enough traffic to warrant its own domain. But if you’re aimed at selling credit card signups and the click - to - action rate is 1 percent, and the action pays $10, if you sell one credit card every two months, you’ll be making 52 bucks a year (((12/2) * 10) - 8). Ten of those sites would be 520. A hundred, well you see where I’m headed.

True, true, that’s not a lot.

What I’m saying, though, is that if your shotgun approach of finding something with 250,000 sites and 3 million searches per month isn’t working, maybe you should work on something smaller.

I should take my own advice. I’ve been complaining about Easy Spyware Removals not getting much traffic. However, it’s not very seo-optimized, and it’s not very sales oriented. It’s just a dumb bunny site. That specific keyword gets about 140 searches per month. There’s about 3000 sites that pop up in google for that search, and honestly, most of them suck. Of course, the niche is pretty competitive.

After reading how to create a twitter bot in 5 minutes, I realized that I could combine this with simplepie and create a twitter spammer in less than ten minutes. Well, ok it took me fifteen.

So the steps were:
Get something to say,
Get someone to say it to,
Say it.

First, I needed an account with something to say. I used my fsbo homes already-spamming twitter feed. ( . http://twitter.com/fsbohomes ). It just takes the new stuff in the fsbo auction site’s rss feed and twitters it via twitterfeed.

As it turns out, the public feed probably isn’t the best tool for this; a better tool would be twitter’s search at http://search.twitter.com/ . Note the rss feed icon on the results page.

Here’s what I came up with:

Ta-daa. I set it to run every five minutes and it caught about one twitter every few hours.

One word followed by the next word is very unreadable, but basing your markoving on two words (and then randomizing the following word) is a better technique.

All code in these examples are PHP code. There’s a flaw somewhere in it, perhaps you can spot it.

Here are the steps that this markov generator goes through;

1. Read in a bunch of topical content.
2. Keeping it in order, split it up into an array of words.
3. Go through the array, taking a pair of words:
   
   From “The quick brown fox” you’d get “the quick”, “quick brown”, “brown fox.”
4. Now that you have the list of word-pairs, get a list of words that follow this pair.
   
   Keep that in an array, but don’t make it unique. This way it’ll weight the random selection. By which I mean, if you have a word pair like “credit card” and the array is “users, debt, debt,” the word “debt” will have twice the weight when you do a random selection from that array. Which is good, because in human-written content, the word “debt” appears twice as often as the word “users” after “credit card.”
5. Start the content. I chose to start with the first word pair of my input, but it could and should be randomized.
6. Take the last two words of the output ( in this case it’s the first word pair) and look up the array of words that follow. Then take a random word out of that array, and append it to the output.
7. Is it long enough yet? If not, take the last two words of the output again, and look up the array of the words that follow those two words. Keep doing this until it’s long enough.

Here’s the script I came up with. Feel free to use it or abuse it. It’s got some extra verbose variables in it because I’m trying to track down a flaw. It randomly (heh) dies because it can’t find the word pair in the input. But that should be impossible, because I’m selecting it from the input. Hmm, it’s possible that the “find me” part doesn’t have the punctuation that the “input” has, and so it dies.

If I keep putting together code for this site, I’ll have to install one of those pretty-code plugins.

So in a perfect world, I’d find that flaw and then maybe substitute some links into it from this stuff to a real site ( for instance, $output = str_replace(”bad debt”, “bad debt“, $output ) ) and then use xmlrpc and post it to a splog and ping pingomatic.

Hopefully this gets you going in writing your own markov generator. It’s important to write your own stuff, so you know how it works and to keep your stuff from looking too much like someone else’s stuff. My markov generator might generate different output ( different looking output ) from Nickycake’s generator. It’ll definitely be different output, because it’s all randomized. You can feel free to use the code above any way you want, I’m sure that since it’s unfinished, your touches will make it unique.

I’ve long been a fan of the National Novel Writing Month, NaNoWriMo. I don’t think I could keep up with the novel writing because of my other “irons in my fire” but the Old Blind Ape is hosting a shindig. We’re writing 50,000 words of content; which could be ( does the math ), 100 500 word articles for various sites.

I have sites that could use a few 500 word articles. It’ll be good to get the content running on them. And if I get the pump primed well enough, I’ll be able to talk myself into writing more often.

For instance, I don’t think I could come up with content for project car catalog, but with this, I should be able to come up with bloody something.

First, thanks and props to the cakes. He’s keepin it real.

First go over to Market Leverage and sign up there. Sure, you can use my affiliate link, or just type it in your self: marketleverage.com. Or use Nicky’s affiliate link. Whatever you want, it’s a free country, you know? You need an ad before you can join MySpace Advertising.

Logging into your Market Leverage account, click on “find offers;” it’s in the top row of the menu, white on blue. Click on that and then you’ll get a filtering sort of page. If you look down a little bit to “search by category,” use the drop down box and display “$ Current top performers” (it’s the top option). Javascript will submit it and you’ll get a list.

Ok, scan down these and find an offer. You’re looking for something you can imagine a demographic for, and something that allows web media types (the blue box in the “short description” column). I picked “TriSlim.” You can too.

Look. Most people are gonna hide their creatives, their programs, their niches, and their bids. With good reason. But I don’t have a lot of expendable cash, and I’ve already spent my allotment. So it’s dead to me. Ok, if you click on the second column link, it’ll take you to a page where you can see the banners. You’re gonna have to go t hrough them page by page; there’s eight pages for the trislim offer.

You need a graphic that’s either 728×90 or 300×250. Just find one. You don’t want one with animation, it has to be .gif or .jpg or whatever. The graphic I picked was an animated gif but when it imported, it was static. So you’ll want something that won’t lose anything if it’s staticized. ( ie, if your graphic flashes “lose weight” and then “free trial” and then “click here” your targets won’t know it’s a free trial, kwim? )

Save the graphic to your local machine. Write a webpage that looks like this

That location should match the href of the link in the advertising you’re pushing.

So if the text box beneath the offer you’ve decided on looks like this

then take out the “allybranding” piece, from http to the empty subids at the end; &s= and put it in your little php page. Save the image source to your hard drive. (right click, durrr ). Upload the php page somewhere so you can access it.

Ok, sign up to MySpace Advertising with your ad. I have to use MS Windows and Internet Explorer to view this page. It sucks ( it’s their flash thing). Anyway, when it asks for your image, upload your image. When it asks for the destination, put in the location of that little php page you wrote (http://somedomain.com/offers/trislim.php for example ). It’ll ask you to save that information. Walk through their little wizard. When it comes to “demographics” you’ll want to look at who would buy it. Since I was selling TriSlim, a diet product, I went with women from 30 to 40. You have to go American, but you can pick either gender ( or both ) and a range of ages. The advertising wizard hung a moment while it checked everything and then asked me how much I wanted to spend.

You can go with the minimum suggested bid. You can go lower, you can go higher. I started higher. My original suggested minimum was $0.36. I went with $0.40 to drive a little traffic. As MySpace figured out that people would actually click on it, the minimum went down. My current minimum is $0.34 and my bid is sitting at $ 0.35. Then you set a limit as to how much you want to spend. Because I was chicken, I set that to $40; 100 clicks. I expected at least one purchase ( that’s a 1 percent success rate. ).

The ad took three business days to be approved. I don’t know why. Maybe they were busy. I submitted it on Friday and it was approved on Wednesday. Thursday morning I had one sale. I dropped my bid by a penny and increased my budget a little. To do this, you have to suspend your campaign and make the changes and re-submit, but approval was automatic ( probably because I didn’t modify the advertisement ).

I got a second sale with my budget capped at 60, so that was $56. My income was slowly catching up to my outlay. I figured I’d boost it to $80 and call it done, so I dropped my bid again, lowered my bid and went to bed. Today, it’s up to four sales, for a total of $112, and I’m done. ( I don’t have any more money to put in ).

You, on the other hand, knock yourself out; go put in a bunch of money and make some dough!

Name Boy has a better spread of optional names; Domainsbot will sometimes wind up with things like “easyhomerenting” and “myhomerenting” which are sort of bland.

It was about five or six years ago when a client encouraged me to consider a name spinning tool. He had a domain registration service and the one he used was, frankly, terrible. I always got hung up on how to break up a domain request ( houserenting) into the component words, but these namespinners rely on the user’s input of keywords.

Give nameboy a try; it’s a lot of fun.

First and foremost, I’m earning about $100/month or so from a ebay site. I’m also earning about $30 from . That’s it.

A buck a night? I’m not so sure. But I can also say that I’m making a fair amount of interesting discoveries. I was doing some research on “auto restoration,” to find new articles and information in a niche when I tripped over “camaro headlights dot com” … fascinating, huh? A specific model of car and a specific thing someone might need to replace. So perhaps I should focus my niches?

I have one niche, for instance, at “project cars” with pages beneath it for makes or states. What if I worked the “ford” angle and aimed at “restoring ford trucks” or something?

My work on this front has stalled over the last four or six months due to working with a very focussed client ( Jeff ). His site, hyperdesk.com is finally complete. And he’s happy, which is the more important issue. His checks ‘ave cleared, and I’ve even put together an affiliate site at hyperdeskthemes.com and I’ve bought a few more domains in preparation for more affiliate sites for jeff’s cool products.