Some time ago, I wrote about a way to randomize ads; that is, in a particular position, show one of a number of different ads. With AdSense channels, you can use it to find out which ads (formats, colors, etc.) work better in a particular position.

But now I’m going to take you one step further.

What about randomizing between two positions? Say you want to see if an ad works better on the top or on the bottom of a page. How to show one, and only one of them?

Making the first ad show 50% of the time is easy. The problem, here, is that the second ad must have a way to tell whether the first ad was shown or not.

There are, of course, several ways, most of which using some kind of variable to store if the first ad appeared or not. But I think that there’s a better method, which has the advantage of being incredibly simple, and doesn’t require you to store anything.

It’s simple: use the clock.

If the current hour is even, you show one of the ads. If it’s odd, you show the other. Each of them “knows” whether it should appear or not, according to the current time.

Sounds complicated? It isn’t. For instance, using PHP, in the first position you put:

function isodd($number) { return($number & 1); }
$x = date('G');
if (isodd($x)) include "adsense-top.php";

And in the second position:

function isodd($number) { return($number & 1); }
$x = date('G');
if (!isodd($x)) include "adsense-bottom.php";

The “adsense-top.php” and “adsense-bottom.php” files can be mere AdSense code, or can branch itself further, using something like my original tip.

By the way, I use hours instead of minutes (which would provide better granularity) because I want to avoid, as much as possible, the small possibility of the hour changing in the miliseconds between the two scripts. Using the hour value, it should virtually never happen (never happened to me, so far), though it wouldn’t be the end of the world anyway.

Some time ago, in more “dubious” sites (porn, cracks, etc.), you’d usually see ads (usually for porn) and, near them, messages telling people to “click here to support the site”, and variants of that.

These days, such messages aren’t so common, especially in more “legitimate” sites, where AdSense is allowed. In fact, those are against the AdSense terms of service, and, so, people who attempt to use them are eventually banned from the service.

Still, there are still people who ask, using other methods, to click on ads: by mail (e.g. mailing lists), or by asking friends.

However, if you don’t want to lose AdSense forever (including any money you have already accumulated), don’t do it!

Why? Because, if I were Google, and I wanted to battle click fraud, I’d easily find a way to fight those two methods of “artificial clicks”. Now, if I can do it, it stands to reason that they can do it much better, right?

In the “mail to a site or forum’s readers asking them to click on ads once or twice a day” case, what would it cause? Some tens or hundreds of particular users clicking on ads in the same site, regularly. That would, of course,create an easily detectable pattern, which they’d be able to detect with a simple script.

Asking friends to click, on the other hand, will often make them think “I’ll click a bunch of times, to help the guy.” An user clicking 20 times, in a couple of seconds, on ads on the same site? I think they can certainly detect that. I would.

In fact, in the past, I’ve had to ask friends not to do that anymore, because, trying to help me, they were doing it on their own initiative… and I, myself, detected it on the AdSense stats. Fortunately, the “guilty parties” told me about it (after all, they were trying to help), and I was able to ask them to stop.

In short: non-natural ad clicks are relatively easy to detect. Never ask anyone to do them, or you risk saying goodbye to AdSense for good.

So, what was the trouble I mentioned yesterday, which made Google warn me by email?

Apparently, there are some programs for generating traffic and/or clicks artificially, which Google doesn’t like, of course.

This appears to be related to referer spam, which sends a lot of hits with fake referers (such as casino sites), because many people and sites have public statistics, which may mean improved PageRank for the spammer’s site. If they spam thousands of sites this way… it can really make a huge difference.

This new case, however, doesn’t appear to be “true” referer spam, but traffic exchangers or autosurf programs.

For instance, last month the top referers for this blog were:

That’s an image, of course. I would never link to such sites.
Note, by the way, how close the numbers are… can you say “automated”?

I have just blocked those, and others (go here for some tips on doing so with Apache).

It’s quite likely that blog exchangers are considered “bad”, too. And if Google is cracking down on such services, it may be a good time to stop using them (and to edit that part of the series). Most poeople really use them just to surf for credits, though I have been lucky to discover a couple of very interesting blogs, which I read to this day, there.

I’m too tired to do it right now, but tomorrow I’ll have to block a bunch of “bad” referrers, which have earned me a nice warning email from Google’s AdSense department. Googling around, I see I’m not the only one. More details tomorrow… but, meanwhile, one suggestion: check your logs, especially your top referrers. If they’re suspicious, it may be a good idea to start blocking them.

This is related to referrer spam, but it’s apparently something much worse – it can cause your AdSense account to be closed. As I said, more to come.

Meanwhile, WordPress 2.0.4 is out.

And I’ve just bought Roger Waters’ Ça Ira. Yup, it’s opera. Real opera, not a metal opera like Avantasia. Something to listen to, tomorrow.

It’s interesting to observe how well the “contextual” part of AdSense works.

For instance, for my mini-site for yesterday, Chess: basic principles, it’s working very well. Most ads are about chess, chess sets, or board games. I’m pretty impressed by this one.

However, it fails completely for the Portuguese version of the above, Princípios básicos de Xadrez. The ads are in Portuguese, and mostly related to games – cell phone games, soccer games and so on. There isn’t a single ad about chess or board games.

Why is that? The content is exactly the same, only in a different language. Even the meta tags are correct. I believe that, in this case, it isn’t Google’s fault; it’s just that there is nobody in Portugal using AdWords to advertise something related to chess. In this way, I sadly have to admit, we’re a more primitive country: soccer rules here. But I digress.

Another one that isn’t working is today’s mini-site, Zombie Types. This one’s in English, so there should be something out there (horror movies, horror games, etc.)… but the ads I’m getting, even after half a day, are “public service” ads – the kind they use when they can’t find keywords in the site. And these don’t pay.

This can happen, sometimes, when you put the AdSense code before there’s any content in the site, because it can take some time for "Mediapartners-Google/2.1" (the AdSense bot) to come back to the page, so it’s the page’s initial version that counts. But that was not the case here; the AdSense code was the last thing to be added. Oh well… it should be fine in a week or two.

Eric has an intriguing post called The Single Page AdSense Site. It’s, in a way, an even simpler variant of Making Money with Niche Websites, and it’s something I’d be doing in spades if I wasn’t such a lazy guy.

No, I’m not implying that they read this blog. But their newest article, “Channels are made for changing“, reads a lot like part 23 of the Blogging Tips series: “Change your ads“.

You can even go a bit further, and randomize your ads, and then use channels to see what variations work better. Not to mention being unpredictable, and working through people’s “mental filters”.

Inside AdSense, the official AdSense blog, has a post called “Advanced site targeting and referrals reporting“, which explains how to see whether any advertisers are targeting your site. Using channels, you can also see which sites.

Apparently two of my blogs have targeted ads: The Tlog and, a bit surprisingly, Cool Stuff.

I also use AdWords to promote some of my sites (just some pocket exchange per day, though), and I may try out targeting some sites for my own ads, soon. Maybe some popular Portuguese or Brazilian blogs, for A Arte de Blogar, for instance.

“I believe in coincidence. Coincidences happen every day. But I don’t trust coincidences.” – Garak, “Star Trek – Deep Space 9″

Just a couple of days after my randomizing ads tip, Darren Rowse mentions a WordPress plugin to do exactly that. Funny how the universe works.

The plugin may be useful to some. However, I find my method much simpler and quicker – the plugin has 4 pages of instructions! I’m not saying it’s badly made, mind you. But, to me, the point of a WordPress plugin is to make things simpler, to prevent one from having to do things manually. Yet, I don’t think the plugin actually saves work or time, or is easier to use for non-technical users.

Looking at it was useful in a way, though: just like that plugin, “my” method also works for anything made in PHP – not just WordPress -, and isn’t limited to AdSense ads – indeed, it’s not limited to ads at all. You can use it to rotate between any kind of content – Chitika ads, static images, static text, a blogroll, whatever.

No, this isn’t my “brilliant idea” – Chris Garrett of Performancing mentioned it first. But, while he gives a PHP example on how to get a simple random number, I’ll go a bit further…

The idea, as Chris says, is to randomize a particular AdSense ad (changing formats, colors, etc.), and have different channels for each version. Doing it, you can easily learn what works best, and what doesn’t, instead of having to wait for a couple of days while you test each format. This also allows you to focus on a single change – one ad – without being affected by other factors, such as the time, the day of the week, and so on – as at any day, any time, every single ad version has a chance of appearing.

This is just an example in PHP, which works for WordPress blogs, for instance. There are, of course, many other ways of doing this.

1- pick a particular ad. Replace that ad’s AdSense code with <?php include 'adsense-ad.php'; ?>.

2- on the WordPress directory for the theme you’re using, create a file adsense-ad.php with the following:

<?php
  $i = rand (1,8);
  $script = "adsense-ad" . $i . ".php";
  include $script;
?>

Replace the “8″ with the number of ad versions you’re going to test.

3- go to the AdSense options, and generate code for the first version of the ad. Remember to set a unique channel. Save it as another file, in the same directory, named adsense-ad1.php (note the “1″, it’s not the same file as before – that one had no number).

4- repeat the previous step for the other ad versions (changing the channel, and the digit in the file name).

Easy, wasn’t it?

Inside AdSense, the official AdSense blog, has a new AdSense optimization tips for bloggers post. Nothing there is rocket science, and there are no surprises there for regular readers of the Blogging Tips series , but it’s still a good, short, informative post.

I especially like the maps:

although I find that lowering the “search” function a bit, and putting a square ad in its former place, works quite well.

Following my comment on his post (it’s always nice to serve as an inspiration for someone more successful than ourselves ), Eric has a new post in which he offers a trick to use temporary URLs to test ad relevance in blogs – which I had said was not as easy as with normal sites.

His trick is based on the fact that you can add “garbage” to the end of URLs (such as “?a=1), which are ignored by your blog’s software, but, to Google, means it’s a different page. So, you… oh well, Eric says it well, so just go there.

Eric has a recent post in which, talking about the completion of his Invisible Fence case study (very interesting reading, by the way – I’ve mentioned it before), he also mentions a nice trick: using different directories to test the relevance of AdSense ads in different configurations.

Google “reads” the content of a page and displays mostly revelant ads quite quickly (you put the AdSense code in a page, then load it a couple of times, and the ads will probably be relevant); however, if you change the content, it can take weeks for the ads to update their subject to the new content. A consequence of this is that your initial tests may “stick” for a long time – even though you’ve updated the content, or keywords, or section targeting, or whatever, the ads are still related to what that page was like when you first loaded it up, after adding the AdSense code to it.

So, Eric suggests testing it in other places – instead of www.something.com, what about www.something.com/test1 , then, after a couple of changes, www.something.com/test2 … and so on. Each new address means Google will “learn” what your page is about, anew. Only when it’s exactly as you want should you move the files to the root directory.

This, of course, applies more to “normal” sites and mini-sites than to blogs, but can still be of use in blogging.

Don’t include the word “blog” in your blog’s title, unless it’s about blogging.

You may have noticed that this site is back to simply “The Tlog” (which is neutral and doesn’t affect what Google will think the blog is about), losing the ” – a technology blog” part. Yup, I make mistakes like everyone else.

This can help in SEO, too.

It’s amazing that this thing has been posted in at least two places (this blog is one of them), and I’ve seen nothing, anywhere, coming from it. No blog entries. No discussions. Not even comments under the two blog entries. And it works – even my 2 “late afternoon” attempts have been actually making me money, which, considering that they don’t cost me anything in terms of time, effort or money, is great.

Let’s put it like this: imagine you can do something today, that, starting 6 months from now, will make you about a dollar a day. That’s $30 a month… insignificant, right?

But you can do it several times. As many as you want.

So, if you do it 10 times, in six months you’ll be getting 10 dollars a day (in average). $300 a month.. you can’t quit your job, but maybe you can pay a few bills with it, or, perhaps, finally buy that Nintendo DS you’ve always wanted, plus a couple of games.

What if you do it 50 times? 100? More? You see, it’s something that you can do relatively quickly, and which requires no maintenance.

If you’ve been reading this blog for some time, you have probably already guessed what I’m talking about: static, niche mini-sites.

And the “six months” estimate is quite conservative on my part. Consider the following: I have two mini-sites, each made in an hour or two. One of them is a joke, which I created mostly for a few experiments with CSS. The other is a recipes site in Portuguese, which only has 4 recipes so far (I intend to add more, but I keep forgetting or delaying it). Both are very simple, only a bunch of static pages, no images, and one AdSense ad at the bottom of each page. Both were created about a month ago.

And yet they make me money. They get very few hits (they’re niche stuff, not very well SEO’d, and probably “sandboxed” in some way), but for some reason CTR (click-through rate) is quite high – much higher than in any of my blogs. Not a lot – about 20-50 cents a day, in average. There’s certainly a lot of room for improvement.

But what if I made more of them? (I certainly intend to do so.) Better ones, too? They cost virtually nothing in terms of time, work, bandwidth, hard disk space, whatever. They require no maintenance – it’s basically “fire and forget”.

And they make money. And anyone can do it.

Strange, isn’t it? It seems that almost nobody knows about it – I certainly haven’t seen it mentioned anywhere, except for the blog entry that I linked to in my own, and which inspired it.

(NOTE: this is part of the “Blogging tips” series)

This is a point I’ve already made in several past posts (that sounds weird…), but it’s so relevant and so often ignored or even unknown, that it deserves its own part in this world-famous-any-day-now series.

As I’ve mentioned before as well, most ad clicks come from first-time visitors, as your regular readers tend to become “blind” to your ads – they filter them, unconsciously. That is mostly unavoidable – though it’s still desirable to convert newcomers into regulars, as there’s still a chance they’ll click on an ad, but, mostly important, they may link to your blog, make it more interesting with their comments, and tell other people about it.

But there’s still something you can do to reduce the “ad blindness”, and that is to change ads from time to time.

Without wanting to go into psychobabble, it’s a bit like this: it’s as if people have a mental “ad filter”, which works for ads in a position they expect, or of a type they expect. So they don’t even look at those.

The idea, then, is to go around that “filter”, and surprise people, so that they see ads where they don’t expect, or in a different format.

So, my suggestion is this: from time to time (I’d say at least monthly), change your ads. You can do one of the following, or a combination of several:

• change an ad’s size or format
• move an ad to a different position
• completely remove an ad for a while
• add a new ad somewhere (assuming you don’t already have too many, of course – you don’t want more ads than content)
• change an ad’s provider: for instance, replace an AdSense ad with a Chitika ad, or a Blogads one.
• change an ad’s colors – if it used the colors of your site, try using different, more visible colors for a while. Conversely, try changing a “blended in” ad to colors that stand out more.
• if using non-contextual ads (such as the default Chitika mode), try changing the keywords – possibly even replacing them completely with a bunch of new ones (but save the old ones somewhere, so you can compare results).

This may sound like a lot of work, but can help a lot in terms of click-through rate (CTR).

One other reason (and a very important one, too) to do this is that you’ll be able to learn a lot about which ad formats, positions, colors, ad providers, keywords, etc. work better or worse. Remember to configure channels, if available, for each type of ad, so you can track which ones give you a better CTR, and more money.

If you read this blog regularly (and if you don’t, why not? ), you’re probably aware that, in the last few days, I’ve made a few changes around here, such as adding the sidebar to individual posts, and changing the ads below each post in the individual post pages – replacing the old “medium rectangle” with 2 green-bordered small squares.

If you haven’t done so, look at them. Notice something wrong?

Yup, they’re all about blogging, or something related. Not like they used to be.

Even old posts (see the top 10 posts list, for instance) have blogging-related ads.

And they’ll probably be this way for a week or so. Why?

Meanwhile, there’s a different phenomenon that I had already noticed, for instance, in my mini-sites. Once, I copied the “external links” page (which already included an AdSense ad) from one site (the stem bolts one) to another (the recipes one), and, naturally, loaded it in the browser (at its new location, using the new CSS file, but still with the old content).

About a week has passed, and that particular page still shows ads about bolts and stuff. Not recipes. Not ads in Portuguese.

Why?

Because AdSense is very quick when reading the content for the first time an URL shows an ad. But, then, apparently, it only updates it from time to time. I’m guessing it’s two weeks or more.

And something similar happens when you change ad formats – for some reason, in my particular case, it “guessed” at some time that it was to show blogging ads, and hasn’t updated yet – even though many of the pages are months old, and had been showing appropriate ads (of other formats) for months.

Less than 3 months of blogging… In the Google Sandbox (which means virtually no visits from Google searches)… and I still have a day job.

Not a fortune, but I’ve just been able to pay the recent repairs of my car with my blogs. Nice.

Today, car repairs. Tomorrow, the world!

EDIT: shameless self-promotion: why not read the Blogging Tips series, now with 37 parts and still growing?

Eric Giguere of An AdSense Blog: Make Easy Money with Google has been doing an AdSense case study from some time now, which I’ve found quite interesting. Basically, he’s been creating a small site about invisible fences for dogs, and each stage of the site is archived, so you can compare them.

For instance, the first stage, “content”, only has text and some AdSense ads; the second stage is SEO’d, the third adds images and the fourth moves all formatting to CSS, incidentally improving the looks a lot.

Along with this article, this case study has been one of my main inspirations for my new mini-sites.

I’ve just found an article at My Blogging Tips called “How To Make Money Online Using Niche Content Websites“.

The author’s suggestion is not related to blogs at all – instead, he suggests discovering a niche that has few or no websites about (as a Groo fan, I’d suggest cheese dip, myself ), then making a simple but useful (and static) site about it, with good SEO techniches, promoting it, then… forgetting about it!

He suggests that, if the niche is well chosen (something people want to know about, but which there are few sites about), after 3-6 months (allowing for getting out of any sandboxes and everything) the site can steadily earn you $1-$2 a day. Now, as that site was probably put together in a week or less (but it’s still supposed to be better than the competition, which, the author suggests, probably consists of just a couple of unprofessional sites made with FrontPage or Word… yuck!), and needs no further attention, you can create another next week. And another, the following week. And another.

50 web sites (doable in a year) would give $50-$100 a day. Without requiring any further work from you. Not bad, huh?

And, as you do them, you’ll probably get better at it, be able to do another site more easily and more quickly, have increased knowledge about which SEO techniques work better, and so on.

The only bad point: you have to search for those niches all the time, and then quickly learn enough about each one to make a decent, informative, useful website – and they’re probably subjects that don’t particularly interest you.

As I said, these are not blogs. You’ll get virtually all earnings from first time visitors, as the sites won’t be updated (unless you get “attached” to a particular one, whose subject actually interests you). Besides, return visitors tend to ignore ads a lot more than first-timers, whether it’s a blog or a static site.

I may try something like this in the future…

Read the article, it goes into a lot more detail.

Eric Giguere has a great tip at his blog: Plumb AdWords for keywords. He shows a quick and easy way for publishers to use AdWords (which is for advertisers) to find high-paying keywords to use in your blog.

Looks like a big coincidence, doesn’t it? After I mentioned my drastic drop in hits from Google in my gaming blog, Darren from ProBlogger makes a brilliant post called “What to do when your Google traffic disappears.

Go there. Read it. Now. But come back here, afterwards.

One of the points Darren makes is very important – in an already successful blog or site, such a drop in Google search rankings (whether caused by the Google Sandbox or something else) can mean a drop in your earnings of 2/3 or more! What if your blog has been slowly but surely growing, and you’re starting to see some money, which causes you to see the “light at the end of the tunnel”… and suddenly, earnings drop down to a trickle, almost nobody comes from Google… it’s enough to make many people quit blogging, or at least hoping to one day make a living from it.

But, as Darren says, all of this – again, whether it’s the Sandbox or not – is temporary. It may take a couple of months, but you’ll get back to good positions in Google, eventually. Meanwhile, keep blogging, keep improving your blog, and get ready for the big comeback, when suddenly you’re flooded with hits. And clicks.

I’ve just finished making some changes to my blogs, including this one. See if you can spot them (assuming you’ve been there before, of course ).

Can’t spot them? Here they are:

• removed the borders around the entire pages, the individual posts, the sidebar, etc.. – I think that just that minor change has made my blogs look much more “fluid”, less “rigid”.
• reduced the importance of the ad in the header, which, anyway, according to Google, isn’t a good spot anyway. In some blogs, I removed it altogether.
• in a couple of blogs, I removed the ads in the sidebar – again, they’re not usually very successful. Left it in Garrafive de 5 Litros, my brother and cousin’s humor blog, as a big vertical ad like that one can work, as long as it doesn’t make the sidebar too big.
• began using my real name in some of the blogs, instead of “Dehumanizer”. People tend to be more talkative, and comment more, if they can call the blogger by name.
• finally, I added some ads between posts in the main page. These don’t feel as much “not part of the content” as those at the top, or the sidebar.

I really think they look better now, and, as for the ad results, the next days should prove interesting…

From ProBlogger, another great post: Making money from blogging takes time.

It’s something I’ve also mentioned before: blogging for money is not a “get rich quick” scheme. It’s hard work, and takes dedication, talent, a bit of luck… and, above all, time. In this particular case, I don’t mean “time” as in a lot of hours a day, but in the sense of “been doing it regularly for years”. Because:

1. it takes time to build a readership of regular readers, who either visit your blog(s) every day, or subscribe to the feed(s) – but still come to your blog when they want to comment, and they will do it often, because they’re used to your blog, they feel comfortable there, and know (even if never face to face) both you and most of the other regulars there.
2. it’s also a matter of numbers – number of posts, that is. If you’ve started your blog a week ago, it has probaly 5-20 posts, but if you’ve been blogging for 2 years, 1500 to 2000 posts are a more likely number. That gives much larger odds of someone googling for something and arriving at your blog – even at an old post. Which impresses the visitor, of course. And, being impressed, he or she then looks around, reads your more recent stuff… and you have a new reader!

(NOTE: this is part of the “Blogging tips” series)

At last, the long awaited (I hope) part 16!

Google AdSense offers many ad formats: big, small, horizontal, vertical, with or without images… so, which to choose for your blog?

The answer, of course, depends on several factors. And there probably isn’t a single one: it’s likely that you’ll be better served by using different ad formats in different places.

Also, if you have other sources of money on your blog, such as affilliate ads, it may affect the number and position of AdSense ads. But affiliate programs are a subject for a later part.

A few tips, then:

1. usually, you want the appropriate ad for a particular section (a vertical ad for a vertical section such as a sidebar, and a horizontal one for the rest).
2. the ad should be as big and noticeable as possible, without, however, drawing attention away from the content. Content is king. You don’t want your blog to look like an “ads blog”. People see those and close the window (or, if they’re intelligent, close the tab ).
3. a large, horizontal ad at the top of the page, either with the page’s color scheme, or with a different one, which fits nicely, tends to work well.
4. vertical ads on the sidebar aren’t as good, but can complement the top ones nicely.
5. ads at the bottom are usually ignored. In general, don’t bother.
6. avoid “Link Unit” ads (go to https://www.google.com/adsense/adformats, they’re the ones at the bottom). My experience with them has been dismal: 0 (zero) clicks in two weeks. It’s possible that in other kinds of blogs, or using them in a different way, they work… If so, I’d like to know how.
7. on individual article pages, an ad after the text can work well.
8. I haven’t used these yet, but I intend to: ads between articles in the main page (but only a few – no more than 2 in the whole page). Some people seem to use them quite successfully. These ones will probably work better either using the site’s color scheme, or a very similar one – they should be different enough so that people notice them, but not so different that people just skip them without even consciously looking.