I do a lot of niche research these days. One of my criteria has been a competitive and reachable first page of SERPs on Google. Those that look unattainable end up getting ruled out. While doing this research I noticed something very interesting…

We’ve all been taught to believe in the power of incoming links. Links pass PR which boosts our ranking, right? But on a recent search across the web I noticed a lot of content getting ranked in the top 10 on highly competitive terms, but yet the page has no pagerank value associated to it. On closer inspection a lot of the text seems to come directly from a lexicological search, like the one you’d get if you looked for related search terms with Google’s Adwords tool. In one case someone mirrored a wikipedia page and ended up only 2 notches below the original page, yet had no links on a link: search and no pagerank at all.

This is leading me to believe some things are changing. When I searched for the URL in question I found numerous web bookmarking services like digg and technorati had the site listed, but all using rel=nofollow links. Could we be seeing some sort of alternate trustrank in the works? Maybe these nofollow tags are being followed afterall.

Or maybe it’s purely on-page optimization. I was always told that on-page text is too easily manipulated, but maybe Google’s algo has gotten a lot better at detecting abnormal text patterns and they’re beginning to shift from the old link vote strategy to a more on-page LSI-based.

This makes me wonder how long it’ll be before SEO folk will be to beat the new system and get high ranks without a single drop of PR or link building. I wonder what that’ll do to all of the link directories out there… or whether they’ll continue to serve a purpose in the new regime. It’s hard to tell, as these things have just begun, but one thing’s for certain… times are changing.

Seems that every compay wishes they could increase their customer loyalty. I hear it all the time… but what about a company’s loyalty to their customers?

Last month I recognized several bad charges on my credit card. Turns out the number had gotten leaked and now was unusable. So naturally I reported the number and got a new one issued. Well… sometimes it just takes forever.

So I’ve been using this service for the last 3 years or so. It’s a very small monthly fee, under $10, never a problem. Well I get an email today saying that my card was declined (big shocker) and that if I don’t reply immediately they’ll freeze the account within 24 hours and start charging a fee.

Now I understand they want their money, but I’ve been a paying customer of theirs for YEARS. They’ve never had a problem with me, and there’s a legitimate reason why the card was declined. Give me a week and the mess will be straightened out… it’s bad timing all around.

Of course I’m quite sure this message was automated. Which brings me to a natural question… Why in the world would ANYONE put hostile, threatening language, in an early run automatic response message? It’s almost enough to make me start shopping around.

I think there’s a lesson to be learned here for businesses with good long term customers. Treat them with a modicum of respect. Be nice to them, maybe even bend a rule here or there. And never start threatening them, anything you have to say can be said nicely.

“Help! I’ve changed something in my domain and now it’s all funky! How long is it going to take before it all works again??!?”

I just spent the last week slowly moving over several domains to a new server. I’m worn out by it all… but now it’s done. Here’s some things I’ve learned from the experience. Make use of them.

When you register a domain name you take control of it for a small fee. But just because you have the domain registered doesn’t mean it’s capable of doing anything. In order for the domain to be useful it has to point somewhere… it has to point to a hosting service of some kind.

Every domain has something called a “domain zone file.” A zone file is where all of these pointers are stored. In this zone you can specify that your domain will point to a particular server, or that a particular subdomain will point to a particular IP and your main domain will point to another IP. You can specify that a domain points to another, or that a domain is nothing more than an alternate name for another. You can add text records like SPF, you can do a lot of stuff.

Zone files are hosted on “name servers.” When you register a domain name you need to provide name servers, or use the registrar’s. That’s where your zone file will be kept.

A typical record set looks like:
skepticalmarketer.com. 14400 A 70.47.127.0
skepticalmarketer.com. 14400 MX 0
www 14400 CNAME skepticalmarketer.com.
ftp 14400 CNAME skepticalmarketer.com.

This says that skepticalmarketer.com points to 70.47.127.0 (A typical “A” record). It says that skepticalmarketer.com is the primary mail transfer server for incomming mail (so that other servers know which server to connect to) and that www and ftp.skepticalmarketer.com both point to the main domain’s IP.

The number, 14400 is a special number called the “TTL” number. TTL, or time to live, specifies the maximum duration (in seconds) that it should be allowed to stay in a cache. 14400 seconds = 240 minutes = 4 hours.

Additionally, the information about which name server is the “authority” is cached as well. Technically this is called the “authority record” or “SOA” (start of authority) resource. It contains the main name server that will be responsible for delivering this zone file to those that ask for it. It, too, has a TTL. Usually it’s a bit longer, 86400 is a typical number. 86400 seconds is 1 day.

What’s a cache?

How many times do you visit warriorforum during the course of a day? Once? Twice? 10 times? How many pages do you load during that visit? 5? 10? 20? Think about all of the images and all of the sub-files that load. Each of these are loaded by domain.

Now your local ISP is a lot closer to you, net-wise, than some all powerful centralized server. And that central server might not even be available due to internet traffic or other problems. If every time you loaded a page you had to make a dozen queries to some all powerful central server, that server would get overloaded and your connection would drop to a crawl. The people that designed the internet realized this, so they decided to decentralize the domain name system by using caches. Your ISP, and all ISPs, maintain a cache of domain names. The cache is checked and updated anytime someone makes a name server lookup.

It works like this:
Your lookup -> Your ISP’s domain name server -> Looks at the TLD (top level domain, ie: .com, .biz, .org) record -> Queries the name server listed in the TLD (the authority record) -> gets the information from that server and saves it.

This way whenever you want to make a domain request you only need to go as far as your local ISP. When the TTL is up, that entry is purged from the cache. If you visit that domain again your ISP will simply cache the info again.

Since windows 98, most operate systems have also kept a small local cache. So the chain goes even deeper. First your OS will ask itself “Do I know this already? before asking your ISP. This speeds up the process too and helps to keep you running smoothly.

Now that you understand why and what a cache is, you can begin to understand what happens when you change domain information.

So what would happen if you suddenly changed, say, the A record above:
skepticalmarketer.com. 14400 A 70.47.127.0

To something like:
skepticalmarketer.com. 14400 A 127.99.57.104

Well first the change would have to happen within the server. That usually only takes a few minutes, if that. Now any server that hasn’t already asked for the record, or that is working on an expired cache, will now receive the new information from the authorized zone the next time they ask for it. So there’d be no delay there. But on any system that has a cache… the old cache will continue to be used until it expires.

In this case the recommended expiration was 4 hours. Local cache servers are free to change this value if they want and some do have minimums or maximums in order to improve performance. Imagine if the cache expired every second, the traffic that’d generate could slow the entire system down quite a bit. Still, the recommended amount was 4 hours, and that’s probably what it’ll be on most servers. So in this situation the old A record cache should expire in just a matter of hours and the new one will be ready to go. Pretty sweet, eh?

Consider this with a semi-static IP. If you have an IP address that changes some you could point a subdomain to that semi-static IP, set a low TTL (like 3600 seconds, 1 hour) and merely update the zone file if things change. This is how “dynamic DNS” services work, except they combine the zone update server with a program that runs on your machine. You can emulate that behavior even without a program like that, tho, if you want.

But that’s just the beginning.

So what if you’re moving hosts? Or if your domain is registered at namecheap and you want to pass zone control off to your host and have it run the DNS service?

Well first, it has to get a completely new authority record. It gets this from the top level domain record, like .com, .biz, .org, .net, etc. Each of those have designated central registries that maintain these records. Those records must be updated (which can take time), individually cached SOA records also must expire, along with local caches.

As you’ve probably guessed, this takes quite a bit more time. This is why some DNS changes can happen almost instantly, while others can take hours or even days.

So how can I put this to good use?

If you’re moving servers, start by lowering the TTL ahead of time. A day or 2 if you can, this will make sure that cached entries don’t stick around very long. You can always change it back when you’re done, or if you’re changing name servers just lower the old zone’s TTLs and keep the new one at the default.

Next, consider whether or not you really need to change name servers. If you’ve registered your domains with a service that provides it’s own name servers you might never need to change them. This will tremendously speed up the change-over.

If you absolutely must change name servers then keep the old zone file around for a little while. You can edit the old zone file and point all of the entries to the new server without completely removing the old zone. This means that queries to the old name server will still resolve to the new records. Meanwhile the new zone will be coming into effect. As the old caches expire the new one will take over. After a few days you can then remove the old zone completely and still benefit from the lowered TTL.

Old zone, new zone? How can I setup a new server if it’s not resolving yet?!?!

The answer is… the HOSTS FILE. Remember how I said that all operating systems these days have their own local DNS cache? Well you can override this cache by using a “hosts file.”

The hosts file looks like:
70.127.88.122.123 www.skepticalmarketer.com
70.127.88.122.123 skepticalmarketer.com
70.127.88.122.123 ftp.skepticalmarketer.com

It’s just IP address and the name. Wildcards are not supported so if you’re working with a lot of subdomains you may need to setup an entry for each.

The location of the hosts file is different in each operating system. If you’re a linux user you probably already know where your’s is (/etc/hosts usually).

Other operating systems:
Windows XP/Vista = C:\WINDOWS\SYSTEM32\DRIVERS\ETC\hosts
Windows NT/2k = C:\WINNT\SYSTEM32\DRIVERS\ETC\hosts
Win 98/ME = C:\WINDOWS\hosts

You can put the new IP address in this file along with your domain name. You may need to refresh your network or reboot your system for this to take effect (depends on your OS), but once it does you’ll now be set to use the new system. This allows you to make fixes on the new server before it goes live to the world.

Well, that’s what I’ve learned. I hope it helps someone. Make use of it, it can save you quite a bit of headache.

(c) 2007, Aaron Colman

A story on Digg today got my attention. It was a discussion on how our insults have decayed from a clever retort to todays “pwnt!”

As I sat thru and read them I got quite a good laugh, and suddenly found that a little wit sure goes a long way.

Little insults like:
“Some cause happiness wherever they go; others, whenever they go.” - Oscar Wilde

“There’s nothing wrong with you that reincarnation won’t cure.” - Jack E. Leonard

“In order to avoid being called a flirt, she always yielded easily.” - Charles, Count Talleyrand

“They never open their mouths without subtracting from the sum of human knowledge.” - Thomas Brackett Reed

When writing your business letters, when writing your sales letters, when just writing at all, remember to add a little wit and humor. It lightens everyone’s day and will keep people reading thru the predictable boringness of modern communication.

Remember the immortal words of Robert Redford:
“He has the attention span of a lightning bolt.”

I just don’t think it’s about providing value anymore, but rather it’s about providing entertainment.

Is your business entertaining enough?

Had an interesting experience today. In the quest to debug one of my apps I was talking to someone with an interesting problem. His email address was getting double encoded.

Example:
user@domain.com

Normally becomes:
user%40domain.com

When it passes thru the browser.

So what happens when, for some strange unexplained reason, the info is double encoded?

The % in %40 becomes encoded too, you get user%2540domain.com

Which is exactly nothing but an invalid email address. The system kicks it out, and it’s unusable.

Now most people will never experience this problem, and perhaps only 5% of your userbase may experience it. But what if this guy, this new subscriber, was a potential big customer? What if you’re working with multi-thousand dollar items and every single person counts?

We could simply blame it on the software, but if you’ve been on the internet for more than a week you’ll notice… there are more potential problems than anyone can ever solve, ever. No matter how great the app is, some stuff will always slip thru the cracks.

It gets worse, and perhaps more ironic. I get a spam on my forums last night. Some web developer that should know better is posting ads on a forum of mine. It’s my general policy to track these guys back when I can, and in this case it was easy. They put their URL all over the place. So I went there and sent a message via their contact form.

I got this error:
Hi. This is the qmail-send program at xxxx.yyyy.net I’m afraid I wasn’t able to deliver your message to the following addresses. This is a permanent error; I’ve given up. Sorry it didn’t work out.

<user@domain.com>
Connected to xx.yy.zz.ww but sender was rejected.
Remote host said: 554 refused mailfrom because of SPF policy

Now let me translate this from “Geek.” You see SPF policy is designed to verify that the sender of a message is really the sender. It looks at the chain of servers a message passes thru and compares it to a record specified by the domain, and if they don’t match… it’s trash’n time. In otherwords SPF helps to make sure that 3rd parties have a hard time sending email pretending to be you (spammers don’t use real “From” addresses anymore, they use another name on the list).

The problem here is simple. I’m using their contact form to send mail, and it’s sending it as me. Their mailer notices that I haven’t given permission to send mail as me to their mailing server and denies the message.

These elements are hilarious. SPF = anti spam. Good right? A strict policy, good too, right? A contact form works great too since you don’t have to post your email addresses on the site. All of these, by themselves, work flawlessly. But when combined… it’s an exercise in hilarity.

Nobody with an SPF record can ever send messages thru their contact form, it’ll bounce and they won’t receive it. I could’ve been a big paying client with thousands of dollars to spend… but a poorly thought-thru config will surely cost them a lot of money.

So why am I going thru all this? Just to point out wierd technical problems?

Because in today’s world we’re used to letting the machines do the work for us. This sounds great, to be sure. But sometimes our machines fail us. They were told to do something wacky, or a set of normal commands mixed together becomes an evil frankenconfig monster. Ever seen 2001, a Space Odyssey? It’s HAL, just a few years late.

It’s unpredictable. And it will cost your business customers and money.

Unless…

The developer above would’ve been smart to test their config. They would’ve also been smart to have a usable phone number, and perhaps an alternate way of sending messages via the internet. This is why I’m a big proponent of support forums, if they can’t reach me thru a contact form they can post to the forums where I’ll still see it.

It’s all about a little preparation and having backups in place.

What’s your backup?

It’s been extremely laggy and slow around here lately. Just a note to let everyone know I’ll be moving servers in the next day or 2. So if I drop off the face of the net for a day, you’ll know why.

Wish me luck!

The power of viral videos is unquestionable.

But how can you use it to sell your product?

Well just look below. Notice the simple product placement. It’s the centerpiece of the entire video.

Just makes ya think “dang that blender is cool!” On their website you can even buy one in a nicely placed “order now” link.

See the video:

Is the web a generative media?

Can the world wide web create a completely new brand presence or does it merely reinforce previous ones?

The vast majority of sites are found by search engines. When people search on a term, they already know that term. They may be looking for a topic, or they may be looking for a brand or term they’ve seen before. But either way, they’ve already encountered the situation or your product before and they aren’t new.

So as a new name, how can you get established and break thru this barrier? How can you create generative buzz that gets your word out to new customers? How can you create a market for a product that has never existed before?

Well here’s 3 sure-fired ways:

1. Ezine placement - Perhaps the easiest way is to get seen by people checking their email. Find people that are likely to match your target consumer and buy space there.

2. Media coverage - Online or offline, media coverage is the #1 generative branding available. Coverage in general media, on the front page of say, cnn.com, is a sure-fired way to get serious attention to your new product.

3. Buy adspace - Avoid PPC, go for impressions. Get your name out and buy space where people will see your ad, even if you aren’t relevant to them. This is a financially risky approach, since in order to capture a new audience you have to reach outside of your established niche. But sometimes it’s the only way to get your name out.

So here we are. At the start of it all.