profitlabinc.com

Line-making and Line-tending in Light of Trustrank

The classification systems we use everyday and all over the world are based almost entirely on arbitrary and subjective diferences between things. That two people can disagree on whether Object A belongs in Category 1 or Category 2 begins to illuminate an important problem faced by the Google Spider, as it is today. If two persons, with all five senses in good working order, can disagree on fundamental classification, how can the Spider, which is blind, serve them both?

To remember Image Search in its begnnings, you'll recall SERPS filled with very generic looking, you might even say idealized, images of things. That was when the Spider was young and wholly literal, operating on the basis of simple text recognition for a pre-programmed list of words. If it encountered a file named apple.jpg, it assumed "apple" -- which is exactly what its programmers wanted it to do. Programmers, who were also responsible for the initial "seed set" of images, chose ideal representations of basic objects, like the perfect red apple, a tree set by itself on a grassy hill, a pen on a desk, etc. Most people would probably do as they did, if given the job of "teaching the meaning of things" to a fledgling intelligence.

Partly because that was the way it was taught, and partly because the idea of "catalog browsing" dominated the early Web user's expectation -- the Web was like a great catalog, and all the images seemed as if from a catalog -- one-dimensional when describing "common", archetypal objects like an egg or a cloud, for which one universal representation would suffice for most people. Or else, a "record of detail", like a particular machine part or something of the sort.

The people responsible for naming files, especially in large quantities, realized it was impossible to be sufficiently descriptive of the image, in just it's file name, and it would take too long anyway. So they switched to automatically generated alphanumeric strings. It's the reason we all have Social Security numbers, because our names are not enough to sufficiently distinguish us.

So, the images in the Index began to polarize into two distinct camps. Big companies tended to have lots of well organized, stably-hosted files with no natural language in their file names. Small companies and individuals, who could take the care to name files thoughtfully, had tons of "yellow_car.jpg"s and "BBQ.gif's" that were here today, gone tomorrow. Of course, big companies had many files with natural language names in them too, but these tended to be permanent interface graphics or the small number of images that could be placed deliberately by a webmaster or editor.

Then a funny thing happened. SEO spammers began to label files with misleading file names and alt tags, in an effort to secure their placement on a greater than merited number of SERPS. And the Spider recoiled with distrust from images given natural language names. But, since the guy in charge of Acme Inc.'s library of 10,000 widget images probably had no interest in pursuing image rank, the Spider's capacity for very specific identification via long-string file name went un-utilized by the spider, and un-exploited by spammers. Everyone was so busy trying to get on page one for "big tits", that no one noticed the opportunity.

Google will always continue to bring images into the index from little sites and bloggers. We know that they are committed to the "little guy" and to preserving the democratic model that brings the best content to the surface. But companies, "businesses", are going to have to adhere to different rules. They are:

To surpass standards of organizational efficiency and "stability" as Google are probably already well on their way to defining for us.

To use file names and other object identification systems as reflect a large, stable organization -- namely, random strings.

To never violate Division Trust. If those are really images of pineapples in that folder then they should never be hotlinked to from certain kinds of sites.

But the biggest change is to totally erase your current conception of how image assets are used to acquire traffic. They don't have to be linked to, to be present in the index, nor do they have to be involved in a high number of links in order to be important in the index. When you have that one perfect image you're determined to see on page one, you are likely competing with tens of thousands of other webmasters with the same idea. What you want is instead to cultivate an ENORMOUS number of images and to pursue having as many of them indexed, as specifically as possible. In this sense, it is better to have 1,000 images on 1,000 different page thirties, than one image on page one. Which seems more stable to you?

Instead of making your image SEO strategy "outward facing", like somebody trying to signal a high-flying plane with a postage stamp. focus internally, on using your own sites, your own links, to show the Spider what and where things are.

Cultivate meaning for your images on your own sites, where presumably they are best described. I've done a lot just going in and taking advantage of sloppy work done by others, who may have good image assets, but no sense how to capitalize on them.

If you have a special image within a non-ordered set, take special care to describe it within a trusted page. You'd be surprised what you can do with an image (not linking to anything) with a nice, long, descriptive alt tag. Why write just "big tits" in the the alt tag, when you can write "Sandra's gargantuan breasts hung pendulously, daring Julius, Tommy, and Alexander to come fuck the shit out of her at will" -- hehe.

This new relationship between traffic acquisition and image assets may suggest a number of lucrative business models as well. For example, instead of selling hard links which likely violate Trustrank, think about renting trusted image use. I see a nice business for the company that has kajillions of images, thourougly indexed, and thouroughly described and provides a search interface to be used by webmasters creating new sites and new pages. I'd love to discuss the mechanics of that, if anyone is interested.

Which bring us to the relationship between affiliates and programs as regards images. Send your traffic to programs that give you benefits -- a good program should give its affiliates the benefits of their image assests (as PR booster and just plain resource) -- hence the importance of hotlinking. When you encourage affiliates to download and host, they are on their own, taking your images and struggling to legitimize them in the Spider's eyes. Fuck that.

HERE IS SOME ADDITIONAL GOOD INFORMATION TO STUDY CAREFULLY

“There’ll be no Fockers in my circle of trust”
Bob Deniro - Meet the Parents.

“Trust is earned not given”

Time + Trust = High Quality Indications for relevance
Why there is no sandbox there is only trustbox

1. Trust is earned not given
2. It’s easy to break trust
3. It’s difficult to regain trust once it’s broken

The Sandbox
is:

* a myth perpetuated by the uninformed
* an Adwords keyword tool
* over dependence on simplistic variables
* baggage from failed SEO’s
* misrepresented; and should be put to sleep

The Trustbox
Speculation on quality relevant site indicators:

* A high likelihood for being quality from quality references - multiple quality references
* Uses sampled user data to verify integrity
* Has checks and balances for quality control threshholds
* Sustained, consistent, natural growth
* Language similarity, reading comprehension and grammar checks to verify integrity
* Improves over time - broken trust is tough to re-earn but takes more to break trust over time
* Being noisy doesn’t make you trustworthy
* Trust is natural and not synthetized - it’s fuzzy and ambigiuous

How do you break into the trust box (out of the sandbox)?

You don’t. You earn a position in it. You’re owed nothing by the trust box until you’ve earned your trust. If you violate the trustbox you are owed nothing once again. You can choose 10 ways to get rich quick or you can earn trust through paced practice over time.

Link development is not dead…you just need ninjas, not monkeys and father time on your side. Market friction creates balance and stabilized quality.

Relative trust
Trust is relative to:

* the industry
* the individual
* the information
* the intent

Think of the people in life you trust most and why. Now apply those concepts to websites...more
The Jagger update seems to have taken the aging factor and combined the TrustRank factor into one, forming a new age for Google.

In other words, sites have to reach a certain age AND acquire relevant links from authoritative sources. Further those links must also be aged before they are attributed to your site.

As you can see, Jagger is quite the update – forcing not only quantity but quality. You need to have both a sufficient number of quality aged content pages as well as a sufficient number of properly aged relevant links.

Jagger is doing what Google has been striving for ever since the Florida update 1 year ago. That is to make the index more relevant than ever.

By forcing sites to endure an aging delay (also called the “Sandbox”) it is attempting to ensure that a site is indeed worthy of joining its regular index.

Also, it is assuming that those sites which are related will want to link to each other without reciprocating links. In other words, you will want to link to another site because it offers more relevant information to your site visitors, and not because it will help artificially boost your rankings.

Further the update also ensures your site is worthy because it assumes that only those that link to you will do so because it’s worthy...more

"Meta-search has this problem of doing its own thing with results and taking control away from the user. TrustRank and Trust-o-Meter give the control back to the user on the search engine preferences. Its quite simple: Your trust level decides the ranking of the common results amongst the two engines. If you trust Google more , Google results get pushed up more than Yahoo's. So you get to keep your beliefs, while you still combine results at the same time...more

How is Combined Trustrank Computed
TrustRank = C*( GoogleTrust / GoogleRank + YahooTrust / YahooRank)

Jack Mardack