Improving Google's use of ODP data

I will make another attempt at starting a constructive discussion on here.. we'll see whether we are capable of doing this without snide comments....

There have been a few discussions recently about Google using the ODP description as its text in search results. Several people are quite upset about this, and I think they have every good reason to be.

Here is where things clash, and here is some ODP guideline stuff for you, so you can understand how things work. This is how the ODP has worked from day 1, as far as I can tell, so this isn't anything new.

Sites in the ODP are put into hierarchical categories, so a listing in the ODP is made up of the following attributes:

* URL
* Site Title (usually the name of the business)
* Category Name
* Description

Editors are told to be quite strict in the Title - no keywords, just the name of the business. The Description, which is obviously free-form text, should not repeat any of the information from the other fields - don't put the URL, the company name, or any part of the category name in there. Put a short description of the nature of the business, if needed, and then describe what is to be found on the site (i.e. photos of products, newsletter, F.A.Q. - that sort of thing)

So which of these parts do Google use?

+ URL
+ Site Title
+ Description

So, in my opinion, a vital piece of information is missing - the category.

Assuming that Google still keeps using the ODP data (that isn't what this thread is about - there are plenty on that topic), wouldn't it drastically improve both the Google results, and also the placings of sites if they were to somehow incorporate the category elements into their search results?

Let me give an example - a Funeral company in Manhattan, NYC. They would get listed in the category Regional: North_America : United_States : New_York : Localities : N : New_York_City : Manhattan : Business_and_Economy : Funeral_Services and the description would (quite properly) contain something like "Includes services offered, pre-arrangement consultation, pricing and contact information." - you don't NEED to put where they are, or that they offer funeral services, because that is already in the category name. Understandably, everyone would be ticked off with Google, seeing that listing out of the context of the category name. It doesn't say where it is, or what it does! If they were to just add the terms from the category, things would instantly be much, much better.

I have seen some very knowledgable posters about SEO and how these things work on this forum. So does anybody with more of a clue than me have an idea why they don't use this? I have asked the question of various ODP editors, but the standard answer is "we have no idea why other people do what they do with our data".

Am hoping for some serious, constructive and non-snide discussion on the topic.

 

Awesome information and insights Alucard - thanks a million for sharing.

It's just my novice opinion, but Google is like a college kid still experimenting and deciding what they want to be when the grow up. No doubt, they are extremely intelligent, or they wouldn't be at the top of their class, but they're still learning along the way.

One thing that would be helpful is that if they're going to use the "sanitized" ODP description of a site that they list the category in which it is contained - and since the Google Directory is an RDF anyway, this seems plausible:

 

We have had a couple of threads in here suggesting Google doing a little re-editing here and there.

 

Like the majority of editors I don't understand SEO either. If we can stay on topic, it will be enlightening to get the opinions of the experts on this.

Excellent topic, alucard. I look forward to learning.

 

I am waiting for google to ignore the directory for results.

 

Wonder if Google checks DP posts

 

Someone didn't understand my comment?

Let me put it this way: Read the title and then read this (I will add clarification LOL)

So I am saying: "why improve the use when they shouldn't use it at all?"

get it?

probably not..

 

Debunked, since the point that you bring up has been discussed in numerous threads on DP in the past, I was hoping this would be something a little different, which is why I wrote in the opening post:

 

OK, now I understand - sorry.... I actually didn't read the whole thing to start with cause I thought it was the same old ODP-google bit. Personally I have our sites listed in ODP, but think it is becoming useless for today's internet.

But back on topic:

I think google just uses the PR ranking of ODP pages to determine whether or not a page will show high, but that google uses it normal algo to find the search term on a page.

 

debunked, I didn't understand your comment either. Even with the clarification I still don't understand what it has to do with the topic, unless it's your way of requesting that we not discuss it?

I joined dmoz before I knew what google was. I don't understand seo. I don't understand all the anger toward a project that I deeply care about and have invested a lot of time in. I began visiting seo forums in an attempt to gain an understanding of the problems that web professionals insist are widespread.

I hear the complaints. I'd like to hear some serious discussion about the basis of those complaints, assuming that the "canned r-z response" to them is incorrect. To help us understand, we need to make a couple of assumptions.

First, alucards assumption: "Assuming that Google still keeps using the ODP data (that isn't what this thread is about - there are plenty on that topic)"

Second, we must assume the "canned r-z responses" to those who have complaints are incorrect.

There are a lot of intelligent people on this forum. With those two issues set aside, maybe we can have a productive discussion. It's happened before and I'd really like to understand this issue.

(added) oops, sorry. I was still writing as debunked posted.

 

LOL, I don't know if my head can think about this to the depth you guys are - I had a great weekend with about 70 guys from our church at a men's retreat. Between going to be late and 3 hour drives, I am too tired.

I didn't know you guys were in DMOZ, I haven't paid much attention to who is in the club or not. I am listed, but to answer part of your question, the listings are not the greatest for links. We can't specify a product name or type for the link, and then we are even told not to put KWP into the description. So as far as SEO goes, there is very little use except for PR and the authority that DMOZ is (or was, since I think it is losing its authority level)

 

assuming the sites around you in that cat are relevant to a similar topic to yours, simply being amongst them probably gives Google a better idea of what neighbourhood and standard your site is, than the category title would - anyone agree?

 

bradley, but if Google completely ignore the category, they have no concept of sites being "around' other sites. Using the category name would be a way of ensuring that mutual affinity. That was my idea, anyway.

 

Using an ODP description without the category path doesn't make any sense as far as I can tell. When we write descriptions we avoid using terms that are already in the category path to avoid redundancy. In other words, in the mind of an ODP editor, the category path is considered part of the site description. --this is one reason why it's so important to us that sites are listed in the proper category.

An ODP description without the category path is incomplete. For any data user to use incomplete ODP descriptions seems counter productive if the data user is trying to deliver good, relevant information. Editors put the information there, but Google chooses to ignore it ...I don't know why but it's not our fault.

The keywords that web owners seem to want so badly are already contained in the category path. That makes it fair to everyone in the category by giving them all the same keywords. Google leaves out those keywords, not ODP.

@links and related category links should further show relevance to a topic, ensuring mutual affinity. The ODP is a wonderful tool, but like any tool it has to be used properly or it can do more harm than good.

(This reminds me of an incident where my husband used a chainsaw while doing a small plumbing job. The chainsaw wasn't bad, the way he used it was bad.)

 

not necessarily - take, for example, http://dmoz.org/Computers/Internet/Web_Design_and_Development/Promotion/Chats_and_Forums/

- let's assume google 'knows' digitalpoint and other sites linked to by that page are about SEO (in part, at least)
- let's assume a site relatively unknown to google is added to that category. Without even knowing what category this is, google can look at the other links on the page, find a recurring theme, and may decide that there is a high probability that this new site is also about SEO, and place it higher up in SEO-related search results
->> therefore, the link to that site, when placed amongst those others and given equal importance on the page, suggests to google its in the same neighbourhood - without looking at the category title, google can look at the content of the page (links to known SEO sites) and presume that this other link is also related to the topic. Of course, a spidering will confirm/dispel that presumption, but i believe its a fair assumption to make.

Of course, that's all assumption on my behalf, with very little I can actually point at. But don't forget there are plenty of ways in which google will be picking up the theme (i.e. category) of the page, and passing that onto the links, even without looking at other links on the page - the category name, after all, is even in the URL and page title! I think it's fair to assume google does in fact realise the links on the page are indeed related to Manhattan funeral services, a lot more than you realise. What evidence is there that it doesn't (i realise I can't put my finger on evidence that it does, either!)

 

Anecdotal evidence, from this forum and others.

Along the lines of "Why does Google not return a value for 'London Real Estate' when my site is listed in the London Real Estate category" - implies to me that it's only searching using the words from the description, not the site name.

Stuff like that.

But I guess this is why I was asking on here - because most ODP editors can't even spell SEO, let alone explain it, and because there are some very knowledgable people on here that can take some pretty decent educated guesses as to how the big G works.

 

Basically since google seems to use the sites on the page for theming, but not necessarily the categories names for search results. If an ODP editor wanted a site to show up for a KWP it would work if you used the KWP as the link to the site. This is called the anchor text. (sorry if am simplifying to much for some)

ie. london real estate - find the home you want in london using greatlondonrealestate.com

you will see I used the KWP as the anchor text

 

The ODP says that the Title (which is used as the "anchor text") is the name of the business. That makes most sense for the listings.

I'm not saying that the ODP should start paying attention to keywords, here - I am saying that a potentially valuable source of keywords for Google is already contained in the RDF files, in the category name field, which Google appears to ignore.

This thread is about improving Google and how they use data which is already available to them.

 

WOW, I understand this! (doing the happy dance )

Problem is, our editing interface won't allow that type of link. There's no way for us to use anchor text with the link in the description field.

Google needs to use the complete data to make this work.


(added) I didn't mind be called a 'peon' but I don't like being called a 'grunt'

 

And the beauty of it is, the ODP makes that available - whether they are spidering the ODP pages (it's in the page URL) or using the RDF files (it's in the files with the listing)