Keywords Meta Tag Research: Significant Effects Found

[EGM]

A lot has been written about the relevance (or lack thereof) of the traditional keywords meta tag. However, there is a
potential revolution afoot with regard to the tried and true keywords meta tag. In order to examine this, we need to
research the complicated interaction of keyword meta tag content, page use or position, and emphasis. Our research
indicates that web pages that carefully craft their keyword meta tag and the corresponding page content can see some
large improvements in their page rankings in the Google index.

For the purposes of this research, we used 41 different web pages from 6 sites. We collected preliminary baseline data
on the pre-test ranking of the primary key phrases for each page. We then examined the effects of some keyword changes
over a period of several months. The factors we examined are discussed in detail below. The effects were compelling,
largely due to a combination of related factors involving both the meta tag itself and a variety of on-page elements.
Statistical analysis of the results showed a significant interaction effect across a number of factors.

Here are the variables we examined:

1) Order of keywords within the meta tag
2) Presence of keywords both in the meta tag and in on-page elements visible to the user
3) Emphasis of keyword use in on-page elements visible to the user

The third variable deserves special consideration. We defined emphasis in several ways. We looked at three specific
instances:

a) Presence of a keyword in the title tag
b) Presence of a keyword in an H1 tag
c) Bolding of a keyword

Mid-way through the project, we added one other variable to the mix which was the simple presence or absence of the
keywords meta tag altogether.

Some of the results will not be surprising for most regular readers of this forum. I present them just for
thoroughness.

1) Presence of a keywords meta tag by itself doesn't help, but its absence hurts. In other words, you need to have
the meta tag, but just throwing any old words into it doesn't help on its own.

2) All words or phrases used in the meta tag need to be present in on-page elements, visible to the user. Use of
plurals and other stemming is included in this finding. Having a singular version of a keyword on the page visible
to the user and having the plural version of the same keyword in the meta tag but not contained in the visible
page elements has a slightly negative effect. This was a surprising finding.

3) The order of words in the keywords meta tag matters. A statistically significant positive effect is present
when the order of the keywords in the meta tag matches the weighting of the combination of emphasis elements defined
above. This is the potentially explosive finding of this research project and is discussed below.

The interaction effect of keyword order in the meta tag and relative emphasis in on-page elements is complex but
could be worth the time for content designers to figure out, especially in highly competitive fields where ranking
for particular keywords and phrases is critical.

We assigned weights to each emphasis factor to come up with an overall emphasis score for each keyword or phrase on
each individual page. Here are our weightings:

a) In title tag = 3 points
b) In H1 tag = 2 points
c) Bolded = 1 point

We assigned points based on only one instance of each factor resulting in a maximum emphasis score of 6 for any
individual keyword. Multiple instances of a keyword in the title, multiple uses of a keyword within an H1 tag and
multiple instances of bolding the same keyword did not demonstrate a statistically significant effect.

Similarly, none of these factors was significant on its own. Only the interaction of all these emphasis elements
combined with the word order in the keywords meta tag produced significant results.

To code the effect of keyword order in the meta tag we assigned values to each position. Our research showed no
effect beyond the third keyword in the meta tag. As a result, we assigned the following values:

a) First position = 3 points
b) Second position = 2 points
c) Third position = 1 point

The following interactions were observed:

1) Keyword position by title tag by H1 interaction: Significant at the .04 level
2) Keyword position by title tag by H1 Bolding interaction: significant at the .01 level
3) Keyword position by title tag by Bolding interaction: significant at the .06 level

The results on SERP ranking were very clear. Pages that employed the combination of keyword meta tag position along
all three emphasis elements improved their rankings by an average of 4.1 positions with a range of 1 to 27. No
negative effects were observed.

Although the basic keywords meta tag has been much maligned over recent years, it appears that there is still some
value for the strategic use of this basic search engine optimization element in conjunction with other on-page
factors. The preliminary results reported here are only based on Google rankings. An analysis of the results using
the other major search engines is in progress.

We will be writing up a full report with more detail on our methodology, analyses and conclusions very soon. But
we felt like this was just too exciting to keep to ourselves any longer.

Our next project looks at a complex interaction of the first letters of each sentence in the first two paragraphs
of any content on a web page to examine this effect on search engine ranking. We believe that this research will
be very illuminating.

Thanks for reading!

EGM

[Stormy S]

Wow, that's really interesting... thanks for sharing the details of your research!

A question though. Given that you conducted this research over a number of months, how were you able to make sure that you were measuring the effects of the changes that you made to your pages, and not the changes/tweaks that Google might have been doing over that time?

Google seems to be changing stuff all the time and recently with Big Daddy hanging around it seems to be more volatile than ever. Were you, like, able to change some pages in a certain sequence, and change other pages in the reverse sequence, to prove that your measurements were independent of Google's fluctuations?

This is an April Fool's joke, isn't it?

[EGM]

Read the paragraph at the end about our next project and then apply this to the first two paragraphs of the original post.

[Metis Online]

hehe sweet!.. good and subtile one

[KylePosey]

I can't believe I was duped...

[simons1321]

thank you for wasting exactly 2.7 minutes of my time

ps. i hate you

lol

[nimbus]

I have had various conflicting information with regard to meta tags, I understand how they work but keep being told the key words must be part of the page text for the key words to be picked up. I personally think this is not the case and that meta tags work behind the site to ensure your postion in the search and aid searching.

Can somebody confirm this please

[jimbeetle]

Welcome, nimbus.

This recent thread might be of help. But basically no, the keywords appearing in the meta need not be on the page itself. I believe that goes back to pre-2000 when Alta Vista was battling keyword stuffing in the meta.

<added>

I just noticed that this is in Google Web Search. As far as we know Google does not look at the meta keywords. As far as we know Yahoo is the only engine that looks at them.

</added>

[nimbus]

Thank you for your response. I am still not sure of the definitive answer.

If I do not include the keywords in the text on all the pages will this affect the search or do the meta tags do all the work of linking the search.

Thanks

[jimbeetle]

Did you read the article referenced in the other thread? It pretty definitively spells out how Yahoo looks at the keyword meta. (And again, as far as I know, only Yahoo looks at it.)