Over Optimization Filter and Age

[GreenDog18]

Background: I have a fairly new website (2 months old) with around 84 articles. 40 of which are in-depth high quality product reviews.

Problem: A lot of my reviews are 1000+ words and I use <h2><h6> and <STRONG> tags to separate the different portions of the article. For example I have a section where I track Widget Recalls so that is in an H2 tag.

This is where it gets good:

Last weekend I was wondering why these post were not ranking very well so I decided to take one of my articles and strip it of any <strong> or <heading> tag. I basically made them all red text (which looked horrible.) so they would stand out.

Logged into Webmaster Tools and Fetched the URL of the page I changed and BOOM my rankings increased by 30-40 spots for my keywords within 5 mins. I was able to replicate this on over 50 pages.

Bigger Problem: I couldn't stand the way the page looked. It looked like amateur hour on my website. It was harder to distinguish the different parts of the article.

Once I put the heading tags back in my rankings dropped again.

Now my real question: If I look at the #1 website in my Niche they have TONS of <strong> <heading> tags. For example here is what he has for one page:

<h1> - 1
<h2> - 4
<h4> - 6
<strong> - 39 (not keyword based)
Affiliate links - 1
Dupicate Content - He has a 218 word block of content that appears in most of his articles. (Describing the way he reviews widgets)
Incoming links: This guy has and gets TONS of links from high quality sources.
Domain age: Around 4-5 years old.

This website ranks for TONS of keywords and roughly brings in about 50k visitors a day.

Final Thoughts: So what I was thinking is that there has to be some kind of domain age/authority threshold that if you pass you can have nice things like headers and strong tags.

Should I continue ranking poorly with a nice looking website or should I have a crap looking website but atleast have visitors.

[Andem]

Very interesting observations, GreenDog18. Thanks for reporting this.

Have you done more testing than simply dumbing down your page structure and changing the text to red? How about just removing certain header tags or just using one header tag for the article title?

I would highly suggest forgetting the red and focusing more on the readability (ie. Black) of the text so that people are likely to stick around and read the text. Since your site is new, you don't want high bounce rates to plague it which may hamper any growth.

I believe Google's algorithms for ranking pages take age and markup into account but in a much more complex way than any of us could really understand enough to provide personalized advice on. Actually what you're doing seems to be the best way to go about things.

To answer your questions: The data from any A/B testing you do is infinitely more valuable to you than any answers on this forum provide and no, I don't think you should continue running a really ugly site unless users are really sticking around.

[JD_Toims]

Have to agree the removal of the document structure increasing rankings is interesting.

Dupicate Content - He has a 218 word block of content that appears in most of his articles. (Describing the way he reviews widgets)

Contrary to popular belief, there is no such thing as a duplicate content penalty for the most part and the block of text will likely be detected and discounted as "boilerplate" content -- Meaning it will basically be ignored as if it wasn't there.

---

Links to the duplicate content detection patents and boilerplate/stub-pages threads are available here: [webmasterworld.com...]

[not2easy]

You could always use css to style the appearance the way you want it to look and make the page organization clear.

[bumpski]

Have to agree the removal of the document structure increasing rankings is interesting.

Another way to look at it is you have now created one long contiguous block of text, a single uninterrupted paragraph. It seems sites that have moved away from the paragraph tag have survived the onslaught ...

You could always use css to style the appearance the way you want it to look and make the page organization clear.

This is a scary thought that could be true, Google has gone so far as to penalize header tags but not span tags?

There does appear to be many sites that get a "free pass" or "bye" on Panda like and subsequent penalties. It would be nice if Google just moved back to ranking pages, not sites and pages. There are many stories of pages relocated to other domains popping right back up in the rankings. Small sites with diverse topics, and limited specialties, seem to be hammered.

[roshaoar]

I'd be interested to know what you have IN your h1-6 tags - maybe it's not the absence of just "the tags" themselves per se, but the absence of whatever is "inside" the tags as h1-6 tagged.

Fwiw, I have a 5 month old site that's now the niche leader, I use H1 & H2, some pages H3, and all pages H4 (just for "related articles"). Barely use bold though, 1 per page on the breadcrumb but that's about it...

Yes I could easily do it in CSS and I'll try that on some random article to see if it makes a difference, but I'd rather have H1-4 because well it's 'proper'...

[rish3]

Fetched the URL of the page I changed and BOOM my rankings increased by 30-40 spots for my keywords within 5 mins

Anyone else ever seen anything like this? I'm not trying to call you out, but I'm highly skeptical of the 5 minute turnaround. You sure you weren't just seeing some personalization effect of having visited the page over and over?

Because if that's the case, Google would be incredibly easy to game with a 5 minute cycle time A/B test.

[GreenDog18]

@rish3 I think my site was suffering from a OOP Penalty so once I used Google Fetch and Google re-crawled the page and noticed it wasn't optimized at all it lifted the filter for that page.

This week I will see if I can't replicate it on video.

[bumpski]

Anyone else ever seen anything like this?

Using "fetch as Googlebot" and submit, you can certainly get five minute turn around on Title changes presented in the SERPs. Very recently I tweaked quite a few titles to reach Google's new slightly sorter title length requirement; so a full title would show instead of a Google hacked two word title. These SERP result changes showed up in minutes.
I didn't pay attention to ranking, but Google pretty much seems to be penalizing anything that appears to be tweaking. Here again one is attempting to improve the visitors experience, but Google, with its SED programs (Search Engine Degradation) is harming the visitors experience both in the SERPs and by hacking titles, basically discouraging visitors from visiting legitimate sites.

[Martin Ice Web]

I saw that Google uses H1 tags for rewriting my title in serps. Maybe to many h-tags confuses the bot? Especially if you have more than one Topic on your page.

[aakk9999]

Once I put the heading tags back in my rankings dropped again.

How long did the page stayed on in the horrible new format before you put back headings/strong? What about changing one page back to horrible red and leaving it for 2-3 months to see if rankings hold?

It could be that by removing headings/strong, the page was sufficiently changed that resulted in some signals being reset so the page had to be re-evaluated again. It is also possible that some signals are folded in later on with the time (e.g. usage? maybe even inbound links?) which impact ranking later on.

Here is why I am saying this:

Few years back I had a page that Google kept getting the language wrong. This was a multilingual site, hosted in a country in Europe (not UK/Ireland), but with the home page in English. Suddenly out of blue Google declared the page language to be of the country it was hosted in and home page got "Translate this page" in SERPs. Ranking dropped a few places and CTR dived.

After a while I realised that every time I changed something more significant on the page (e.g. page title or more significant things in page structure), Google "fixed" the language and recognised it to be in English.... for the next 8 weeks. Then suddenly it said it was in foreign language again. I would change the title by swapping 2 words around, do a fetch and "Translate this page" would disappear... again for the next 6-8 weeks. So such changes were my work-around.

This had been going on for a year and was incredibly frustrating. I discussed this with Tedster whose theory was that when something significant on the page changes, it needs to be evaluated again. But this evaluation cannot use signals that would build over the time.. until some time passed.

I am wondering whether the similar thing could be here - that by removing headings and strong you have so much changed page structure that initially ranks well but could drop later on if the ranking is influenced by signals that build over time.

[White_Dove]

I have established pages with h3 tags where the h3 content ranks very wells, it is hardly over optimized; Rather the material specific to the h3 tag has only a paragraph or two related to the h3 topic and the page title is relates to a larger theme and does not contain the words used in the h3 tags.

All long tail stuff IE
title:widget
h1:Page headline similar to title.
h3:blue
h3:red
h3:big
h3:small

Search looking for blue widget shows widget in the title and blue in the description.

The h3 content changes on a regular bases. Again no effort is made to optimize other page materials around the h3 content.