Is My Site Relevant or Not

[trabis]

Hi,
I have a UGC website about widgets with a domain name similar to country-widgets.net

It was founded on 2006 and it grew linearly until 2012 when it peaked at 40000 visitors a day. After November 2012 it linearly fell and now it is getting about 15000 visitors a day. Never had problems with Google animals but the strangeness of it had brought me to sites such as this one, where I try to find an answer to the WHY question.

Today I just want to share a strange experience with Google search results hoping to get some help making sense of it.

My site, during it's peak months, was #1 result in SERPS for the query WIDGETS. Nowadays I keep bouncing between #6 and #12.
But...
When I use the date search filters I get:

Past hour: if there was a post made in the past hour, I usually rank #1, sometimes I even get top 5 spots on the page. The other results are mainly from news sites that are not competitors.

Past 24 hours: #1 or #2, sometimes I find a competitor in this page, the rest are just facebook pages or random blogs.

Past week: Sames as above, sometimes I'm out ranked by my own facebook page.

Past month: #1, no competitors on this page.

Past year: #1 to #5, sometimes I see competitors.

I've been seeing this behavior for about a year now and my site is the only one that is always there, that is always relevant (or isn't it). I don't understand why using ANY TIME, it is not relevant ANY MORE.

Thank you.

[rainborick]

It sounds like your pages benefit from the Query Deserves Freshness (QDF) effect, but once that passes, you're back to normal ranking levels - at least for the queries you've been testing.

I'd suggest looking for a correlation between the starting date of the drop-off in your rankings and the history of Google's algorithm updates to see if you can find the problem(s). I'd start with looking at how the Penguin updates work, since they involve issues that are common to UGC sites, and the dates might highlight particular aspects of how Penguin might affect your site.

[trabis]

Thanks rainborick for replying.

I'd suggest looking for a correlation between the starting date of the drop-off in your rankings and the history of Google's algorithm updates to see if you can find the problem(s)

I've mapped the updates (and every change I made) in my GA account and could not found a correlation with the drop-off. I never had a sudden drop-off! I.E. when I expected the site to rise 20% it would only rise 5% and when I expected it to drop 5% it would drop 20%. Imagine a hat or a snake that ate an elephant(The Little Prince), that is how my graph looks like.

I can argue/fantasize downward trend began after:
- Sudden spike in traffic due to some upcoming holiday
- Testing an Adwords campaign because I got a free trial invitation
- Temporary ban from Adsense due to images with nude content
- Facebook new world order
- MetaFilter contagion
- Spooky action at a distance

It sounds like your pages benefit from the Query Deserves Freshness (QDF) effect, but once that passes, you're back to normal ranking levels - at least for the queries you've been testing.

With the exception of "Last 24 hours" filter, it is my root url that is being listed, not the recent posts. Even when searching with "Any Time", it is my root url that is listed. BTW, I like to note that when my competitors are listed (using any of the filters), they get internal pages such as "category pages" or "top 10 pages". I'm the only one who gets the root page listed. Some how this looks like authority to me.

I'm not really obsessing over this particular query because most of the traffic comes from long tail queries. It just puzzles me this discrepancy of RELEVANCE when using/not time filters.

[rainborick]

Showing the main page when the query is more closely related to an interior page sounds like the interior page is seen as having much lower quality - which is not uncommon for Google and still smacks of the QDF effect. That is, Google would ordinarily serve the interior page due to QDF and general relevance, but that internal page has issues so they show the main page instead. It happens.

If you can't directly match the changes in your rankings to specific Google updates, it could just be your site fell below some quality threshold at that time and as Google continued to re-crawl your site it continued to see more of the same issue.

I'd still say the first step is to look for Penguin-related issues on your site.

[tangor]

Short query:

Is this Info page or ecommerce?

If latter, you're up against a boat load of competitors. If former, you best be the best of the bunch.

The fact that you are tracking G results THIS closely indicates this is a bread and butter site. I'd look at site SALES for the answer, not your G "time" etc. Sales holding up? If so, don't stress over the other stuff! (There's too many other things in the world we can stress about!)

[seoholic]

@trabis
I am in a similar situation and watching my trophy keyword rankings manually for years now. Slow death spiral, excellent ranking with search date filter, mostly long tail traffic, no impacts from updates, keyword in domain name, you name it.

I assume that freshness as a ranking factor decreased in importance over time since introduction. We are constantly adding some content, but it doesn't seem to be enough to influence the overall process, it just keeps us alive.

My traffic decrease can't really be compared to yours but also started in 2012 and in my opinion due to a lack of acceleration(!) in link acquisition and adding content because we are working on a huge update since then in the background, but also due to an decrease of user engagement.

My page targeting the trophy keyword basically never showed up in this SERP although it should be way more relevant. I guess, that this depends on the query and the number of links.

How are you doing regarding pageviews per visit and repeating visitors over the years?
Do you have some kind of responsive design?
My assumption would be a lack of user engagement.

@rainborick
Freshness can also be on the homepage, as in my case. I almost always have this "XX hours ago" notice attached to it. And it always outranks the more relevant page. So that doesn't have to be problem for overall rankings for a query.

[trabis]

Showing the main page when the query is more closely related to an interior page sounds like the interior page is seen as having much lower quality

This specific query(WIDGETS) is related with the main page and it is what the site is about. The site is about widgets and users post widgets. The main page shows a list of the last published widgets so I would say this page is always fresh.

I'd still say the first step is to look for Penguin-related issues on your site.

As in link issues? Most of the links I get are from blogs. People love to copy/paste widgets into their blogs and leave a link to the original page. I also use a famous tool that automatically appends a link to the content being copy/pasted. I never did link exchanges with other sites.

@tangor, thanks for the reply

Is this Info page or ecommerce?

None of the above. Site is supported by Adsense. All services/tools provided are free of charge. The site is in art/literature category. Users post their widgets hoping to get feedback from other users.

@seoholic, thanks for the reply

How are you doing regarding pageviews per visit and repeating visitors over the years?
Do you have some kind of responsive design?
My assumption would be a lack of user engagement.

I had an average of 3 pageviews/visit until late 2010
Then I had a linear drop until march 2012 hitting 2 pageviews/visit
After that it start rising again and it is now at 2,5 p/v
The pageviews for registered users is much higher.
I guess the drop may be related with ads above the fold which caused more bouncing. I do still display ads above the fold.

I don't use any kind of responsive design. Site is based on table/tr/td. Yeah, I know, but I like to use 3 columns.

Regarding user engagement, perhaps. I lost some users to facebook but so did my competitors right? Some users go, some users come. People still register everyday and some of them became very popular.

[Planet13]

I don't use any kind of responsive design. Site is based on table/tr/td. Yeah, I know, but I like to use 3 columns.

What percentage of your users are mobile device users?

And have you compared your bounce rate and other usage statistics between mobile and desktop users?

And there are CMS out there that allow for responsive design AND have three columns.

[trabis]

Hi Planet13!

What percentage of your users are mobile device users?

30%

And have you compared your bounce rate and other usage statistics between mobile and desktop users?

No I did not, let me check...

Desktop bounce: 74,73%
Mobile bounce: 79,05%
Desktop p/v: 2,57
Mobile p/v: 1,58
Desktop time on site: 00:02:40
Mobile time on site: 00:01:31

Percentage of new visitors is the same for mobile and desktop: 78%
Tablet stats are roughly between desktop and mobile.
I have some conversion goals I keep track (I.E. : posts, registrations) which account for 0,12% in desktop vs 0,09% in mobile.

And there are CMS out there that allow for responsive design AND have three columns.

Sure, I use Xoops CMS (I was a core developer there for some time) and they are releasing html5 responsive themes based on Twitter's bootstrap that are just wonderful.

I'm not a theme designer but I did attempt to modify one of this new themes to match mine. Thing is, it is pretty hard to set up 3 columns, each having different colors, and have the columns colored from top to bottom. For example, the right column should be grey from top to bottom but the grey stops at the last block that happens to be there.
There are solutions to this but they compromise the responsiveness of the theme.

And it is not just the theme, along with it I would have to fix every module templates/css. It's a tremendous amount of work and I'm not 100% sure it would worth the trouble. Developers are lazy (that is a good thing).

What I did to assist mobile users was to give up on dropdown menus and use contextual navigation on the left bar.

BTW, this forum is also old fashion.

Perhaps I should have used another title for this topic. I did not want it to be about my raking issues. My idea was to debate the relevance of the results when using time/date filters. I thought that searching for WIDGETS from "last year" might gives us a good indication of trending sites, upcoming competitors but apparently, it does not.

[seoholic]

In my niche besides my real competitors, there are also a lot of authorities and newspapers in the normal results. If I change the time span to a year, there are mostly authorities and newspapers and the content is less relevant to the query. Maybe some informational sites just don't get a lot of "XX hours ago" notices or some kind of "freshness rank" due to their structure, size and overall lack of changes to their content and therefore no good rankings with the date filter activated?

[Planet13]

@ trabis:

Is the nature of your topic something that one would normally expect to have updates on an hourly basis?

Meaning, is your topic generally a "news-oriented" site?

Or does it cover topics that are more or less stable but you get regular new user-generated posts about those topics?

[trabis]

@Planet13

I read the posting guidelines and I don't see a problem in telling you exactly what my niche is about. Poetry.

The biggest share of my competitors do not have UGC, they are built by a small team and one could argue they post better content and less frequently. They buy images, scrap famous authors poems, post some of their own and some that they don't know the authorship. They heavily use social networks and have great followers/fan bases.

The small share (me included) is based on UGC, we share part of the user base(some users post in more than one site) and one can also argue we provide the best content but with more frequency. We don't buy images, we get scrapped often and we promote copyright rights. We do not heavily use social networks because they compete with us.

Claiming quality for this topic is, I would say, impossible. I don't know how Google decides what of my pages is better for certain queries. I have content refereed by Google 100.000 times that does not have a single comment or up vote! That is why in the other thread about "making a new search engine" I suggested a system that would ask webmasters to deliver the response. Although I could not tell what was the best page, I could tell what was the most popular one according to my user base.

@seoholic

Maybe some informational sites just don't get a lot of "XX hours ago" notices or some kind of "freshness rank" due to their structure, size and overall lack of changes to their content and therefore no good rankings with the date filter activated?

I can give you that for the competitors that do not use UGC.
Perhaps Google sees them more as an informational site and mine more like a forum. They tend to have pages listing 10 items (like top 10 lists) posted on a daily/weekly base while mine is more focused on individual items posted every 5/10 minutes. I also make heavily use of date/time on posts and they don't. My search results snippets almost always have the date in it and they often date from 2 or 3 years ago.

But what about "last week" and "last month" and "last year". Should not informational sites that rank above me in the "Any time" also rank above me when using this filters? You see what puzzles me? I could rank #1 for "last year" forever and never rank #1 for "any time". It is not like Google is using a smaller data set, it is like it is using a new algorithm.

[Planet13]

I really don't know if I have an answer for you regarding why you rank one way with date filters applied, and why you don't when rank filters are not applied.

I would also once again encourage you to make your site as mobile friendly as possible, even if that means forgo some of the design elements.

I know you said that you don't really do social, but you DO make it easy for people to SHARE your site, right?

Also, you mentioned that it is user-generated, but do you do a lot of curating your site? Maybe make the top 10 user submitted poems about subject XYZ, or top 10 poems written by [demographic XYZ]

I don't really know though...

[trabis]

@Planet13

I know you said that you don't really do social, but you DO make it easy for people to SHARE your site, right?

Sure, I use social buttons under the content and html tags to inform the service about the title, description and image I want go with the url(canonical) being shared. I also have set up feeds to post the most popular content of the day to several social services. I don't manually/personally interact with this services except to invite people in. I have about 4500 fans on fb but that is nothing compared with the tens of thousands some competitors have. Perhaps they buy them? Don't know. I also have a "link to us" page where I show users the html code that they can copy/paste in their blogs, the available rss feeds, a box showing the most popular content of the day and/or a rhymer widget.

Also, you mentioned that it is user-generated, but do you do a lot of curating your site? Maybe make the top 10 user submitted poems about subject XYZ, or top 10 poems written by [demographic XYZ]

When I noticed that those competitors were ranking with top 10 lists and after knowing about Panda and thin content, I went into programming mode! In the beginning of this year I've added a new set of pages that pick up categories, tags and users and create pages with the most popular content to match them. I can also create pages using any filter I want. For example, I have top content by year, I create contests and have pages listing the winners, I have pages dedicated to special days of the year. Content is ranked by users using signals from "favorites", "like/dislike" and "comments" so this resulting pages can be considered as high quality.

It took 6 months to start noticing this pages in search results and I'm now steadily gaining 100 new visits each weak (not much but it is the few stats I see growing). I hope this pages can surpass the traffic I get from individual pages in the future.

Regarding my site performance let me just mention:
The year of 2013 was of constant decline.
In the beginning of 2014 I noticed an upward trend.
This last month of July was the first month(looking 2 years back) that I did better when compared to the same period a year before.
I just hope that in the next 2 years I can get the 50% traffic I lost.
It seems as if Google discarded all my history and I'm now crawling my way to the top (again).

I would also once again encourage you to make your site as mobile friendly as possible, even if that means forgo some of the design elements.

Noted.

I really don't know if I have an answer for you regarding why you rank one way with date filters applied, and why you don't when rank filters are not applied.

Perhaps looking at other sets of queries could gives us an answer.
Is Google using the same algorithm in a smaller data set?
Does Google use ETA gathered from the smaller data set?
What signals are dropped when using filters?
If we could reverse engineer the filtered results we could find out why we do not rank as well in the unfiltered ones.