User talk:Prioryman/Use of SEO techniques on Wikipedia

Comments

Interesting study and analysis. LadyofShalott 18:19, 16 August 2011 (UTC)

As of 18:49, 16 August 2011 (UTC) Whorlton Castle appears to be ranked #5 on Google results and #1 on Bing and Yahoo. I don't know enough to suggest why this might be, though I do note that the page view statistics tool indicates all but 282 of the to-date ~10k page views took place on August 15, when the DYK was displayed. Would this be responsible for pushing up the page rank? Or has Google's marketing / localization / homing skewed my own personal search results? ☯.ZenSwashbuckler.☠ 18:49, 16 August 2011 (UTC)

Googling from the UK, entering whorlton castle as the search term, it shows up in tenth place for me, as Prioryman says. (The same is true when entering "whorlton castle" as the search term.) Cheers, --JN466 19:08, 16 August 2011 (UTC)

OK, it makes sense that the WP page would be higher-ranked in Massachusetts (and much of the rest of the U.S.?) than in the UK. Still shows up #5 for me after clearing cookies and browser cache, so it must be location-specific rather than tailored to previous searches. Thanks for helping clear that up. And please disregard my previous edit summary. ☯.ZenSwashbuckler.☠ 19:17, 16 August 2011 (UTC)

It's also in 5th place for me (Southeastern US). LadyofShalott 21:01, 16 August 2011 (UTC)

It makes sense that Google would have some degree of regional weighting; if you're searching from the UK, it would be more likely that UK websites would be more relevant to you. Prioryman (talk) 23:17, 16 August 2011 (UTC)

I'm using Google UK; other regional Googles may produce different results. I'm aware of Google's use of customisation to refine search results for individual users based on your previous searches and have tried to allow for this by performing searches from multiple networks, without logging in to Google accounts of anything like that. The results have been identical in all cases. In any case, though, it isn't the absolute ranking that is important here but changes in the rank. My experiment suggested that nothing I did to the article in terms of adding links, templates etc caused any movement. Prioryman (talk) 19:13, 16 August 2011 (UTC)

The article is the 12th result for my searches on Google.com, both with and with quotation marks. Delicious carbuncle (talk) 19:57, 16 August 2011 (UTC)

Overall this is a nice bit of analysis. While this is useful as a counterexample to puncture some existing theories, I think the experiment should be repeated a few more times before new conclusions can be drawn. It shouldn't be necessary to create a new article for this, the effect of adding a link to a template can be measured with an existing article as well. It would be interesting to know if Google factors in cleanup templates. For example does the the ranking go down if an article is PRODed?--RDBury (talk) 02:51, 18 August 2011 (UTC)

• I find that Whorlton Castle has now moved from being the no. 10 Google result to being the no. 2 Google result for google.co.uk searches using "Whorlton Castle" as the search term (conducted from a UK location).
• Of the 61 links to the page presently shown on Yahoo! Site Explorer, 60 are Wikipedia links: [1]. (The other one is a Danish Wikiquote link.)
• In google.co.uk searches for Whorlton Castle (without quotation marks) the Wikipedia article is now in 11th place (one lower than two weeks ago).
• In google.co.uk searches for Whorlton, the article Whorlton Castle is in 120th place. [2]
• Do other editors obtain the same results? --JN466 03:56, 29 August 2011 (UTC)

No. Still #11 for me, as it has been for some days (tested from multiple networks, cookies turned off etc). No difference between searches with quotations marks and without. Mathsci is getting similar results [3]. Prioryman (talk) 07:21, 29 August 2011 (UTC)

Additional data

Interesting study indeed. Thanks for performing this experiment. Here's a few data that may help tracking this article's further rise in Google:

• Presently, Google indicates the article page has 29 external links from .edu, .co.uk, .com, .co.uk, .co.au, .net and .org sites (excluding Wikipedia itself), or 53 external links with the "very similar" results included.
• Presently, Google indicates the article has 55 internal links from other Wikipedia pages.

Interestingly, the 29 external links Google thinks exist include this page, which (as far as I can see – am I missing anything?) does not link to Whorlton Castle at all, but only includes a link to Blickling Hall. (The reason seems to be that Whorlton Castle was mentioned in a question on that site that was present in the left hand sidebar of pages that linked to Wikipedia: [4]) Cheers, --JN466 19:03, 16 August 2011 (UTC)

Don't forget that quite a few websites scrape content from Wikipedia for their own purposes. I'm sure that accounts for most if not all of the external links. It is unclear whether this has any effect on ranking; I suspect not, otherwise we would see Wikipedia knockoff pages being given a far higher rank than they seem to do. Regarding internal links, many were added by myself through the addition of a template and a few through direct editing. As documented in the study, this linking had no observable effect on page ranking. Prioryman (talk) 19:19, 16 August 2011 (UTC)

Note: I added some details to my above post after Prioryman's reply: Diff. --JN466 19:22, 16 August 2011 (UTC)

I have found that there is a time lag in Google's response. For example, after renaming an article, the old name may still outrank the new for several days after the renaming, or both may show up next to each other. It will be interesting to see how the page rank of this page develops, along with the increase in both internal and external links to it. --JN466 19:27, 16 August 2011 (UTC)

Looking at the above results, especially the ones from the Tudor site, I have a feeling the "link:" argument in Google no longer works as advertised. FWIW, Yahoo! Site Explorer shows 18 links to Whorlton Castle right now, all from within Wikipedia: [5] --JN466 20:50, 16 August 2011 (UTC)

The number of inlinks in YSE has now gone up to 27, all bar one from within Wikipedia. Interestingly, about half the pages that include the Castles in North Yorkshire navbox still aren't listed among the inlinks. Whorlton Castle is still at number 10 for Google searches performed on google.co.uk, and from a UK location. (It's number 11 for google searches performed on google.com from a location in the UK.) --JN466 18:11, 17 August 2011 (UTC)

The increase seems to be largely due to bot activity on Wikipedia. No obvious sign of effects on page ranking though. Prioryman (talk) 19:25, 17 August 2011 (UTC)

Agreed. We are now at 48 inlinks, still all internal except one, and no change to page ranking so far. --JN466 01:34, 22 August 2011 (UTC)

Is a longer test possible?

Excellent test, Prioryman. Thank you for looking in to this. I was wondering if you might run a longer test. Specifically, I would like to see what happens to the rank if the page is left to its normal evolution (i.e. people update it as normal not knowing this test exists). Possibly letting it go 30 days like this and following the daily ranking. Then try again to increase the rank using the techniques suggested at Wikipedia:Wikibombing (SEO) once again to see if there is any effect. I may be asking too much, but I think it could be informative. Thanks for your consideration. - Hydroxonium (T•C•V) 02:37, 17 August 2011 (UTC)

I already have something like that in mind. I added as much information about the castle as I could from available online sources. I've tracked down offline sources that may provide substantial additional information. That should allow me to expand the article further. It will be interesting to see whether that has any effect on the rankings. Prioryman (talk) 06:56, 17 August 2011 (UTC)

That's great. Thanks very much. I appreciate your efforts. - Hydroxonium (T•C•V) 04:35, 18 August 2011 (UTC)

Internal links – important or not?

What is editors' view of the analysis in this video, by an SEO consultant, claiming that internal links play a very significant role in the high ranking of Wikipedia pages? --JN466 18:01, 17 August 2011 (UTC)

The narrator mumbles and doesn't clearly explain what he's doing at each step. But one general comment I can make is that the video is two years old, and Google is known to make constant changes to their algorithms.   Will Beback  talk  00:06, 19 August 2011 (UTC)

Explanation of "santorum" rise and fall in search engine rankings?

As several editors observed, the article now known as Campaign for "santorum" neologism rose to the first result on Google, surpassing even the "spreading santorum" site. The article is no longer the first result for me. I am curious if the conclusions drawn by this study are consistent with the observed behaviour of that particular article in Google ranking. Can someone with SEO expertise could comment on this? Delicious carbuncle (talk) 17:39, 18 August 2011 (UTC)

You're merely citing anecdotes – the plural of anecdotes is not "data". Observing a change in ranking on an unrecorded date is useless without having any data to show whether there is any correlation to changes to the article, or any external occurrences (like Colbert) that may have influenced ranking. The purpose of my study was to specifically correlate article rankings over time with editorial actions. The data I recorded enable us to know exactly what the rank was on any given day and what if anything done on Wikipedia might have affected it. Unless you can provide similar data for Santorum, no inferences can be drawn. Prioryman (talk) 19:18, 18 August 2011 (UTC)

It happened, was commented on at the time, and still seems inexplicable. I've added Ravensworth Castle (North Yorkshire) to the template.

• Ravensworth Castle (North Yorkshire) is currently at no. 17/18 in Google.co.uk, and no. 23 in Google.com (accessed from the UK, searching for Ravensworth Castle with/without quotes).
• Ravensworth Castle (North Yorkshire) is currently at no. 2 in Google.co.uk and Google.com (accessed from the UK, searching for "Ravensworth Castle North Yorkshire" with quotes)
• Ravensworth Castle (North Yorkshire) is currently at no. 4 in Google.co.uk and Google.com (accessed from the UK, searching for "Ravensworth Castle North Yorkshire" without quotes) --JN466 13:19, 27 August 2011 (UTC)

YSE inlinks: [6] (showing none at present). --JN466 13:33, 27 August 2011 (UTC)

Old article

See here, I haven't read it yet. Count Iblis (talk) 16:30, 13 September 2011 (UTC)