There was a discussion somewhere about what leads people to view topics.
I decided to try adding a lot of keywords to a topic title to see if this would lead to more google search hits and hence more views.
I made this topic
and decided to record each night at about 10pm, the cumulatiive number of views.
After doing this for a while , I decided I needed a baseline for comparison, so I also recorded views of this topic with a short title
Then , after about a month, I thought maybe I neede a measure of whether any of the recorded views were just web crawlers, so I recorded two topics which were definitely only of local interest.
Graph was constructed in R.
What does it show?
Well,
the topic with the keyword-loaded title had a lower rate of viewing than the short titled Dinit topic. Ignore the initial jump… that is local viewers… the long term view rate of Chimera topic is slower.
the two ‘local interest only’ topics went to about 20 views then did not increase. Obviously web crawlersare not inflating the view records.
So I would conclude that keyword-loading a topic title is not going to help get more views. What helps is a good broad interest topic…more people are interested in Dinit than Chimera.
I believe that google searches both the title and the content of itsFOSS topics. So while a good title may help, the content is important too.
I want to apologise to the authors of the two local interest topics… I am not trying to outdo anyone there. All topics are of interest locally.
It is not a very well designed experiment.
Does anyone wish to disagree with the conclusion?
Well done !
It used to be keywords in a seperate section of html header, but now the google search engines (and others) not only look for titles but through the document and in pdf attached files. The indexing cost must be high in computer use given how many sites and documents there are and on so many subjects.
Glad now i have stepped away from search engine optimisation
I thought so. Some of my google searches bring up things not in the title.
There is something else interesting about that graph… it is very smooth… there are no surges of interest… nothing went viral… I guess the topic precludes that
and
As far as I can tell none of the external views generated a reply. There were a few local replies, then nothing.
Its more complex, depends on the site and if the pages are included in the research posibilities, you can exclude.
Also if they are set as pages in the index site map
Then the revisit option, you can tell the bot to only visit once per month or more often, so if you say once per month then update it could be a while before it comes back
Then if the page is formatted correctly no errors and no duplications
So you say that bots may be selective in their activities?
So I cant conclude there are no bots in the views by looking at static topics?
OK, so are bots likely to jack up views every day? If so my result is biased.
I dint know ehether bots register as views at all… it would depend on how they access the forum?
Lots of questions?
.
This is the google training program for search engine optimisations
I have been looking at the courses on the learning site
Trouble is, dont want to pay, dont want another certificate or qualification so undecided… the government is offering me money to take a course but cannot justify at the end of my career to take it.
Thanks for your meticulous work @nevj
That is indeed an interesting approach and I agree with the conclusion that keyword stuffed title doesn’t work as good.
The title should have keyword(s) but it should also be natural.
I tested something new this weekend (vague, doesn’t tell much so it will only have some traction with regular members/readers)
Exploring Chimera Linux in virt-manager with Plasma (straightforward, could work moderately for both existing members and web searchers)
My Experiment with Chimera Linux in virt-manager taught me a special thing (intriguing for both existing readers as well as web searcher)
Chimera Linux in virt-manager with Plasma | Wayland | APK | Clang/LLVM | musl | dinit | BSD core tools (too many keywords make it a weird sounding vague title)
What you did here is called A/B testing and it works a lot better if the contents are on the same topic.
In fact, pro-SEOs and bigger publishers use specialized tools to check which headline works better (generating more clicks on Google/selected platform… called CTR). That will run one headline for a few days and the second one for the next few days… whichever works better is then used permanently.
But things are a lot more complex as there are way too many factors involved on what makes a page hit.
The biggest factor for me is that an article should match searcher’s intent. If the titles says something, the content should be able to justify the title.
In other words, title should be clickworthy (enticing enough for the searcher to click) than clickbaity (searcher gets irked with the content as they were tricked into clicking).
That’s what I have to add on this topic from my end
Thanks @abhishek , you have added a lot.
I will try in future to make titles that arouse interest rather than be stuffed with keywords.
I do believe that google searches the topic content as well as the title, but what it presents to the searcher is the title plus a content extract.
Therefore the title is important, because it is a large part of what the searcher sees.
