A bug with %keyword% inserting
Hey @sven,
I think I have found a bug with the %keyword% insert functionality.
I have a setup in which there are 5 keywords used to create what I call 'dynamic content templates'.
The keywords are in an order like this:
New Zealand,New Zealand,New Zealand dollar,New Zealand dollars,NZ$
In this H-tag file, %keyword5% is the only one mentioned. I checked the file 100 times, looking at every single mention manually, and there are no mentions of anything besides %keyword5%, yet SER is using %keyword3% (and/or 4? not sure) which makes the text no make sense.
I initially thought that maybe it was because keyword1 and keyword2 were the same for this particular project, but was able to reproduce the bug in another country using different ones (i.e., Germany, German).
The H3 tag is being pulled in like this:
An example of the spin inside this file would be something like:
%keyword5%{100|200|300|400|500|600}
It's hard to reproduce this bug consistently, but I've now done 6 network posts with my new setup, and the bug exists in 2/6, so it seems like a pretty common occurrence. Note: It's appearing in live outputs, not just the preview (I have yet to see it in the preview window, though admittedly, I haven't done a lot of previews to see).
Last note: Just realized the bug is worse than I thought. It's now inserting %keyword3% in all spots for pretty much all keywords. I can't figure out what would be causing this.
This output was supposed to be %keyword1%. Yet it's always using keyword3. In my articles, I can confirm that 100% of the keyword inserts are being output as %keyword3% or %keyword4%.
Thanks!
I think I have found a bug with the %keyword% insert functionality.
I have a setup in which there are 5 keywords used to create what I call 'dynamic content templates'.
The keywords are in an order like this:
New Zealand,New Zealand,New Zealand dollar,New Zealand dollars,NZ$
In this H-tag file, %keyword5% is the only one mentioned. I checked the file 100 times, looking at every single mention manually, and there are no mentions of anything besides %keyword5%, yet SER is using %keyword3% (and/or 4? not sure) which makes the text no make sense.
I initially thought that maybe it was because keyword1 and keyword2 were the same for this particular project, but was able to reproduce the bug in another country using different ones (i.e., Germany, German).
The H3 tag is being pulled in like this:
<h3>%file-C:\Program Files (x86)\GSA Search Engine Ranker\folder\h3\generic.dat%</h3>
An example of the spin inside this file would be something like:
%keyword5%{100|200|300|400|500|600}
It's hard to reproduce this bug consistently, but I've now done 6 network posts with my new setup, and the bug exists in 2/6, so it seems like a pretty common occurrence. Note: It's appearing in live outputs, not just the preview (I have yet to see it in the preview window, though admittedly, I haven't done a lot of previews to see).
Last note: Just realized the bug is worse than I thought. It's now inserting %keyword3% in all spots for pretty much all keywords. I can't figure out what would be causing this.
This output was supposed to be %keyword1%. Yet it's always using keyword3. In my articles, I can confirm that 100% of the keyword inserts are being output as %keyword3% or %keyword4%.
Thanks!
Comments
1) The lines of content must be pulled using #file_links from an external source, using lines to create a content template.
2) I can confirm that this issue is related to words not being included in the article when SER tries to stuff. Meaning, if you have an anchor "the cool games," SER will replace "the" with this anchor. When SER cannot find a matching word, for example, "Domain.com" this bug happens. But again, only when using content templates so it seems.
A more important bug directly related to this:
(This one is much easier to reproduce)
When using content templates, SER now only inserts non-word matching anchors in the top 4 paragraphs. If you use a static article, it's random, but when using lines of content templates / #file_links to pull in content, SER will stuff the anchor in the first paragraph ~50% of the time, and then only the top 4 parahraphs 100% of the time.
Upon testing, this is directly related to the bug above.
1) This bug only occurs when using content templates #file_links to pull in lines to create articles.
2) It only occurs when SER cannot find a "matching word". So, as with the above example, if the anchor is "the cool website," SER will find "the" and replace it randomly with this anchor. However, if the anchor is "Domain.com," SER cannot find a matching "stuff word," which causes the bug -- it only stuffs it in the top 4 paragraphs, losing a huge amount of randomization.
This one is very easy to reproduce as it happens consistently, 100% of the time.
But overall, I can say with certainty that both of these bugs are related to:
1) Content templates / #file_links for pulling in lines of content/content templates.
2) SER not finding a "stuff word" to replace when inserting the anchor text.
Update:
I can now confirm with 100% certainty that both of these bugs did not exist in SER version v13.95. I have done 50+ previews on 13.95 using the same project(s) as 13.96, and the issue has not happened once, the inserts are randomized properly and the URL/"non-matching stuff word" anchors never end up on the bottom of the articles.