When we prospect for links and resource pages – the classic target for the white-hat style link begging tactic – we often wondered which advanced queries are most-likely to return to us the pages we’re after.
Here are the 10 most-productive queries in case you have links-page prospecting to go and do right now. Read on to learn where the numbers came from…
- links – 95
- resources – 72
- inurl:links.html – 54
- inurl:links.htm – 52
- links to resources – 51
- of links for – 50
- intitle:”helpful links” – 48
- intitle:links – 47
- other helpful links – 46
- links to websites – 46
So if link prospecting is new to you, you’d take the 10 lines above (yes, do remove the numbers) and paste them into the middle box of a tool like this. Then put your topic in the first box and hit submit. Then search all those queries that come out in Google in a tool like this (warning – hard sale ahead at that link).
But wait Garrett… why would you want to know the most productive queries? Why not just combine with your topic keyword and run all 700 of our queries (which is often what we do).
Our motivation for this exercise was to isolate the key queries to run when we’re doing LNR topic discovery as well as Topic Keyword accuracy assessments. Basically we’re trying to become better informed about productive topics.
But also wait… Garrett… Where did those numbers come from?
Ah yes! So glad you reminded me.
So here’s how we figured this out.
First we took 5 topics that I am unwilling to share with you at the moment (probably they are on this list though). Then we combined them with our 700 or so known query stems that bring back potential links and resource pages. So from those 3500 queries we then recorded by hand all 35,000 pages in a spreadsheet – keeping careful track of which queries brought back which pages.
We then ran those 35,000 potential pages through the following test to determine if it’s a probable link page:
- The page must have greater than 7 unique outbound domains
- The page must have greater than 15% link-to-text density (weeding out articles, etc… yes they ARE good targets but not what we were technically after)
And THEN we were able to assign a count of probable links pages discovered to each query.
So in case you were wondering, of those 35,000 pages analyzed, 14973 looked like LNR pages according to the above criteria. Note: that’s not a unique page count – there are almost certainly duplicate domains and URLs across the queries.
Great! So, actually Garrett, who cares? What’s the big deal? Well honestly I’m not sure but this is the work that’s paving the way for us to have a better understanding of the relationship between any given topic and the likelihood of LNR pages actually existing for it… or rather, of finding where and how any given target keyword fits into the LNR topic space. So it’s not really that cool now except we did a test and have data. Also I like it.