Keyword
Clustering:
Some cool shit you might not
have thought about
Slide 3
Slide 3 text
Who am I?
3
• Worked in SEO since 2013
• Worked both in-house and agency
side.
• Grew my own start-up to a turnover of
£2.5 million when I was 25.
• Co-founder of Keyword Insights - a
topical authority platform
@digitalquokka
Slide 4
Slide 4 text
What is keyword
clustering?
4
Keyword clustering is the
process of gathering similar
keywords into groups.
@digitalquokka
Slide 5
Slide 5 text
It makes turning large, unorganised
keyword lists into something more
manageable
5
@digitalquokka
Slide 6
Slide 6 text
There are 2 methods of keyword
clustering:
6
Serp Clustering
Semantic\NLP
Clustering
@digitalquokka
Slide 7
Slide 7 text
NLP/Semantic Clustering:
7
Pros Cons
Cheap/free.
Can be less confusing
to look at the results.
We operate in
Googles world. NLP
isn’t accurate for that.
@digitalquokka
Slide 8
Slide 8 text
SERP Clustering:
8
Pros Cons
We operate in
Googles world. It’s
much more accurate
for that.
Will always cost
something (scraping
isn’t free).
Results can
sometimes be
confusing.
@digitalquokka
Slide 9
Slide 9 text
Let me explain this one:
9
Pros Cons
We operate in
Googles world. It’s
much more accurate
for that.
Will always cost
something (scraping
isn’t free).
Results can
sometimes be
confusing.
@digitalquokka
Slide 10
Slide 10 text
Keywords can sometimes fall into
multiple clusters (keyword
fragmentation)
10
@digitalquokka
Slide 11
Slide 11 text
But there are some cool
things you can do with a
SERP based clustering
tool… 🤩
11
@digitalquokka
Slide 12
Slide 12 text
12
Solving Keyword
Cannibalisation
Slide 13
Slide 13 text
Case Study 1
13
● Homes for sale in …
● Houses for sale in…
● Properties for sale in…
● Log cabins for sale in…
● Wood cabins for sale in…
+ another 400-odd more.
= >50m URLS!
@digitalquokka
Slide 14
Slide 14 text
We believed keyword
cannibalization and crawl budget
were an issue
14
@digitalquokka
Slide 15
Slide 15 text
So we clustered all the category
variations
15
And redirected
over 15m URLs!
Slide 16
Slide 16 text
The result? 110% rise in organic
traffic
16
Slide 17
Slide 17 text
You can do something similar for
other niches:
17
E-commerce Travel
Slide 18
Slide 18 text
Case Study 2
18
@digitalquokka
Slide 19
Slide 19 text
We started with their sitemap
19
@digitalquokka
Slide 20
Slide 20 text
And turned the URLs to “keywords”
20
=SUBSTITUTE(SUBSTITUTE(REPLACE(A2,1,LEN("https://buzzoid.com/"),"") , "-", " "), "/", "")
Slide 21
Slide 21 text
We then clustered the “keywords”
21
@digitalquokka
Slide 22
Slide 22 text
Used vlookup to turn the
“keyword” back into a URL
22
@digitalquokka
Slide 23
Slide 23 text
We see this a lot with companies
whose content production isn’t
centralised
23
@digitalquokka
Slide 24
Slide 24 text
24
Keyword Mapping
Slide 25
Slide 25 text
Using the same
technique as before, we
can also map keywords
to URLs 💡
25
@digitalquokka
Slide 26
Slide 26 text
Start by turning the URLs into
keywords
26
@digitalquokka
Could also be
the H1 or
page title if
they’re better
Slide 27
Slide 27 text
Add a “we have this” identifier
27
Slide 28
Slide 28 text
Give them a realllllyy high
search volume
28
Slide 29
Slide 29 text
Finally, cluster them with your
keyword research
29
Slide 30
Slide 30 text
With a bit of vlookup magic, all
your keywords are mapped
30
Slide 31
Slide 31 text
31
Keyword
Un-Cannibalisation
Slide 32
Slide 32 text
Keyword Clustering/SERP Similarity
showed pages we didn’t think we
needed
32
Slide 33
Slide 33 text
33
Discovering
Content
Opportunities
Slide 34
Slide 34 text
Idea 1 - We can use keyword
clustering + rank to find content
gaps
34
@digitalquokka
Let’s say Hubspot wanted to own the topic of “keyword research”.
How do we know what content they already have and don’t have?
Slide 35
Slide 35 text
Start by collecting an
indiscriminate keyword universe
on the subject
35
@digitalquokka
Slide 36
Slide 36 text
Then whack them all into a
clustering tool that also pulls in
rank
36
@digitalquokka
Slide 37
Slide 37 text
Quickly filter in the report for
“clusters” of keywords where we
don’t rank
37
@digitalquokka
Slide 38
Slide 38 text
Idea 2 - “Content Fragmentation”
38
Let’s say I want some
quick and dirty content
ideas around this
product...
@digitalquokka
Slide 39
Slide 39 text
Simply google a related keyword 👇
39
Slide 40
Slide 40 text
Grab all the keywords for the top
ranking URL(s)
40
@digitalquokka
Slide 41
Slide 41 text
Reupload them into a clustering
tool (with rank)
41
Bang! 25 new ideas.
We’ve fragmented
the top article out
into at least 25
separate content
ideas.
@digitalquokka
Slide 42
Slide 42 text
This is a great strategy when you
want to compete with the big
leagues
42
@digitalquokka
Slide 43
Slide 43 text
Idea 3 - “Zero volume keywords”
43
Challenge: How can a new
site with no authority enter a
competitive niche?
@digitalquokka
Slide 44
Slide 44 text
Everyone has access to keyword
tools. So we turned to other sources
of inspiration
44
50M+ Daily Active Users
100K+ Active Communities
13B+ Posts and comments
@digitalquokka
Slide 45
Slide 45 text
There’s nearly always a related
subreddit for a given niche 👇
45
Slide 46
Slide 46 text
When we cluster all the questions we’ve
scraped, “zero volume keywords” aren’t
so zero volume
+ 1000’s
more
ideas
Slide 47
Slide 47 text
Bonus: Enrich your questions by feeding
them back through PAA
Slide 48
Slide 48 text
The strategy works well…
Slide 49
Slide 49 text
We actually made a free tool to
make this super easy to do 🤯
Slide 50
Slide 50 text
But, in the first instance you could
do this with your search console
queries
50
Slide 51
Slide 51 text
If you do decide to use Keyword Insights,
we have an integration to make that
easy too 🤩
51
Slide 52
Slide 52 text
We also make it easy for you to
actually write the content
52
@digitalquokka
Slide 53
Slide 53 text
53
Summary
Keyword clustering can (and should) be used in some pretty
clever ways.
● Finding keyword cannibalization instances
● Finding keyword un-cannibalization opportunities
● Keyword Mapping
● Helping with your keyword research and content planning by:
○ Letting you group loads of “zero volume” keywords
together
○ Quickly find content gaps without much planning
○ Fragmenting content out