Why I don’t use KD [And the June 2022 Report]

Here's why I don't use the KD metric (and what I use instead)

A few days ago, I tweeted about KD (keyword difficulty) scores and why they’re not necessarily an indication of the level of competition.

The replies turned into an interesting little discussion, with some people asking what I thought a good alternative would be to the KD score.

My answer was too long to fit into a single tweet, so I want to try and address it here in a post. And I’ll also share the monthly report about four of my sites.

Why I don’t use KD scores when choosing topics

My go-to method for finding new topics is good old Google Autocomplete. I do use Ahrefs occasionally when I run out of topics and need some new ideas.

When filtering Ahrefs results, I pretty much ignore that colorful column of KD (Keyword Difficulty). In other words, I don’t deliberately pursue low-KD terms. Nor do I avoid high-KD terms.

Let me try and explain why.

As always, everything in this post is just my opinion, and your mileage may vary.

If you think I’m wrong, leave me a comment and explain why. I’m always looking to learn, and nothing would make me happier than improving my keyword analysis methods.

Automated keyword research?

My definition of a perfect topic is pretty much this:

Lots of searches but no other site competing over the query.

When I encounter a topic like that, I know I hit gold. I get a dopamine rush. It’s the best thing ever.

I happily settle for low competition topics – with only 1-2 sites going after the same query.

The golden nuggets are out there. The question is, how do you find them?

We’re looking at two aspects here –

  1. High search volume (potential traffic)
  2. Low competition (easier to rank your content)

There are many tools out there that offer ways to predict both of these aspects. Ahrefs is one of the more popular ones.

You can use Ahrefs to find keywords/topics in various ways and generate reports with tens of thousands of potential queries.

But how to choose which topics to cover on your website?

The metrics most people look at are Volume (or traffic potential – that one makes more sense to me) and Keyword Difficulty.

In theory, you could filter the above list of 120,739 keywords generated for “Golden Retriever” and narrow the list to those with high search volume yet low competition.

I went ahead and limited the report to Keyword Difficulty of 5 and lower and a search volume of at least 100. I also filtered for questions because I deal with informational websites.

Voila – I have my list of topics for a new niche site!

Keyword difficulty is all bright green here. It shouldn’t be a problem to rank with a new site, right?

Well, not really.

I wouldn’t rush to buy a Golden Retriever domain just yet. Not based on this list, at least.

Here’s why.

Why the KD score is problematic

The KD metric is tempting because assessing competition levels can be so tricky. I can see why relying on a precise scale with colorful grades is appealing.

But those colors can be misleading. They don’t reflect the level of competition.

And here’s the thing. If you read Ahrefs’ blog post on KD, you’ll see they actually say as much. They don’t suggest using it blindly or taking the score at face value.

What Ahrefs does is simple enough –

Here at Ahrefs, we use a simple method for calculating KD. We pull the top 10 ranking pages for your keyword and look up how many websites link to each of them. The more links the top-ranking pages for your keyword have, the higher its KD score.

Unfortunately, the backlinks profile is not the only factor at play. When assessing the competition, there’s more to look at.

Title/post relevance matters

People want a quick answer to their questions. Google wants to deliver the best possible search results for a query.

Therefore, Google prefers low DR pages that answer the specific question over high DR pages that don’t.

IMHO, whether or not a topic meets the user’s search intent is the most important thing. Keyword Difficulty can be bright orange, and I would still go after that query if no other site has answered the question.

Niche authority seems to matter.

I sometimes see niche-specific sites win over larger, more general sites with the same post title. It’s not always the case, and with so many other variables at play, it’s hard to tell.

If it is a factor, then if your site is more focused, it may be able to compete against high DR sites that are more general in scope.

Low DR sites can easily outrank high DR sites

I’m seeing this more and more lately.

Both sites have on-topic pages. The content is similar, and so is the level of niche authority.

And yet, the low DR sites rank higher than the high DR sites.

I’m not sure why this is happening. It could very well be that the low DR site had a focused link campaign promoting the specific article.

Or it could be something else with the complex Google algorithm.

All I know is that I see it more often these days.

Low KD does not mean low competition.

The bottom line is that KD in itself is not a good way to assess how likely you are to rank for the term. Even Ahrefs tells you as much.

KD merely reflects the backlink profiles of the top 10 sites in the SERPs.

Low KD means those sites may have relatively few quality backlinks. They may still have –

  • On-topic content
  • High-quality content
  • Niche authority
  • Enough links to outweigh a new site anyway

And most importantly: There may be quite a few of them!

Even KD zero doesn’t mean you’re dealing with low competition.

I just Googled the question. Guess what – one snippet + 8 existing sites hit the search intent heads on. These sites have a Moz DA of 11 to 35. For a new site, that’s a lot of competition.

Could you compete over this query?

Maybe. Just be prepared for a struggle. This is not low-hanging fruit in my book.

So, what’s my alternative to KD?

Ahrefs is a great tool. I do use it occasionally when I run out of ideas for topics.

However, I analyze each query manually by Googling the question and reviewing the top 10 results.

  • Do they address the search intent?
  • Are they good quality?
  • Are the sites known players in the niche?

I still haven’t found an automated tool that can answer the above questions. If you know of one, please let me know in a comment.

Is my method foolproof?

I wish it were.

Depending on the niche, I can probably get half of our content to rank in the top 3 places on Google. Maybe 70% to rank on the first page. The rest of the posts are duds.

The model still works for me because I just need an average of 300 pageviews per post to make a decent profit.

I have found that my ability to assess queries improved over the years.

Keep in mind that I have published more than 15,000 posts to date. I constantly go over Google Analytics and see the results.

I think that by looking at what worked and didn’t work, I developed some level of intuition. In other words, there’s value in being experienced.

Speaking of looking at what worked and what didn’t, let’s move on to this month’s report.

The June 2022 Monthly Traffic and Revenue Report

A quick recap:

I own a portfolio of more than 20 content sites, monetized mainly with display ads. Today, my web publishing business makes more than $190K in monthly revenue (Yes, that number kept going up in June!).

I documented my progress in this blog, including detailed monthly reports covering all my sites. In 2022, I switched to reporting traffic and revenue on four new sites.

You can read the initial report here. That post includes more information about my web publishing business and an FAQ. Please check it out first if you have any questions.

My system in a nutshell

My system is not very complicated.

  1. Find suitable topics.
  2. Create good content.
  3. Scale by outsourcing, using a good workflow.
  4. Rinse, repeat.

I blogged here about my workflow.

If you want to learn how to find suitable topics and produce the right content, check out the courses on my resources page.

I won’t repeat each site’s story – only the basic stats. Please refer to the initial report to learn more about a site.

Site #1

  • Niche: General (the site does have a unifying theme)
  • The first post was published on April 2, 2021
  • Number of published posts: 210
  • Monetization: Display ads + Amazon

Traffic went down in June. I blame the Google update. Our larger sites increased in traffic, so I’m not complaining. Overall, June was our best month in terms of traffic and revenue.

However, some of the smaller sites were hit by the update, including site #1 and site #2 in this report. I should not be complaining since these two sites only lost around 5% of their traffic. So many web publishers suffered much more significant losses.

Fortunately, June has high RPM rates, so site #1 actually made a little bit more money – with less traffic.

Site #2

  • Niche: Home & DIY
  • The first post was published on June 11, 2021
  • Number of published posts: 226
  • Monetization: Amazon & display ads

It’s a similar story with Site #2, with a decline in traffic due to the algorithm update. Here too, higher RPM keeps overall revenue relatively high.

We decided to pump some more content into both sites during Quarter 3. They justify the additional investment with a “dollar pet post” average of $8-$9. Whatever we put into them will start to turn a profit in under a year.

Site #3

  • Niche: Pets
  • The first post was published on May 18, 2021
  • Number of published posts: 119
  • Monetization: Amazon
  • Special circumstances: I created the site on an aged domain I’ve had for about 20 years as an in-house experiment. I knew the domain was clean, and all incoming links were on topic.

This site remains stable. We didn’t notice any effect from the algorithm update here, and higher RPM rates meant more money came in. Yes, overall, the revenue per post is low.

Site #4

  • Niche: Home & DIY
  • The first post was published on August 23, 2021
  • Number of published posts: 214
  • Monetization: Amazon

Site #4 keeps slowly growing but is still at the low end of the traffic spectrum.

I’m glad site #3 and site #4 are included in this report. It helps to show that not all sites are successful right away. I feel that I managed to avoid a survivor bias by randomly selecting four sites to follow this year.

If you’re new to web publishing, you should be aware of that. Too many people try to sell you a course by showing off their successful sites. The ones that don’t do so well remain out of sight.

That’s why I promote courses by Jon Dykstra and Morten Storgaard. Both share site reports on their blogs/channels without cherry-picking sites after the fact.

So, there you have it. That was the report for June 2022 on these four sites.

As always, I’d love to get feedback from readers!

I’d love to hear more about your keyword research experiences – the good, the bad, and the ugly! Let me know what you think about keyword research tools. What are your favorites, and how do they compare when assessing keyword difficulty?

22 Comments

  1. Anne,
    You are absolutely right about ahrefs KD. It is based on the external links pointing to the pages listed on the first page of serps.

    Semrush seems to have improved their keyword difficulty score and is more reliable that before.

    One thing I like about these tools is the ability to see the snapshot of the serp preview. By looking at the title tag and url, we will get a rough idea on how many of the search result pages are directly targetting the keyword.

    • Hi Sunny,
      I agree seeing the snapshot of the results is helpful. For me, seeing it on Google makes more sense, but that’s just me. I should give Semrush another go sometime – thanks for the feedback!

  2. Hi Anne,

    Thanks for the write up. I agree about KD scores.

    After browsing Yeys and learning about your astonishing success, the brain of mine generated a question – What’s the end game for your publishing empire?

    Is there a point when you will stop publishing in volume? do you plan to sell your portfolio?

    I wouldn’t mind on reading a post about this.

    Thanks.

    Keep up the work)

    • Hi Dan,

      That’s a great question – I think everyone should have a goal in mind! In my case, we’re in the US based on our business investment, so I want to keep it viable and profitable for the next few years, until my son get his PhD degree here. Then we’ll see.

      I want to sell off some of the sites, but keep a robust portfolio of self-owned sites as well. Selling sites would provide an additional revenue source, and is not intended as an “exit”.

  3. Hi Anne,

    Thanks again for sharing your knowledge with us, and in addition, all praise for your fantastic earnings report. You are a real inspiration for me, and I think for many others.

    I was curious about your take on large sites versus sites focused on a specific niche.

    Take for example a site about musical instruments, versus a niche site focused on a specific instrument.

    When both sites go equally deep into each specific instrument, with clusters, etc. Which site do you think has more potential, and more chance of ranking high on Google? And what are the pros and cons of both concepts?

    I’m very curious to hear your take on this.

    Thanks for the inspiration!!!

  4. I have two questions:

    1) What extensions do you use when doing the autocomplete method? Do you look at Moz rank and keyword volume with keywordseverywhere?
    2) Do you update old content? If yes, how? Must be a pain to update those 15k articles.

    • Hi M Guy,
      1. Just the usual question words – is,can,does,what,where, etc. I see the Moz rank and keyword volume with Keywordseverywhere but I don’t base my decision on those.
      2. No, I don’t update old content. Have tried that several times, using various strategies. It did not help with rankings (compared to articles in a control group).

  5. Wow, another record month Anne, simply amazing!

    Question:

    You will have undoubtedly worked with a large number of writers now. Based on your experience, on average what percentage of writers that you hire would you save are excellent, good, average and bad?

    For each writer that you end up hiring, how many people on average do you usually need to interview?

    • Hey Viral,
      That’s an interesting question there. We still use the same method described here. The initial filtering is done based on the application form. Then we have test posts. We only interview people based on a successful test post.

      I’d say about half of our test posts get rejected. Then we talk to the ones who passed the test post. They need to be good enough to make it to that point, so I’d say about 90% are at least good. Maybe 20% are great at that point. We end up working with most of the people we interview – at least 90% of them. The ones who aren’t good enough don’t make it to the interview.

      We then offer paid training and our editors keep working with the writers on improving their writing.
      Overall, if I had to put numbers/percentages on this, I’d say that writers who have been with us for at least a couple of months – 30% are great and the other 70% are good. And that’s all we need, really. Anyone who is not good, we’ll part ways with quickly enough.

  6. Hello Anne,

    Insightful read. Signed up here from Jon Dykstra’s forum and was present for your AMA run. Great value and appreciated your answer to my question and the other your answers there too.

    On the side of “low-hanging fruits”. Have you tried lowfruits.io? Instead of KD you can find topics that are user generated that you may find useful and be able to beat.

    Would love to know your thoughts.

    • Hi Evonne,
      I’m glad you enjoyed the AMA! That was quite a thrill to run the AMA on that weekend πŸ˜€

      I have tried lowfruits.io. I think I bought $25 worth of credits and I still have half of them, left unused. I think it’s a good tool and I like that it’s focused and affordable. The support is great too. I had a technical issue there at first, and the owner worked with me on fixing everything.

      I believe Lowfruits, just like other tools, basically looks at backlink profiles (DR/DA). They have a nice touch there, with looking at UGC sites, but I still need to judge for myself and see whether user intent is met on the search results. Personally, it doesn’t save me time. I needs to see the Google results either way, so might as well just use Google autocomplete.

  7. Hi Anne,
    Thank you for sharing your knowledge. It’s very valuable.
    Quick question:
    Do you do any internal linkings between your posts? If so, what process do you follow? Do you use a plugin?

    Thanks for your time

    JC

    • Hi JC,
      We ask writers to add 2-3 internal links within a post. We tell them to make sure those links are helpful to the reader and that’s about it. No tools used.
      I’ve tried other approaches and tools. Didn’t find them helpful in terms of SEO. So we just link in an effort to help the reader and hopefully keep them on our site viewing another page.

      • Thanks for your help Anne.
        Do you have the internal links open in a new window? Or the same one?
        What about external links?

        Thanks

        • Hey JC,
          We never did anything different about the internal links, so it’s just the WordPress default (same window, I believe). With external links, we make sure they open in the same window and tab which means an interstitial ad may show up when the user is leaving the site (a bit more revenue – though not very significant).

          And no worries on the duplicate comments – I deleted those. Sorry about the issue with that!

    • Hi Pax,
      I believe I tried Surfer but I don’t recall much about the specific features. Have you had good experience with that one?

  8. Great insights. Evaluating SERPs manually is the lost art of keyword research. Nothing beats checking (manually) if the current results meet the search intent. I use keyword research tools to get cues but I never base my decisions on keyword tools.

    Thanks Anne. I’m new here. Absolutely appreciate your comprehensive insights.

  9. I enjoyed your AMA and am slowly going through some old posts. Although I first heard about YEYS when Shawna Newman mentioned it on a podcast quite a while ago.

    In the revenue per post column, I am assuming that it is monthly. Is it usually evenly spread out across the whole site? Or do the top 10 posts get much more, and some get pennies?
    Have you ever had a site that gets $3 (revenue per post) or more per day?

    Just wondering….

  10. Anne! This take is spot on. There are lots of sketchy metrics in the digital marketing space. Most tool providers’ KD #s really just look at backlinks vs the keyword itself.

    The relevance and ‘niche authority’ (niche focus, not just how many links) appear to be increasingly important factors to Ggl, and these should also naturally correlate with user intent. So, I definitely agree that you have to look at the competition in the SERP according to these dimensions. And, unfortunately, almost all KD-type metrics depend entirely on how many backlinks the ranking pages have.

    There is one out there that mixes things up a bit: Keywords Everywhere actually incorporates some onpage into it KD score, by looking at keyword frequency on ranking pages. Of course, this is still somewhat simplistic, but at least they are trying to innovate a bit.

    No substitute for actual humans doing actual SERP analysis! Love all your knowledge drops here and on Twitter.. Thank you!

Leave a Reply

Your email address will not be published. Required fields are marked *