SEO experiment: high informational bias niches

In a previous post the fundamental elements of search and intent were introduced. In that post we went through the three major types of intent behind keywords or phrases that people type into Google:

  1. Transactional (want to buy something)
  2. Informational (want to know something)
  3. Navigational (want to go somewhere online)

Today this post will address the topic of doing SEO for high information biased niches. The topic is broad and a comprehensive treatise is impossible. However some basic techniques can be discussed with a case study.

Business, SEO and Transactional Intent

Let’s face it; SEO would not exist were it not for commercial application. So it should come as no surprise that SEO is primarily concerned with transactional intent key words and phrases:

  1. Businesses need customers,
  2. Customers search online for products and services that they need,
  3. Businesses want to be found by customers when they’re searching online to spend money;
  4. Viola, SEO is born and married to transactional intent.

But the story of this post doesn’t end here. Not all businesses are the same or as unequivocally transactional as others. For instance, while an eCommerce store selling shoes fits into that neat, simple model just described, on the other hand a company selling a service, especially a service that is not well known or where education is a fundamental factor, may need more than just knowledge of transactional SEO.

Business, SEO and Informational Intent

On the other extreme there are businesses who never deal with transactional intent for their SEO. A good example of this would be content producers. In most cases a news company like the Sydney Morning Herald would only care for informational SEO, not transactional. For content producers there is an entire realm of informational SEO that won’t be approached in this article.

Now we approach a kind of “middle ground” SEO, which is what this article is about. For businesses who offer a product or service that IS transactional by nature, but for some reason also need to lean across and straddle the line with informational SEO, and not only for pure “content marketing” purposes.

Example and case study: a liquidation company

It’s hard to continue writing generally without going into a specific example, so the one chosen for this article is a liquidation company which is my client: Dissolve and Restructuring Works.

Liquidation companies specialise in taking other companies through the process of liquidation to wind up their affairs. In Australia, often they often offer this as a fixed-price service.

This service is obviously transactional. Liquidating a company is a specific service and it is paid for directly. But here’s something interesting, even beyond what could be classically called “content marketing”, liquidation is a niche where informational SEO can be quite effective.

The rest of this article will outline this specific case study where such a case was identified and SEO work done on it. This is still an ongoing process but the initial results are encouraging, other case studies will be alluded to.

Initial analysis and identification of problem

We had this client approach us and inform us that they used to rank well a few years back and have since lost organic ranking in Google. We knew they had worked with another SEO agency so the default thought that comes to mind is: “they probably spammed backlinks and were hit by Penguin, let’s do A… B… C”.

However, as we began to analyse their website, it’s previous ranking, their niche and competitors, a different story emerged.

Clue #1: Cross-checking site ranking and algorithm update dates

The first clue to their dip in performance not being due to spam is the dates of the dip itself. This graph shows their site ranking over time. As you can see, there is a decline in results starting from around March – May 2013. Cross checking those dates with the algorithm update history, can can see that it might have been a Panda update (it would line up nicely), but the site and layout etc. does not feel like it would be hit by Panda. Besides it looked like a non-major Panda update.

Clue #2: Cross-checking site ranking competitor site ranking

The real revealer was to look at our client’s ranking and compare it with their competitors. Here are some graphs below:

(our client, http://www.dissolve.com.au/)

(Competitor 1: http://www.amosinsolvency.com.au/)

(Competitor 2: http://insolvencyexperts.com.au/)

(Competitor 3: http://insolvencyappointments.com.au. at the time of initial analysis this website was still live, it appears to be gone now)

(Competitor 4: https://insolvencyguardian.com.au/)

From these several competitors, we can notice a general downward trend. Many of them began to lose traffic at about the same time, a time of no Penguin update, possible Panda update.

Clue #3: Terms Dissolve used to rank for, but no longer does

SEMrush is the tool used for all above graphs. In SEMrush one can click on a particular month to see which terms a website ranked organically for at that time. Doing so for February 2013 shows that they used to rank for terms such as:

  1. Liquidation
  2. Insolvent
  3. Liquidators (location)
  4. Voluntary liquidation
  5. Company liquidation

Taking time to review the current SERPS for each of these keywords, we can notice an overall pattern. Given the analysis principles outlined in the fundamentals article, we can notice that these keywords have transformed over time: Google has updated their ranking algorithm to flush out transactional results and promote more basic informational results.

Conclusion: Google’s core has changed in this niche to give way to informational results

Google is constantly analysing the SEPRs they present and the intent behind user queries in an attempt to do a better job at addressing user intent. A fundamental aspect of Google’s functionality is to figure out what users mean when they type stuff in (splitting into those three types of results).

It’s possible that around March – May 2013 they introduced some core algorithm changes to interpret queries in this space (liquidation), and presumably similar areas, to promote information over transaction.

In a sentence, they could be saying:

“We think people who type in ‘liquidation’ probably don’t know much about liquidation, and are more likely to be looking for basic information than for someone to help them liquidate their company.”

In fact, even queries in this niche that seem to be more transactional, are still being flooded with informational results, for instance: company liquidators: (ads removed)

Notice: yes, there were Google My Business results (3X) but for the rest of the SERP, even though “company liquidators” seems clearly transactional, not a single remaining result was purely transactional. They are all informational.

Hence: information bias. Google has clearly prioritised informational results over transactional results, even if the search term seems to be transactional, because of its conclusions about searcher intent in this niche (rightly or wrongly).

(BTW, looking at the PPC ads for this and related SERPS, I’d be surprised if Google didn’t make any money from core changes like this 🙂 )

Solution & Experiment: Definitional landing pages.

Looking at a related search term for our client, “Voluntary Administration”, the SERP looks like this (I’ve highlighted in red everything that shows information bias, as you can see the SERP screams information all over the place):

Something we notice in some of these results is they are plain and simple definitions, for instance a competitor Worrells’ article, which begins “what is voluntary administration?” — simple definition.

So, after some additional research in and out of this niche (which included a brief backwards-engineering of Salesforce and how they rank for “CRM” with their “what is a CRM” page and analysis of Wikipedia sitelinks) we did our experiment.

What is Voluntary Administration? Experiment

We noticed through our research that many of the pages that were ranking well for these type of “definition” searches were “mega-guide” type pages, similar to lengthy Wikipedia articles. So we built one to target “Voluntary Administration”:

What is Voluntary Administration?

Characteristics of this page & experiment:

  1. The first words on the page are: “Voluntary Administration is X”. This was a consistent pattern observed throughout the analysis. Similar to Salesforce’s page which starts “CRM stands for ‘customer relationship management’ and it’s…” This clearly indicates a definition.
  2. We made a category of pages on the site just for definitional pages like this one, /definitions/. Something consistently found in the analysis were information words in the URLs (“factsheet” or “information” or “information centre” or “definition” etc. )
  3. Table of Contents: helps search engines give smaller site links, like in this image: 
  4. Sub-questions: It’s more than just a small page defining a term, but we try to anticipate other related questions that users might have, and we create the page in such a way that they can be skipped to, such as the below. This also helps this page rank for these sub-questions as Wikipedia articles often do (the main article will rank for a sub-question with a sitelink).
    1. How often are Voluntary Administrations successful?
    2. What must be in a DOCA?
    3. What’s a Creditors Trust?
  5. Videos: we’re in the process of building additional videos for certain sub-topics to enhance the quality of information and better compete with existing content out there.

Preliminary results

At the time of writing this article this new page ranks about ~10 in the SERP “Voluntary Administration”. It fluctuates and hovers while Google is processing it. But that was the initial ranking, upon creation and indexation of the content, which is not bad. We’re encouraged by early results and will continue to monitor and work on promoting the page.

Other stuff

There was more work done on the pages with regards to existing content and canonical links, but these won’t be explained in detail here. This article is already long enough!

UPDATE (5-6 months later)

This experiment was a success with flying colours, however we have not claimed the answer box.

 

You may also like...

Leave a Reply

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