Keyword Relevance and The Stories They Could Tell Us
The painstaking work by SEO’s performing keyword research has intrigued me ever since Google came along to kick its competition out of the pool, creating competitive war games for top positions in search engine results.
Unpredictable Search Engine Playbook Changes
Have you ever seen a football team’s playbook? Each page has a plan, with arrows, numbers, rules and code names to be memorized by the team. Headphone wearing, yelling Coaches call out the secret names and if all goes according to plan, everyone will remember what to do to get the expected result. If not, there could be a flag on the play and swearing, screaming fans in the stands.
It is this search engine playbook that SEO’s grabble with. It’s reviewed by the marketing team, memorized and practiced and just at the precise moment a web page scores a goal, an algorithm is updated and a new set of rules and plays have to be figured out. After all, the goal of marketing is to win and the prize is revenue.
Google assigns its search engine algorithms names, often after animals that begin with the letter “P”. Every SEO worth their paycheck tracks whatever Google changes or invents next, but what they do less and less of is studying the “P” word I like to think about.
People balk at change. This is one of the most basic of user interface design rules ever taught and Google’s weapon of choice against SEO’s. In Google’s defense, it’s not about “getting pages ranked high in search engines”. Their business goal is generating enough revenue to build driver-less cars, buy the Internet and know the blood type and DNA of every human being on the planet.
It didn’t start out that way. Search engine optimization began in the age of innocence, where new tools were tested, results shared, and relationships were formed out of knowledge sharing, whether civil or competitive.
When I was a fully invested SEO in the mid-1990's, our main tool was WebPosition Gold for keyword research and a few free site submission websites used to get into, if you were to believe the scam artists, 60,000 directories out there. (In those days, challenging fake submission services could bring you threats of death or lawsuits if you asked to see proof of their lists. Ask me how I know that.) New search engines were always popping up, each one with its own set of rules to get in, rank and live there forever.
· Alta Vista
· Yahoo! (Both engine and directory)
· Northern Light
· MSN (Ms. Dewey)
· Others specific by country.
BruceClay Inc. had a colorful diagram of search engines we followed, and my first official website had a text version that described the same search engine playbook rules. I kept track of every submission link, rules for rank and any fees.
There were web rings and link farms and text link miracle tools, black hat cloaking and company bribery behind the scenes to rank higher than the encroaching competition. There were publications written on the mathematical equations behind Google’s algorithm, and a few gigantic books written by such brilliant minds as Mike Grehan, who crawled underneath the belly of internet search engines to teach it to the world.
Eventually nothing mattered except Google, Bing and a few stubborn directories like Ask, DMOZ and Yahoo! You would think the work load would have decreased when the bulk of link purchases, content writing, keyword tracking, PPC, advertising, black hat secret sauces and corporate bribery focused on just two search engines and a few struggling directories.
However, I feel it’s because people are so unique and unpredictable that you, my SEO friends, will keep your jobs. For starters, most everybody chooses to use Google first and it’s not done messing with you yet.
The Uniqueness of Search Query Relevance by People
Researchers have evaluated the search engine retrieval process going back to the 1990’s. There are studies that evaluate the studies that evaluate search engine retrieval processes and to be blunt, they gave up trying to sample SERPS because, well, you try sampling data from the global population and see how much fun that is.
One paper, Evaluating the Retrieval Effectiveness of Web Search Engines Using a Representative Query Sample, compared Google and Bing with 1000 sample search queries for two types of search patterns, informational and navigational. Informational is what we do the most of. If we search for one or two words from Google, it uses ESP to figure out what we want to find. (I’m kidding.)
I decided to compare Google and Bing on a Sunday when the rest of my area was watching football. The first example is a one-word informational search to see if a search engine could read my mind.
Figure 1 — One word, and I wasn’t searching for the bird or the band. How did Google know I was asking about the football team?
Figure 2 — Bing, on the other hand, is a chatterbox.
We search smarter with voice activated search and we’re getting better at researching, which is why asking questions to machines is now normal. For Google, ever eager to please you, providing the answer through Knowledge Panels is how they show off how smart they are, but for web site owners, this presents a slight concern since there is no reason for anyone to visit websites to get answers if Google is hogging up everything.
Figure 3 — Sorry Eagles football franchise website all the way down there. Google beat you.
Figure 4 — Bing has less ego and though it provides the answers to the question too, it also presents us with additional resources, logos, social links and the coveted link to the actual franchise website with the words, “Official”. Cool beans.
Figure 5 — If we continue to scroll, the informational search switches to navigational search by presenting possible pages to navigate to. Google wants to control that too by limiting what we will see.
Figure 6 — Google generously directs searchers to “Shopping”. As a user experience person, I immediately visualize someone who had tickets for the game but couldn’t go and some kind, caring person decides to purchase a team shirt to cheer them up with. This is more fun than “Maps” because if you live near Philadelphia like I do, you already know it will take 3 hours to go 30 miles into the city.
Figure 7 — Bing, not so much into the shopping angle.
Figure 8 — However, if you do decide to shop, Google needs a tad more information to help you buy that shirt. So much for “retrieval effectiveness” for searchers on a mission that goes beyond the realm of the immediate search query results.
The third way we search is transactional and this is avoided by researchers because they need to buy stuff. But I was curious and went shopping for my obsession.
Both Google and Bing have interesting ways of helping us make purchases.
Figure 9 — Google’s first priority is to get the money from advertisers up front, so searchers are gifted with whatever website had the money to cough up for the coveted top spots. After that the big shots appear at the top, may the best SEO win. In this example, the top search results paid to be there and the images to the right are cousins to the left column, with very little variety.
Figure 10 — Bing won this round because it presents search results for human’s who want to make informed decisions. The images are in the left column and show a variety of products, even though they are technically paying to be there rather than one item by one manufacturer shown in different colors and angles. Bing doesn’t blast “Hey these are ads!” with the little green icon like Google does, which is perfect for sight impaired, contrast loving people who won’t ever see the little light gray “Ad” and believe those websites genuinely ranked there. As a bonus for Bing, they present alternative searches in the right column.
Bing’s display just feels a bit less desperate from Google’s.
For most of us, Google always win the battle of the search engines, but Bing is the default search engine for Amazon’s Kindle, and if you are spy or don’t want Google to know about your secret life as an erotic writer in need of specific research, there is Duck Duck Go. Not that I would know anything about that.
Do you agonize over the user experience of people figuring out how to get cell phone voice search assistants to speak with a British accent? If we’re going to perform voice searches, we may as well get a sexy voice delivering the results to us. SEO’s focus on the keywords, but I want to know how searchers react to search results spoken by the sexy husky voiced woman I call, “the girlfriend” or “the butler, Charles”. In a way it’s a shame the general public has no idea that SEO’s are managing keyword bundles and chasing Possum, plus wrapping their heads around AMP, Schema, competitive research and AI on their behalf.
Ignoring site performance in search engines or within the user interface translates to revenue loss, low conversion rates, reputation management problems, search engine indexing issues, page abandonment and so much more.
Keywords Are Mysterious Creatures
For anyone who can’t spell or read with ease, a search engine is only as helpful as its auto complete. There is a study on Google, autocomplete functions and dyslexia called, Do Autocomplete Functions Reduce the Impact of Dyslexia on Information-Searching Behavior? SEO’s are always thinking of plurals and related terms for keyword research. They plan for misspelled words that they can predict. To actually know for sure who the users are, where they are and what may be causing the misspelled words is what I think about.
Drunk searching is one. And so is dementia, poor education, hand tremors, poor eyesight, high stress, emotional upset and dyslexia. Below is a comparison between Google and Bing of the same search with a phonetic query.
Figures 11 & 12 — Bing not only has no ESP powers but is thoroughly confused about what the searcher could possibly be thinking. Google is totally Zen.
Figures 13 and 14 — Google has figured how to perform remote viewing and knows what I’m doing, while Bing is completely out of touch with reality.
Dyslexia is said to affect 3–10% of “any” population points out the study. The outcome of the study on dyslexia focused on the human behavior responses of the searcher rather than how Google’s autocomplete performed.
“Participants with dyslexia made more misspellings and looked less at the screen and the auto complete suggestions lists while entering the queries. The results indicate that although the autocomplete function supported the participants in the search process, a more extensive use of the autocomplete would have reduced misspellings.”
Which brings me to why I encourage keyword research in the wild. Keyword research and competitive analysis may be directly tied to someone who simply wants to find something in a universe of human possibilities and distractions that math and bots can’t grasp.
Keyword research in real world situations is a keyword user experience with a story to tell.
Berget, G. and Sandnes, F. E. (2016), Do autocomplete functions reduce the impact of dyslexia on information-searching behavior? The case of Google. J Assn Inf Sci Tec, 67: 2320–2328. doi:10.1002/asi.23572