Every day, across the web, searchers create new opportunities for marketers to conduct strategic keyword research. How do they do it? By searching for things they are almost entirely clueless about.
The web is filled with great articles on keyword research. There are keyword research articles which are niche specific, strategy-oriented, intended for ‘dummies,’ and intended for ‘thought leaders.’ There are articles telling you how to capture broad match traffic (where most of the traffic exists, but where the competition is fiercest), as well as articles which focus on long tail search (longer search terms for which traffic is less but easier to rank for.)
There’s one type of Google search query, though, about which very little has been written. And, yet, this query type makes up a large percentage of all search engine traffic. Got any idea what we’re talking about? (Go ahead, use the paragraph break to guess what that type of query might be…)
The answer, actually, is the ‘never-been-used’ search query. Every day, 15% of all Google searches have never been used before. Ever. That’s over half a billion searches, every single day.
“Every single day 15 per cent of the questions people ask of Google are questions we’ve never seen before,” John Wiley, Google
So how do you develop a keyword strategy which accounts for searches which haven’t been made yet? The more typical tricks in our bag – competitive research or research tools such as this or this – all fall short in this circumstance, because they’re all based on historical data. In order to capture the search volume we’re looking for, we’ll need to use a more predictive method.
Predicting What Searchers Don’t Already Know
Before we answer that question, though, it might first be useful to explore the factors accounting for why so many unique searches occur. In doing research for this post, we found that there is very little research out there answering this question with scientific precision. Google is mum. However, some experts think have drawn some general observations. Most types of ‘new tail’ searches are usually expressions of ignorance on the part of their users. Namely, most of these searches represent queries issued in the following situations:
New Products or Services. Someone sees a new tech device. They’re not quite sure what it does, or maybe they saw an ad for it, but can’t remember its name. Absent any amount of precision, they begin to fish around for descriptive language, which loosely describes what they’re looking for. Take a look at the following picture, for instance.
If all you knew about the image above was that it was an underwater camera (spying a couple of humpbacks), you might type in any number of keywords, such as ‘device for taking underwater photography’ or ‘self-powered remote submarine for fish finding.’ (It does all these things.) The actual device in question is actually the PowerVision PowerRay Underwater Robot. So, then, if you were writing SEO copy for PowerVision, your task would be to anticipate all the potential keywords that your ignorant, (but interested) potential buyers might use.
New (to the searcher) objects, things or creatures. Ignorance is also the source of the second reason most often given for brand new Google searches. For instance, take this interesting creature below:
If you had come across it somewhere, and didn’t know its name, you might be at a loss for words. Descriptions like, “bug-eyed monkey” might come to mind. Or “satanic looking were-bat.” And those made-up-in-the-moment descriptions very well might be new search queries. (Actually, turns out the creature above is called a tarsier and, almost certainly, searchers have already used the expression “bug-eyed monkey” to describe it.) They hide on tree branches, by the way, and use those huge eyes to scan for delicious insects. When they find them, the pounce and eat dinner.
Developing Keyword Discovery Methods for New Tail Searches
Since we have no historical data to develop keyword lists from, probably the most straightforward way to begin developing keyword lists is to conduct some market research. We want to learn how our search pool would describe our products and services, without actually knowing their names. If we had deep pockets, of course we could hire somebody like these guys. Lucky for us, though, there are less expensive and quicker ways to get actionable insights.
Google Surveys is one such method. (There are actually several. Survey Monkey is another.) For only pennies a response, these tools allow us to ask a predefined group of respondents almost anything. Using our furry friend above, for instance, we can make a survey which includes only an image and a question. Our question might read, “If you were conducting a Google search, what words would you use to describe the animal shown in the picture?” Google Surveys is not only cheap but easy to set up. For this type of exercise, you’ll need to specify the “image + open ended” survey model. This will allow our respondents to answer our question in their own way.
Once we have our survey results, we will next compare them to two factors. First, we will compare our new set of keywords to actual keyword metrics. In other words, what terms are searchers currently using to find our product? (If you don’t know how to do this, I’d recommend your getting familiar with either this tool, or this tool, or this tool.) Secondly, we’ll want to look at the existing copywriting on all these pages, (and the keywords we originally used) and see how closely those keywords align with actual search keywords.
At this point, we’ve got three buckets of data:
- Our new list of keywords, derived from market research,
- The keyword research data, which we’ve collected from third party tools, which tell us actual keywords which searchers are using to discover our content, and
- The keywords our content is currently written for.
From here on, our task can either be very straightforward or more complicated, depending on our goals and resources. The simplest method involves just three steps:
- Compare our new list of survey-derived keywords to the actual keywords searchers are already using,
- Eliminate any duplicates, and
- Add all the keywords deemed ‘high value’ in our survey data to all the product’s current web content. (In some instances, our lists might be long enough to warrant the creation of new product or landing pages, to support the expanded keyword list.)
However, the keyword process is ongoing and iterative. For instance, in our steps above – when we actually collect the keywords searchers are currently using to find our content – we can use that opportunity to clean up our existing content. Our research, for instance, might find that searchers are actually using different keywords to get to our content than we originally developed. If so, we need to change our content to bring it into alignment with those keywords. Then, we will move on to the task of adding our new crowdsourced keywords to our content and landing pages.
It should also be noted that our original list of keywords from our surveys can also be expanded. Suppose we developed a list of 25 keywords, each of which contained, on average, 4 words. Each one of these 25 strings can be evaluated for potential synonyms and related terms, using a tool like UberSuggest or the amusingly named Keyword Shitter. There are a wealth of other tools which can be used to expand keyword lists; these are only a couple.
Wrapping Things Up
As I mentioned at the beginning of this post, there are half a billion searches every day which fall under the radar of most search professionals. In an era of extreme competition, planning for those searches gives marketers a much-needed edge over their rivals.