By now you have probably heard of Answer the Public (ATP). It is a well-known, free tool that marketers use to discover the questions people are asking on Google related to their niche. People love it for its ease of use.
It is a useful tool, but it is not the best tool.
Answer the Public works by scraping Google Suggest aka Autocomplete. This is the feature on Google that suggests different queries for you to try as you are typing into the search bar. There are a couple problems with this.
1. Answer the Public does not provide search volume data
If you take the data from ATP at face value, you might assume that these are questions that people around the world are regularly searching. That is not always the case. To see what kind of search volume some of the suggested queries had, we entered the term “martial arts” into ATP and then imported the results into Google’s Keyword Planner to get average monthly search volume for each question.
I’m deliberately using keyword planner because it shows volume based on broad match, meaning close synonyms for your search are included in the monthly volume estimate. This normally gives inflated numbers, but in this case I expect to get a lot of long tail searches, for which I prefer the broad match data set.
Of the 700 keywords that ATP gave me, 90 had no data. This means either no one is searching for it, or Google does not have enough data to display monthly volume. Either way, we can make the safe assumption that nearly 13% do not have significant search volume behind it.
Some queries that seem similar also vary in search volume greatly. For example, “martial arts vs weight lifting” only has on average 10 searches/month, while “martial arts for weight loss” averages 210 searches/month. You cannot get these insights from solely using ATP.
2. It does not work well for local searches
You are going to get a lot of phrases with local intent. Using kickboxing classes as an example, I get a ton of “kickboxing classes + geography” type results. Because we already expect this to be a local result, this is not useful for us, and is a common problem with ATP when trying to research highly local searches.
Had enough? Time to show you three ways to do better keyword research without ATP.
Mine Quora using Advanced Search Operators
This is one of my favorite tactics. Quora is a Q&A site, but at its core it is a search engine. If you pay close attention, you will notice that the URLs and title tags are structured to be identical to the question being asked. This is great, because we can take advantage of that in our research.
Go to Google and try these search operators:
site:quora.com intitle:”kickboxing classes”
site:quora.com intext:”kickboxing classes” -intitle:”kickboxing classes”
The first search operator gives you explicit questions about kickboxing, while the second one will include some broader topics that you may miss with the first operator. For example, the second operator gave me a result for “How to combat the fears of joining a group fitness class”.
You can also see the number of views the answers to these questions get which can be a better indicator of the popularity of a question than a broad match volume estimate from Google. I recommend setting Google to view 100 results by default when you’re doing this so you can quickly scroll and select the best questions. Organizing the URLs in excel is a fast way to group questions so you can identify popular questions easily.
Use Moz Pro like a Pro
I use Moz’s Keyword Explorer (KWE) frequently. One of the nice things about KWE is that you can get a lot of insight very quickly by playing with the different filters. Let’s use “martial arts” for our example. Follow these steps:
1. Type martial arts in KWE
2. Go to Keyword Suggestions
3. Set your first filter to “are questions”
4. Group keywords by “low lexical similarity”
5. Set volume filter to “11” and above
You will not always get a ton of results using this method because Moz uses clickstream data to provide exact match search volumes, and I like to filter out the “no data” questions. Sometimes if you get nothing worthwhile it pays to remove the volume filter. I have found that grouping the keywords by lexical similarity I can quickly identify content opportunities, like the one below.
Here’s a fun fact about the lexical similarity filter. If you group by low lexical similarity, you will get larger groups with more loosely related terms while with high lexical similarity, the terms are more closely related so you get smaller groups and it is a bit more accurate. The ideal keyword group often has under ten keywords in it, because they seem to be closely related enough that you could target all of them with one piece of content, and it is manageable to research. A group of 117 keywords is not.
So play with it a little. Switch between each level of similarity, jot down each topic idea you find, and then go create some awesome content. Why is this better than ATP? It is a faster method for choosing what topic to write about and it has volume estimates to back up your content pitch to the team. It also does a good job of avoiding the localized keywords that clutter your ATP results.
The main reason I am featuring Moz here is that they use clickstream data and give exact match estimates. Many SEOs prefer or insist on exact match estimates, and there is merit to their argument. It tells them each query they need to track and optimize for, which is a more focused approach. An SEO that uses broad match would say that they are targeting an array of queries and tracking a specific one is too narrow. Neither is wrong, just different.
Using Keyword Planner for Long Tail Research
There is a tendency for search volume to drop off the longer the search becomes. This is because there are just so many ways to say something the longer the phrase gets. So when you are doing long tail keyword research, or researching questions, broad match is sometimes more accurate. So for this we can use Google’s Keyword Planner, even though the way they match keywords is sometimes questionable. They once matched a client’s brand name with a Spanish rice dish. So it is not perfect, but some volume data with inaccuracies is better than none.
Start by selecting your head term and entering it into Keyword Planner. Change the keyword options so “Only show closely related ideas” is the only one switched on.
Select the blue “Get ideas” button and download the excel file. Highlight your first row, select data in the ribbon and select filter. Many questions start with a “W” (who, what, where, why, which, will, etc.). You can sort the keyword column alphabetically and easily find the questions. Plenty of these questions do not have search volume, so select the avg. monthly searches filter, scroll down, and uncheck “blank”. This ensures you are only looking at questions that have search volume associated with it. Make sure you also look at the “how” and “can” questions. You now have many questions with clear volume estimates.
There you are, three ways to do better long tail keyword research and find the questions people are actually asking. To be clear, I like Answer the Public well enough. It does something automagically that many SEOs used to do manually. But the lack of search volume data and its common usage among content marketers puts it near the bottom of my stack.
Remember as you do your research that volume estimates are based on historical data and they are not a promise or guarantee. It is an estimate, because even if you could see true volume for the prior year, it is not proof that there will be the same search volume next year. Searches for VHS dropped year after year until they evened out around 2009, while searches for vinyl records have been increasing year over year since then. Emerging trends, dying technologies, and current events all impact yearly search volumes and throw off estimates. Use your head, do your research, and pick your keywords wisely.
Thanks for reading! Contact us using the form below if you are interested in a complimentary consultation.