201 – Boolean Mastery
201.2: Boolean Preparation Process
Please watch the video, read the lesson below and take a quiz at the end of this lesson (bottom of this page).
So now you know the different components that make up a Boolean query for social listening – which are keywords, operators, and location. Under keywords, there are also different kinds that you need to consider when preparing Booleans such as basic keywords, unique keywords, local keywords, and other variations of related terms. But how do you exactly search for the keywords that you should include in your Booleans?
1. Desk Research
There are various tools that you can use to look for keyword suggestions. When you’re looking for synonyms of keywords, you can use keyword suggestion sites. For basic translations, you can use translation tools like Google Translate. And when you want to see how people are talking about a topic on social media, you search for posts using search bars in social media sites. We’ll show you how these tools can help you prepare Boolean queries for social listening.
a. Searching for Keywords Using Keyword Suggestion Sites
The easiest thing that you must do when you start creating Booleans is searching for related terms with the use of keyword suggestion sites. There are many sites that provide this information, but these three sites are good to start with:
For example, you’re creating a Boolean to track mentions about “Valentine’s Day”. There may be terms and hashtags that you’re familiar with such as the following:
- “heart’s day”
- “hearts day”
We’ve listed down eleven keywords, but the best practice for creating Booleans for social listening is to be as comprehensive as possible in your query. Make sure that you have covered all terms that are relevant to the topic. You could use Power Thesaurus to find other terms that you don’t have yet by searching “valentine’s day”:
What we need to add to our initial list of keywords are those terms that cannot be captured using the keywords that we already have. We don’t need to add terms that have “valentine” or “valentines” or “valentine’s” because those will already be covered by the first three terms on our list. However, we need to add the following because we don’t have these on our list yet:
- February fourteenth
- Fourteenth of February
- V day
We can also add other variations of these terms like:
- February 14th
- Feb 14th
- 14th of February
- 14th of Feb
- 14th February
- 14th Feb
After this, you can go open up other relevant websites such as Urban Thesaurus:
There are lots of terms on this list, but we don’t need to include all of them. We still need to qualify those that are relevant for the topic and the locations that we are tracking. For example, we want to track Valentine’s Day mentions only from Southeast Asia, can then list down only those terms that will be relevant to this region that we don’t have yet:
- Singles awareness day
- Single awareness day
- Valentime’s day
- Galentine’s day
- Forever alone day
We can then add the variations of these terms, such as hashtags:
After this, you can continue checking other search results until you think you have exhausted all relevant terms. But for the purpose of our course, we will just show how you can do the process for two search results.
b. Doing Basic Translations Using Google Translate
Another tool that you can use to search for other variations of terms is a translation tool like Google Translate. Since we’re searching mentions from Southeast Asia, we also need translations from languages of that region. As an example, we’ll translate “Valentine’s Day” to Thai, Bahasa, and Vietnamese. After running that search in Google Translate, we got these translations:
- วันวาเลนไทน์ (Thai)
- Hari Valentine (Bahasa)
- Ngày Valentine (Vietnamese)
Since Google Translate is not 100%, you should also copy and paste the translations you get in Google Translate to Google search engine to see if relevant results come up. For example, the Vietnamese translation we got from Google Translate was “ngày lễ tình nhân”, but when we pasted that on Google, we got this result:
The translated term in Wikipedia is likely more accurate than the one from Google Translate because this site translates articles to cater to people from different regions. So we’ll use this translation instead.
Just a note though that Google Translate is best used for basic translations only. It will still be best to consult with a local translator/analyst to ensure that you have all slang, colloquial words, and country-specific hashtags for a particular market.
So taking into account all of the keywords that we have, for now, our Boolean can be something like this:
Valentine OR “Valentines” OR “Valentine’s” OR “#valentine” OR “#valentines” OR Vday OR “#vday” OR “heart’s day” OR “hearts day” OR Heartsday OR “#heartsday” OR “February fourteenth” OR “Fourteenth of February” OR “V day” OR “V-day” OR “February 14th” OR “Feb 14th” OR “14th of February” OR “14th of Feb” OR “14th February” OR “14th Feb” OR “Singles awareness day” OR “Single awareness day” OR “Valentime’s day” OR “Galentine’s day” OR “Forever alone day” OR “#SinglesAwarenessDay” OR “#SingleAwarenessDay” OR “#ValentimesDay” OR “#GalentinesDay” OR “#ForeverAloneDay” OR “วันวาเลนไทน์” OR “#วันวาเลนไทน์” OR “#HariValentine” OR “#NgàyValentine”
C. Social Mentions Scan
Another way to look for variations of terms is by doing a quick scan of social media posts. This is done to see what terms or hashtags people use to discuss the topic that you want to track. Social media sites like Instagram, Twitter, Facebook, and YouTube have search bars where you can type keywords related to the topic and see what terms people are using in their captions.
For example, on Instagram, you can search for posts about a certain topic with the use of hashtags. So for “Valentine’s Day”, we can search for posts with the hashtag “#valentinesday”:
After this, you can click on each post and see if there are any terms/hashtags people use that aren’t in your Boolean yet. We saw these posts with different hashtags:
These posts have the hashtags “#happyvalentinesday”, “#valentinesday2021”, “#valentinescookies”, and “#valentinesdaycookies”. We can add these to our existing Boolean so now we have:
Valentine OR “Valentines” OR “Valentine’s” OR “#valentine” OR “#valentines” OR Vday OR “#vday” OR “heart’s day” OR “hearts day” OR Heartsday OR “#heartsday” OR “February fourteenth” OR “Fourteenth of February” OR “V day” OR “V-day” OR “February 14th” OR “Feb 14th” OR “14th of February” OR “14th of Feb” OR “14th February” OR “14th Feb” OR “Singles awareness day” OR “Single awareness day” OR “Valentime’s day” OR “Galentine’s day” OR “Forever alone day” OR “#SinglesAwarenessDay” OR “#SingleAwarenessDay” OR “#ValentimesDay” OR “#GalentinesDay” OR “#ForeverAloneDay” OR “วันวาเลนไทน์” OR “#วันวาเลนไทน์” OR “#HariValentine” OR “#NgàyValentine” OR “#happyvalentinesday” OR “#valentinesday2021” OR “#valentinescookies” OR “#valentinesdaycookies”
After this, you can do the same process in other social media platforms like Twitter, Facebook, and YouTube to ensure that you’re including all relevant terms in your tracking.
2. Google Search Trends
Another way of getting ideas for keywords to include in your Booleans is to find out what keywords people use to search for the topic in internet browsers such as Google Search. Google has Google Trends which is a free site you can use for this purpose.
This tool is useful to get an overview of Google Search trends about a particular topic/keyword.
Access it by going to this URL trends.google.com.
Then, type the main keyword for the topic you’re tracking on the search bar. For example, let’s input “Valentine’s Day”. It will give you a few options, but you can just select the “Search Term” one:
Once it has loaded, go all the way to the bottom of the page to view the “Related Queries” about the topic. You can choose to sort the keywords either by their growth rate (Rising) or Search volume (Top). For example, in the below result, we can see that people search for “Squishmallow” which is a type of plush stuffed toy that they give as gifts to their valentines.
Tip: You can also use this tool to uncover top themes that people are interested in about a certain topic.
We have shown what it looks like for results from the United States, but you can go ahead and change the location and duration that you want to see from the drop down menu:
B.Writing the Boolean
- Tools of the Trade: Text Editors
In writing Booleans, there are a lot of things that you need to ensure such as the proper use of operators, that the quotation marks or parentheses that you’re using are matching, etc. These can be hard to track if you’re typing your Boolean directly into the social listening tool. Preparing your Booleans in text editor software before inputting them into the tool (like Sublime Text) makes it easier for you to ensure that there are no errors in your Boolean query.
There are many text editor tools that you can use, but we’re going to show you how to use Sublime Text. You can download it for free in this link.
When you open a file, it will look something like this. At the bottom right of the window, you will see the text “Plain Text”, click on it and select “SQL”:
Then, you can go ahead and write your Boolean query on the file. I have pasted our Valentine’s Day query and it shows me this:
If you can see, it color codes terms into white if they don’t have quotation marks, yellow if they have quotation marks, and red if they are operators. But, the terms towards the end of the Boolean are showing as grey, which means there are errors with them. Since I used quotation marks with those terms, the error could be because the type of quotation marks that I used are wrong. So let’s try changing them:
Now they’re showing up correctly because I have used proper quotation marks.
It’s also useful if you have parentheses or quotation marks in your Boolean because it shows you if they are paired correctly, so you can easily spot if you have a missing mark. If you put the cursor beside a certain parenthesis, it will underline which parenthesis it’s paired with:
Make it a practice to edit your Booleans in text editors first before inputting them in your social listening tool to avoid getting frustrated trying to diagnose errors. This is also more helpful when you’re creating more complicated Booleans where you’re using so many operators and keywords.
2.Creating Display Topics in 20/Twenty
You can create a Display Topic using two methods:
Creating a Topic from the Search Bar
- Just access the Search Bar either from the home page:
- Or from the “Conversations” page:
After clicking on “Save as Topic”, you will be directed to the Query Builder in the Admin Panel where you can create your Display Topic.
Accessing the Admin Panel
- Go to the Admin Panel by clicking on the “Admin Settings” button on the 20/Twenty homepage
2.Once you’re in the Admin Panel, click on “New Topic”
3.The Query Builder will open up, follow the below steps to create a Display Topic
- Identifying Irrelevant Keywords
Due to the nature of how people use terms in social media and also because data capturing in social listening is purely based on keywords, it is almost certain that you will get irrelevant mentions with every Boolean query that you create. There are three main types of keywords that will capture noise in social listening:
Any mention that is not related to the keywords/topics that you’re tracking are considered spam.
One example of this are online sellers using different brand hashtags to promote their products, even though their product is not of those brands. Say you’re tracking the brand “Estee Lauder” and add their hashtag “#esteelauder” in your query. The post below would have been captured because it has the hashtag “#esteelauder” even though it is an image of a Maybelline Lip Gloss. To get attention for their posts, this online seller also used other brand hashtags like “#esteelauder”, “#hudabeauty”, or “#tarte” so that their post will appear when users search for these hashtags on Instagram.
The definition of spam can also be different depending on the social listening use case. For example, you want to know consumer feedback about the brand Estee Lauder. Online selling mentions like below may not be relevant for this objective because even though it is about an Estee Lauder product, it’s not about consumer perceptions about the brand.
B. Terms that may capture irrelevant mentions from another market
For example, you’re tracking mentions of the skincare brand “La Mer”, it might capture irrelevant mentions in French because it is a term that means “the sea”.
C. Non-brand related mentions for generic-sounding brand names
It is tougher to create Booleans for brand names that are common words like “Apple”, “Shell”, or “Subway” because simply putting the main brand names on your query will return mentions that aren’t related to these brands.
- Applying Exclusions
As we have seen from the previous section, there are various types of keywords that could cause capturing of irrelevant mentions. There are two main ways of preventing pulling irrelevant mentions – one is applying exclusions and the other is tightening Boolean queries.
Applying exclusions is best done when you have specific keywords in your Boolean and are getting mostly relevant mentions, but you just want to exclude a few recurring irrelevant mentions.
In our example of tracking mentions of “Estee Lauder”, since we only want to exclude those non-brand related posts as well as online selling conversations, we can just exclude keywords like “online selling” or “online shopping” and other related terms in our Boolean. As we have learned in the “Key Components of a Boolean” section of the course, the way to exclude keywords in a Boolean is by adding them after the operator NOT. So maybe our Boolean can be something like this:
(“Estee Lauder” OR “#esteelauder” OR “@esteelauder”) NOT (“online selling” OR “online shopping” OR “online shop” OR “#onlineshop” OR “#discount” OR “discount code” OR “free shipping” OR “free delivery”)
- Tightening Boolean Queries
Adding exclusions may work for Booleans where you have very specific keywords, but when you’re tracking generic-sounding terms or brand names, it would be impossible for you to identify all irrelevant keywords that you should exclude. In those cases, it’s better to tighten your Boolean query by adding AND keywords.
For example, you’re tracking mentions of the sandwich brand “Subway”. If you try to go the route of listing down exclusion keywords, then you wouldn’t be able to finish as there are countless keywords associated with the word “subway” that are not related to the sandwich brand. These words include transportation/transit related terms or actual names of subway lines, etc. So in this case, it’s just better to add AND keywords related to the sandwich brand. Our Boolean query can be something like this:
(“Subway” OR “#subway” OR “@subway”) AND (“sandwich” OR “sandwiches” OR “subs” OR “salad” OR “salads” OR healthy OR “chicken bacon ranch” OR “chicken teriyaki” OR “cold cut trio” OR “meatball marinara” OR “chicken ham”… plus other sandwich/wrap/salad variants that they offer)
For both methods (applying exclusions and tightening Boolean queries), the only way to list down terms to add/exclude in your Boolean is to scan through the data. It’s a trial-and-error method whereby you would really need to look through your data so that you can improve your Boolean. And Boolean optimization is not a one-time task – it’s an exercise that you need to regularly check on as you track your topics. People’s usage of social media change every day and the way in which they use terms online will just continue evolving and changing in terms of context. So you would need to be updated on new spam keywords that might come up.