Adaptive Listening Walkthrough

Adaptive Listening Walkthrough from MutualMind on Vimeo.


Video Transcript:

Hi, my name is Eric and I’m excited to take you on this tour of the latest set of features available from MutualMind, called Adaptive Listening.


We created Adaptive Listening because we saw that a lot of our customers had created simple decision rules to handle the day to day of social business, so they could jump on opportunities or crises as they arose. With Adaptive Listening, we let you automate those rules so your team can focus on the human side of Social Business, and actually engage with other people!


Adaptive Listening starts with creating a rule. To create a rule, go to your Update Campaign page, scroll down to the Adaptive Listening section, and click “Add Adaptive Rule.”


You’re now inside the Rules Editor. Let’s create a rule to identify Influencers. We’ll name our rule “Influencer Alert” and click inside the rule text box. You’ll notice a couple things happen: first, the progress bar fills to indicate that we’ve started building the clauses of our rule. Keep an eye on it as we go through the process - you’ll note how it guides you through the steps and tells you when you’re done. Second, you’ll see that the Editor has brought up a set of suggestions for what I can add to my rule. You’ll notice this auto-complete function showing up throughout the creation process.


To begin our rule, let’s select Klout Score. We can then type in the number, or use the slider to select the score we want. Let’s set it at greater than 65 and hit spacebar to continue our rule. You’ll see the auto-complete options box show up again indicating what I can do next. Your rule can have an unlimited number of conditions that have to be met - we call these Dimensions.


We want to add another Dimension, so let’s select “and.” Let me introduce you to some of the other Dimensions we have available:


Category lets you select content that matches the “Brand” or “Competition” groups of keywords within your campaign.


Keyword lets you match content to a specific keyword within your campaign. The auto-complete list pulls the names of each keyword created, so it’s important to name your keywords something that best describes what content it’s capturing.


Publisher lets you apply a rule to a specific type of social network. You’ll note here that, as in our Content Browser, we’ve broken out Instagram and Reddit as separate Publisher types.


Sentiment lets you segment content that has been identified by our natural language processing engine as “positive,” “neutral,” or “negative.”


Display Name lets you choose content from users based on the name they’ve chosen to be shown on their account. This can be anything the user wants, including “Willy Wonka.” If all your brand employee social accounts use a similar naming structure, this can be a useful tool for finding them.


Followers Count and Friends Count let you select content from users on Twitter who are followed by or who follow a set number of other Twitter users.


Profile Description refers to the short blurb most social networks allow users to include with their user account. In the case of Twitter, this is the Twitter Bio, and it can be extremely useful for identifying accounts that are employees of a company, or ones that are the official outlets for brands.


Author Username lets you select content by the actual screen name of their social account. This can let you target content from a single user, or from a list of users - more on that part coming up later.


Finally, Content lets you input a specific word or phrase you’d like to identify within posts, or check a post to see if it matches a preconfigured boolean query. More on that later as well.



Ok, let’s say that this completes all the Dimensions we want in our rule, and we’re ready to have it trigger an action. I’ll select “then,” and we can see that we have five potential types of actions that this rule can trigger.


Create Task automatically opens a ticket within the Work section of the platform, and assigns it to the user you designate. As long as they have their notifications turned on, they’ll receive an email as soon as the task is opened.


Reclassify As lets you alter the sentiment score for content using the rule you’ve created. Combining this with our Content dimension can really help with brands and proper names with terms that normally throw off sentiment scoring engines.


Tag With lets you apply a custom tag to content that matches this rule. This is incredibly powerful, as tags can enable you to rapidly segment and analyze content across many of the same metrics as regular campaign keywords.


Send Email lets you send matching content to any user within the campaign, without creating a task for them. Each email contains a link to the original post, as well as a quick analysis of publishers included. Emails are sent at a maximum rate of one per 10 minutes, so we won’t flood your inbox.


And finally, Add to List lets you curate a list of Twitter Handles publishing content that matches this rule using our Match Lists feature.




So let’s take a look at two brand new tools we’re launching as a part of Adaptive Listening: Match Lists, and Match Expressions.


Match Expressions let you target any content within your campaign as a part of an Adaptive Listening Rule using an unlimited Boolean expression. And, when we say unlimited, we mean it! Here’s an example I created using Wikipedia’s listing of color names - and you can see it’s pretty extensive!

Match Expressions can be used to analyze text anywhere you might find it. For example, here’s a nifty expression searching for a specific set of job titles, and I can use that in an Adaptive Listening Rule any tweets that might be coming from those titles.


Our other new feature is called Match Lists. This tool lets you collect a list of Twitter handles and act on them within the Adaptive Listening Engine. When you create a list, you have two options: Human Curated and Machine Learned.


Human Curated lists are ones you create, such as this one I’ve entered here. At any time you can add new Twitter handles, or “mute” specific handles, so they can’t be acted upon by any Adaptive Listening Rules.


The other type of Match List is Machine Learned. This type of list basically works the opposite way: you can use an Adaptive Listening rule to segment out content using a set of Dimensions, then add any Twitter handles that create matching content to your Match List.


This is incredibly powerful for many use cases, one of which being ad segmentation. You can use an Adaptive Listening rule such as this one to find an audience meeting a very narrow set of criteria, put them into a Match List, then use that list to create a custom audience for Twitter ad targeting.


It’s also important to note that you can use rules to create Match Lists and then trigger off of them. For example, in these two rules the first adds a Twitter handle to a list if one of our customer service reps with these initials responds to someone from our brand handle. The second rule automatically creates a task for the same representative if any of the people they responded to Tweets about our brand again. You’ll note here that Adaptive Listening rules are acted on from first to last, so use this order of operations to your advantage.


I hope you’ve enjoyed our tour of the amazing power behind Adaptive Listening. We’re incredibly excited to see what our customers do with these features, and how they use them to help their business Listen Smarter.

Feedback and Knowledge Base