Sentiment analysis on Twitter

A Guide to Sentiment Analysis Using Twitter

Sentiment analysis is the process of determining the feeling behind a piece of text, conversation or a social media update. It has been used on Twitter and other social media channels as a way of judging public attitude for many years and 86% of marketers are said to value it highly.

On Twilert and indeed in Twitter search, it is possible to find tweets obtaining to positive or negative sentiment using the search operators :) and :( – indicating happiness and unhappiness. This basic type of search not only helps you to gain brand feeling and the areas which need improvement, but can also help you search out the topics, competitors and services your customers like and dislike!

Variations of these emoticons :( and :) are also taken into account which is useful, given how many we now tend to use on social media.

Why use sentiment search on Twitter

As marketers we often think we know what our customers are thinking and how they are feeling. But do we really? Expedia Canada used sentiment analysis to react quickly and creatively to an advertisement blunder involving an ‘annoying violin sound’ that was widely reported on Twitter. Without gaining intel on the ‘sentiment’ of their customer responses on Twitter, they might have carried on thinking that the sound in question was a great idea.

Studies that looked at sentiment analysis of tweets around Hilary Clinton and Donald Trump on the recent presidential election in fact showed up much more accurate results obtaining to the final result than many of the poles that outrightly asked people. Monkeylearn found that not only was Trump mentioned nearly double as many times as Clinton but also that Trump had a better positive to negative ratio than Clinton.

This could even be dated to certain events happening around the election. Remember when the FBI decided to not recommend prosecution into the email case surrounding Hilary? Negative sentiment soared on Twitter as a result, as you can see in the graph below.

clinton sentiment on Twitter

This shows how sentiment analysis can be a great indicator of early wins and losses, as well as general public feeling around a topic or area. When customers are polled, they can sometimes pre-empt what they think they should say or how they should act. When you track sentiment, you’re gaining insight into how they really feel.

Benefits of using sentiment analysis in business

As we know, Twitter is a reservoir for a vast amount of data. This data can be extremely useful for business decisions, new product research and deciding on what content to share with your audience. Pre social media you know what we relied on? Instinct. This is kind of risky when it’s being used to determine where the money moves.

As valuable as workshops, user testing groups and customer surveys are, they take up time, effort and money that sentiment analysis on social media platforms can provide in minutes. Especially when you save the results using Twilert.

How to use sentiment analysis with Twilert

In Twilert, using sentiment analysis is simple! Log in to your account (or setup a free trial at www.twilert.com) and you’ll see the Twilert search panel which looks like the below. Here, you can choose the sentiment keys by using the dropdown under ‘misc’ where you will see ‘:) positive’ and ‘:( negative’ which are your gateway to positive and negative search on Twitter.

sentiment analysis tool

You can also use any of Twilert’s other advanced search operators to compile a more specific sentiment search. For example, you could add the exact phrase “phone provider” and the sentiment operator :( in order to find any negative tweets relating to that term.

Negative Twitter search operator

Useful sentiment analysis Twitter searches

One of the things we’re pretty good at is seeing the gaps in the market where Twitter search can help you uncover some key information that others may miss. With sentiment search, the right Twitter search terms and formats can help you to uncover tweets that scale out the competition, find unhappy customers and leverage additional positive sentiment for your business. Here are the ones you need to know –

The brand name Twitter search

To find out what people are saying about your brand simply enter your brand handle and the sentiment operator. To ensure you don’t miss a mention you could also use the OR operator in order to combine search terms. For example the search below would find instances mentioning ‘Pepsi’ or the brand handle @Pepsi with either the negative or the positive sentiment search signal.

pepsi OR @pepsi  + :) OR :(

The competitor search

A super useful search for industry monitoring is combining the negative search operator with your competitors’ handle. This way, you see exactly who is unhappy with their service and why! Leading to better behaviours and the chance to connect with more potential employees.

The potential customer search

One sure fire way to find customers is to uncover the key language terms they use to search for your product. For example searching ‘wish there was’ :( followed by your product or service keyword, could help you to zone in on potential customers who don’t even know you exist.

Get started

To start setting up your sentiment searches today head to www.twilert.com

Related Articles