Tag Archives: sentiment analysis

Introducing Lava – The Emotional Search Engine

Lava emotional search engine

Many of us are familiar with using tools like BuzzSumo to estimate the share-ability of a content idea, or Google Keyword Planner to estimate search demand. But how do you find out how people actually feel about a topic?

For those of you who didn’t catch James Finlayson’s and Lisa Myers’s talk at BrightonSEO last week, we have just launched a brand new tool which does just this. We proudly present to you Lava, the world’s first emotional search engine.

It’s time to get emotional

By creating content that resonates emotionally, we can increase our chances of getting read, watched, shared or featured by the press.

Lava works by indexing every major UK newspaper (currently at around 8 million articles), soon to include US publications as well. By inputting a search term such as a person, place or thing, you can see how sentiment around that subject has changed over time.

For example, here is how the media feels about three US presidential candidates over the last six years:

sentiment analysis of US politicians

Lava assigns every article a score based on how positive or negative the sentiment is at a sentence level, and then calculates the overall sentiment of the article as a whole.

Strong negative sentiment does not necessarily mean that journalists hate a particular person or thing. It could just mean that there is negative sentiment occurring in the same articles as the subject, as would be the case if I was cited as a source in articles about a tragic event.

Some cool examples

Historic sentiment analysis can be used to tell some interesting stories about how we feel about certain brands, politicians or celebrities – and how this has changed over time.

Here are a few possible things you could look at with Lava:

Is sentiment towards professional footballers linked to their performance?

Does sentiment towards the characters in Breaking Bad or House of Cards mirror the plot?

Does sentiment towards famous musicians like David Bowie peak after they die?

How can Lava be used in content creation?

cat looking at painting

You’re probably wondering what all this has to do with content creation.

Well, Lava can be used to gauge the emotional impact of your early concepts and help you decide which ideas to take further. So, by inputting a few different content ideas, you can get a sense of how your audience feels about a subject over time.

A spiky profile or consistently high positive or negative sentiment would suggest an idea worth pursuing; a flat profile that’s hovering around zero would suggest apathy towards the topic.

In this respect, Lava can used alongside other indicators to predict content performance, such as:

Shareability – How many shares has other content published on this subject got?

Search Demand – Are people searching for this topic on Google, and is it trending up or down?

Audience Fatigue – Has there already been a glut of content published about this?

Interest – Are people asking questions on sites like Quora or Reddit?


Ultimately, Lava is about helping you better understand your audience and what makes them tick. You can then use this information to enhance your content campaigns.

I hope you have fun playing around with it. It’s still in beta, so please email me if you have any questions or suggestions.

Check it out for yourself here: lava.vervesearch.com