Social Media

Can you recommend a good social listening api product?

Jason Gorham Corporate Executive Recruiter at Aflac

June 26th, 2015

We are trying to find a good social listening api that allows us to track positive sentiment for brands.  We would need a tool that can track by country and also allow us to push a message back to that person expressing positive sentiment.

Any recommendations?    
You don’t need thousands of dollars to acquire new customers and move them through the funnel. In this course, you’ll learn lean marketing strategies that work with budgets as small as $20, giving you the tools to optimize your budget, test and adjust campaigns, and increase ROI.

Todd Stein Vice President, Amendola Communications

June 26th, 2015

Most Expensive Option is probably Meltwater.

 There is always Sprout Social.

 Free and small to media priced tools (FREE are highlighted in yellow):

This is by no means a definitive list but it is a good place to start your search for the tool that best matches your specific needs.

 

Addictomatic shows you “the latest buzz” on your chosen topic from Twitter, Flickr, YouTube, Google, WordPress, Bing News, and more. After you search, you can personalize your results dashboard by moving around the source boxes. When you're done, bookmark the page and keep coming back to your personalized results dashboard for that search.

 

Attentio: This multi-lingual Belgium-based Attentio tracks global conversations taking place across social media (blogs, forums, social networks, Twitter, YouTube) and online news sites.

 

Brandwatch: UK-based Brandwatch trawls the Internet looking at news, blogs, forums, wikis and social networking sites and finding mentions of brands, companies, products and people. Clients define keywords (brands, topics, people names, products) and receive reports and brand summaries that they can take action on. Features mutilingual coverage.

 

Crimson Hexagon: Based on a technology licensed from Harvard, Crimson Hexagon taps into billions of conversations taking place in online media and turns them into actionable data for better brand understanding and improvement.

 

Crowdbooster focuses on gathering data from your Facebook and Twitter feeds. It will give you some vital information regarding your social media strategy, such as when people are most likely to view your latest images or video uploads. Other features include being able to analyze impressions, total reach, and engagement.

 

Google Alerts are email updates of the latest relevant Google results (web, news, etc.) based on your queries.

 

Google Analytics provides integrated reporting for the Google +1 button "out of the box." To track networks including Facebook and Twitter with Google Analytics, you will need to integrate the Google platform with each network button. Social Engagement reports allow users to see how content is being shared through social interactions such as Likes.

 

Hootsuite: This Twitter tool allows you to create multiple tabs with 10 columns to help you monitor your mentions, direct messages and keyword searches. There are paid and free versions to choose from.

 

HowSociable: A free account allows you to track 12 social sites, including Tumblr and WordPress. However, if you’re interested in more, such as Facebook, Pinterest, Twitter, etc. a pro account is a must. One of the unique aspects of HowSociable is the analytics - you get a breakdown of scores for individual platforms, showing you which social media sites are working out best for you.

 

IceRocket: This tool offers blog, Twitter and Facebook monitoring in 20 languages, as well as allowing you to choose the periods of time you are interested in monitoring.

 

Radian 6: Tracks keywords across all types of social media, monitoring everything from location to sentiment and providing real-time data on a dashboard widget. You can create, schedule and control where your content is published. It uses a monthly subscription based pricing model, with the monthly fee varying depending on the number of topics monitored each month

 

Reachli: A tool that measures and optimizes video and image content. It offers various features for measuring image and video effectiveness, and is particularly famous for its Pinterest analytics.

 

Social Mention: Easy to use with in-depth results, social Mention monitors over one hundred social media sites providing detailed analytics and shows your social influence across four categories: strength, sentiment, passion, and reach.

 

Spredfast: Measures data gathered from Facebook, Twitter, YouTube, and Flickr, to show how many people you’ve reached and whether or not your target audience is engaging with you. The data is clearly presented in formatted graphs. Other useful features include a calendar that shows you when the optimal tweeting times occur.

 

Sprout Social lets users manage, monitor, and track their presence on social networks including Twitter, Facebook, and LinkedIn. Social Reports provide messaging analytics, a drill-down into user activity and demographic information. Among many other features are separate tabs for Twitter and Facebook that let users capture data in a browser or export it to PDF. Message summaries break down updates into type and show how widely they were shared.

 

Spiral16: Provides an in-depth look at who is saying what about a brand and compares results with those of top competitors. -

 

Sysomos: A real-time monitoring and measurement tool that provides constantly updated snapshots of social media conversations delivered using a variety of user-friendly graphics.

 

Trackur: Tracks sentiments and trends in social media mentions with alerts via email and RSS Boolean keyword query with option to exclude specified sites.

 

Twazzup: A good choice for social media beginners looking for Twitter monitoring. You just enter the name you want to track and you instantly get real-time updates including top retweeted images and links, most active top influencers, and the top 10 keywords relating to your search.

 

TweetDeck: Twitter tool similar to Hootsuite, with features for monitoring multiple Twitter accounts, creating and managing custom timelines, automated searches for keywords and hashtags, and sending out scheduled tweets, and more.

 

TweetReach: Allows you to track the impact and implications of your brand’s social media discussions. You can search by Twitter name, URL, hashtag, or keyword phrase and view reports that show the reach and exposure data for relevant tweets.

 

TwentyFeet: This social tracking tool gives you monitoring tools and analytics for social networks, as well as other services your business uses, such as link shortener tracking. Also aggregates your data and lets you set up email alerts.

 

Twitonomy: Australian-based monitoring and analytics tool that lets you measure mentions, reach, and even conversions. You can track follower growth, click-throughs on your tweeted links, mentions, retweets, favorites, and more.

 

Visible Technologies: Helps clients monitor, analyze and participate in social media conversations as well as protect their executive and corporate brands online.

Anonymous

June 26th, 2015

If you're doing multiple countries, then you need a platform that includes translation. Even if your countries are US, Canada, UK, etc, the way English is used varies. Think about someone Tweeting "your product is sh*t" vs "your product is the sh*t": Two very different sentiments, and ones where knowing the slang of the country really affects meaning.  I understand Radian 6 is highly recommended, but as a former translation company owner who programmed an early social media monitoring tool I can tell you their ability to analyze in multiple languages and dialects is horrible. Look into Reach 7 instead or something else that actually takes into account that we don't all speak the same language.

Greg Armshaw

June 26th, 2015

There is a lot of good advice here, especially from Tom.  Building on this...

Automated text based sentiment analysis is not great even in the best platforms, and when you move away from English it is even worse.  Usually you will find that the majority of systems will give you predominantly neutral sentiment, false positives and false negatives regularly occur.
e.g. [Sneakers] " Blisters again! -awesome!"

In my experience there is NO system that comes anywhere near close to giving robust enough analysis to facilitate AUTOMATED response workflow.

With some understanding of the category you can start to build queries to highlight some of the most likely sentence construction around positive mentions of the brand and negative mentions of the brand.  This tends to be the most effective way of picking up real brand sentiment.
You would need to look carefully at the platforms to ensure they facilitate this kind of query through an API.

The larger brands would pay very little attention to an automated sentiment score these days.  For those that are running customer service through social there are a number of tools that build workflows around alerts and customer service conversation hotspots and these can straddle a number of tools. There is a high level of human interaction.  Automated responses get flamed badly if they are irrelevant, making a bad situation worse.

Additionally with the increase in image and video based social sharing the text element of what people share is becoming less likely to have a brand mention within it 
e.g. [Sneakers] 


It is not clear what you actually want to do in your workflow but if you are looking at providing an SLA that picks up ALL positive brand mentions then you need to be careful about costs for brands that peak in social media. For instance Nike had 1.8M mentions on twitter in the last month.

Lastly be careful about platform pricing.  Many of them are individual brand oriented and scale badly for mother brands with multiple portfolio brands (e.g Unilever).

I haven't mentioned a platform brand because the ones I would suggest have already been mentioned and really you need to look at specific markets, very few do well globally. I have done multi-market international deals with platforms before but usually there are exceptions on a market by market, language by language basis.

Have you heard of dark social?  Many of the social platforms now cannot even be crawled by listening platforms and the trend is moving towards chat apps (whatsapp, Messenger).  So be careful about a business model that relies on open sharing.

Sorry for the long response but hopefully it is all relevant.  Just to re-iterate please do not plan for any publishing process which is NOT approved by a human just before response.

David Neff Digital Strategy Manager at PwC

June 26th, 2015

I know the folks at Radian6 have a solid API tuned to work with applications. What are you building that will use this in the product?

Cindy James Life Hacker & Self Directed Careerist

June 26th, 2015

I would take a look at this one www.affin.io

A.J. Lawrence Data focused revenue growth executive from startups to Global1000s

June 27th, 2015

Throwing another one on the pile, ishttps://discoapi.com/More of a social discovery Tool, with a very wide open dev and api environment. I've run some small social sentiment analysis projects on it and they worked like a charm. Probably worth chatting with EliMandelbaumrecently of Pluggedin Ventures, who's taken a keen interest in helping them expand. 

David Neff Digital Strategy Manager at PwC

June 28th, 2015

That's a GREAT list. Thanks for sharing!

Jason Rosengren Co Founder and CEO at Airzurv Technologies

June 29th, 2015

Take a look at this one:

https://api.nexalogy.com/

Grant Simic Co-founder & CEO, Astute Doctor Education. Tech entrepreneur focused on health, education and sports.

June 30th, 2015

I'd also recommend taking a look at BrandCare by SocialFigures http://socialfigures.com - it was built to directly compete with Radian6, just a powerful, full featured and a lot cheaper. SocialFigures is a startup that's gaining traction in the US, south america and australasia - I know the co-founders, so if you want an intro please let me know.

Tom Maiaroto Full Stack Consultant

June 26th, 2015

Terena makes a very good point and there are very few people out there doing sentiment analysis for multiple languages. However, you can train the classifiers manually (or Amazon Mechanical Turk or hire some other workforce) if you were to approach it all yourself.

This is an area I've worked in and researched for some time now. Here's the (again open-source) Golang package that Social Harvest uses for sentiment analysis (in memory). https://github.com/SocialHarvest/sentiment - as you'll see in the notes, I cited sources for the training sets. Though it's also possible to train more in there, it has primarily English and then some Russian. Like I said, not many people out there are doing more than English.

You need to train thousands of domain specific examples for it to become viable...At that point you're talking about an 80% accuracy rate +/-. In case you were curious was was involved.