Please enter a valid Email Address
Predict the likely traffic your article headline will yield + Score your overall headline quality & it’s ability to drive engagements, reach & SEO Value.
I’m sure you’ve heard this before, but just to put it out there – If you’re a blogger or publisher you need to be on Facebook. It’s really that simple because the world’s largest social media network is ripe with users sharing content which offers a plethora of insights on what to talk about, create content that has the potential to go viral & amplify your reach exponentially.
“10 Things You Didn’t Know Happened in 2016”
“Here’s Everything You Need To Wrap Up 2016 in Style”
“Here Are The Most Entertaining Stories of 2016 That You Need To Know”
Believe it or not, one of these headlines yielded 8 times the traffic compared to the others. Any guesses on which one – it wasn’t the third. Surprised?
I’ve worked with several such publishers on helping them adopt data driven approaches to power their content strategy and while everyone talks about engagements, shares, reach and so on, let’s face it, the one metric everyone is watching most closely is traffic – how many people clicked (notice the past tense – this will be important in a second), on your link & actually went back to your site to read an article.
With the growing crack down on the organic reach & the increasing costs to promote content – the ability to predict whether a certain post will work towards your goal seems like a godsend.
What if I told you that you can now *accurately* forecast (aka ‘predict’) how many people will visit your site based on the headline of the article promoted in a social media post?
Yes, now you can!
After working on hundreds of thousands of publisher metrics & testing, we’re launching a content & headline analyser – this simple but powerful new feature will forecast how many visits your article is likely to generate from social posts among the following other factors (PS. the end of this post has a technical/methodology section on how this works in case you’re interested!):
Readability: Insights into the ideal length & characters for your headline so that it maximizes SEO value & piques audience interest.
Nifty previews also allow you to see exactly how your users will skim it across social media, search results or email subject lines.
Word Balance: Use proven research to leverage the right mix common, uncommon, emotional & power words that ensure that your headlines are readable & that they catch audience attention.
Focus, Hashtags & Trends: PropheSee will pull up words & topics you should focus on based on your headline & also grade it based on the whether or not users are talking about this on social networks or searching across the web. All of this is used to give you insights on what hashtags to use, what to focus & even the kind of sentiment & emotions you should drive through your content.
How It Works
The overall architecture of the headline & content analyser is composed of different natural language processing and information retrieval modules.
First, the headline is preprocessed and cleaned by the headline pre-processor which removes the textual noise (stop words, slangs etc). The original and clean headlines are then passed into other modules (word balance, text readability, text statistics, views prediction, recommendation engine) to surface the relevant information.
These modules are described below:
Vocabulary Analysis: This module analyses the overall words used in the headline using different text cleaning and text matching algorithms such as keyword normalization, levenshtein distance, fuzzy-wuzzy match etc. A set of in-house word dictionaries are used to ensure a right balance of words in the headline.
Textual Readability Analysis: The readability is computed as per the standard readability metrics such as syllable percentage, average difficult words, Flesch reading Grade Formulas etc. These metrics are wrapped in the form of a package which is accessible by REST API.
Text Mining and Analysis (Content Statistics, Entities Detection and Tone analysis): Based on several text mining practices different statistical parameters from are obtained (frequency based features, part of speech, entity detection) from the text. The entity detection system uses grammar based dependency parsing, lexical analysis and pos tagging to obtain right entities with an accuracy of 96%.
The tone analysis module calculates in-depth sentiment, sarcasm, psychometric tone and emotional tone using hierarchical grammatical parsing based rules linked with sentimentally loaded dictionaries and flexible morphological analysis.
Suggestions and Recommendations: This module tracks social and digital media channels. The key entities and focus keywords obtained from the headline are linked with the information obtained from these channels. Using aggregation analysis, trend analysis and correlation analysis different hashtags and keyword suggestions are recommended based on what people are using across social media platforms, press coverage & so on.
Views Prediction: This module consists of a machine learning pipeline that performs feature engineering, data cleansing and regression modelling to estimate the number of views for a headline. The feature engineering includes key entities present in the content, their social media engagements and sentiment, their related search engine rankings, recent news and media coverage, the website Alexa rankings of a brand and the past performance of a brand in terms of headlines used & traffic yielded.
The results of every module are passed in the Insights Generation Engine that makes sense of the data and generate textual insights. These insights are validated against the standard benchmarks created for respective features of the headline.
A limited version of PropheSee’s Content & Headline Analyser is available for free on the Free Tools section of our website. The complete feature is available for Free on any paid plan, which start at just $29!
Have questions? Interested in learning more? Drop us a line at email@example.com & we’d be happy to chat!