Facebook Topic Data Meets the Power of Insightpool

Title  Area  

The Social Relationship Intelligence Platform

#FBdata

Insightpool x Datasift Webinar

FACEBOOK TOPIC DATA Meets the Power of

INSIGHTPOOL

#TopicData

Devon WijesingheCEO of Insightpool

@DevonWijesinghe

Speaker

Co-founded Insightpool in 2012, and has led the company from two to 60+ employees, acquired a Silicon Valley start-up, Next Principles, and is currently revolutionizing marketing and sales across social. 

#TopicData

Tim BarkerCPO at Datasift

@timbarker

Speaker

Joined DataSift after serving as VP EMEA Marketing atSalesforce, leading a world-class marketing team to create the social enterprise and cloud-computing industries. Tim has entrepreneur DNA, having founded 3 successful startups.

#TopicData

Facebook Topic Data #TopicData

#TopicData

What data are we analysing

Facebook Page

Topic Data

Posts, Likes and Comments on brand-owned page globally

Posts, Likes and Comments on Facebook

#TopicData

What’s on your mind?

Content and Behaviours

7

CONTENT DEMOGRAPHICS LIKES and SHARES

Anonymized and aggregate topic data •  Posts •  Pages Posts Plus engagement data •  Likes on Posts •  Shares on Posts •  Comments (no text) on Posts

Data enriched with •  Demographics •  Topics •  Sentiment

#TopicData

Topic/Entity Data from the Facebook Open Graph

8

Entity/Topic Detection is applied to every post Leverages Facebook’s massive Open Graph data set to identify topics and events. The same technology is used to surface Public Trends on Facebook. Enable exploration of topics related to a brand Topics enable exploration of the data without access to raw posts. For example - cluster the topics related to a product or event. Topics can be used within filters to remove noise. Create a CSDL filter for posts for the movie “The Interview”.

fb.topics.category == “Movie" and

fb.topics.name contains “The Interview"

CONTENT

TOPICS

Sony (Organization) The Interview (Movie)

Kudos to Sony for distributing The Interview

#TopicData

Demographics and Engagement

9

Gender Age Range Location

Male Female

18-24 25-34 35-44 45-54 55-64 65+

Country State / Region

Self-declared, not

derived.

Enables exploration of audience segments Demographics for every post, comment and like. Understand the audience engaging with your brand.

#TopicData

Privacy-First Data Management Controls

Social data never leaves Facebook Social data is processed by DataSift technology running inside Facebook’s network. User identity is removed before processing User identity is removed from social data before processing by DataSift technology. Results provided in aggregate, anonymized Only summary results containing insights from 100 or more people are delivered by DataSift. 30-day retention period for underlying social data Data deleted from DataSift’s technology after 30 days. Prevents analysis of minors Minimum age applies for data collected for analysis.

#TopicData

Facebook Topic Data is a Killer App for Research

Demographic Context

38M people in UK, 210M

people in North America

Share with friends

Gender, Age, Location,

Education, Relationship

Status

Data Structured for Standardized

Analysis

Representativity Lack of demographics Unstructured data Self-promotion bias

#TopicData

How it Works

12

DataSift platform connects to the real-time feed of Posts, Comments, Likes.

2 • CATEGORIZE

Filtered data can be classified with customer-specific rules.

4 • QUERY

The Index is sub-queried and further processed using CSDL against 60+ attributes.

5 • ANALYTICS

Aggregated and anonymized data returned for developers to create applications, analysis and visualizations.

Each customer defines their specific filter based on their criteria.

1 • FILTER

Filtered and categorized data is indexed into a real-time analysis engine.

3 • INDEX 6 • MINING AND VISUALISATION The data is then mined and visualised by using a range filters, classifiers and interactive visualisations to help you craft insights.

#TopicData

Why PYLON for Facebook Topic Data?

Data science via PYLON for Facebook topic data is at the center. Information fuels Insightpool’s targeting, messaging, and activation flow.

#TopicData

Client Objective

Major auto parts retailer is interested in partnering with NASCAR. Major auto parts retailer needs to find out what makes NASCAR fans tick -Who are they? -What topics spark influencer engagement?

Insightpool creates a PYLON for Facebook topic data filter to find Facebook posts relevant to NASCAR

FIL

TE

R

Include Tags of “NHRA, NASCAR, Motocross, drag racing, top fuel, funny car, PEAK racing” And NOT Tags of “racing, race driver, ski, bicycle, professional, pro, BMX, obstacle, spartan, champion, champ, bike, raced”

#TopicData

Data Analysis

Peak: Wednesday, August 26

All sharing and engaging on Facebook is uncovered when a filter is applied. We discovered a peak and further investigated.

#TopicData

Deeper Dive

We focused in on the specific peak to uncover topics/demographics that sparked increase in social sharing and engagement.

PYLON can bring detailed breakdowns on:

Sentiment

Post’s author’s region Age

Country of Origin

Top Link Shared

Gender

Language

Hashtags

Media Type (Posts/Links)

Facebook Type (Comments, Likes, Stories,)

Topics

#TopicData

Topics and Data Analysis

A multitude of anonymised and aggregated data can be drawn out from the overall NASCAR discussion, or from any specific demographic or tag. PYLON has the capability to create customizable tags to further filter the sharing and engagement around NASCAR based off of phrasing, tone and sentiment. For example, PYLON can filter out only the negative sentiment posts, to compare to the positive sentiment posts. We can then determine the topics that are frequent in positive posts as well as negative posts.

#TopicData

Sentiment Example

Positive Sentiment

Negative Sentiment

Tag.sentiment of fb content contains any: “love, obsessed, can’t wait, excited, pumped, awesome, great, amazing, excitement, exciting”

Tag.sentiment of fb content contains any: “hate, stupid, dumb, don’t like, lame, bad”

Sentiment of the overall filter Negative sentiment of towards a specific topic

#TopicData

More Insights

With PYLON for Facebook topic data, you can visualize:

Post author’s region Post’s author’s gender

Post author’s gender

#TopicData

Key Insights

Top 3 links shared

•  The top 3 links being shared revolving the topic of NASCAR are all on FoxSports URLs.

•  This can be used to inform ad buys, or content to be pushed by the brand

Takeaway: The most amount of sharing and engagement was coming from Fox Sports around a tribute to a driver.

#TopicData

Conclusion

•  This information informs major auto parts retailer to be empathetic towards NASCAR fans and also create more sharing and engagement around tribute content.

•  This also informs them to place media buys on Fox Sports where a large percentage of traffic is coming from.

http://foxs.pt/1NAb2pX

#TopicData

PYLON Benefits

The major auto parts retailer can now amplify it’s targeting, messaging, and activations around the tribute story. Valuable Hidden Insight: The top traffic drivers were coming from Fox Sports and the content was a tribute to a specific driver. Client also knows where to spend media dollars and put their brand in front of the Fox Sports audience to drive greater conversions

#TopicData

#TopicData

Facebook is the foundation.

Q & A

#TopicData

Title  Area  

Thank You!

#TopicData