Next Generation Social Media Monitoring: Canadian Technology Solves

[email protected]
Next Generation Social Media Monitoring:
Canadian Technology Solves
The Social Media Riddle For Retailers
Document produced by InfoTrellis Inc. and Missing Links Insights Ltd.
September/October 2013
1|P a g e
[email protected]
Why is social media important?
The world of marketing has significantly shifted over the last five years, both as it relates to understanding
the customer and as it relates to advertising and promotion best practices.
Traditionally, companies used market research reports, internally conducted research, customer surveys,
and focus groups to understand what their customers liked or disliked, how they reacted to brands,
products, and promotions, and what demand there was for new potential brands and products.
To promote or advertise, companies largely used channels like TV, radio, catalogs, posters, flyers or email
advertisements to deliver their message.
In the last few years new approaches have become possible with which to both understand the customer
and to reach them with promotional materials.
These changes are due to the widespread adoption of social media channels, platforms ranging from
Twitter to Facebook to Pinterest which allow customers to exchange information and to communicate
with each other or with retailers in real time. The same channels can be used by retailers; they can
monitor what customers talk about and how they talk about it, both related to their own brands and
products and related to their competitors’. This gives retailers insight into customer preference and
sentiment with a speed and frankness which was not possible a few years ago.
Over 80% of North American consumers use some kind of social media input for purchase decisions.
Retailers can no longer afford not to monitor what customers are saying. Studies have shown that many
customers today actually expect retailers to be monitoring social media. Rather than call customer service,
they put their questions or concerns on social media channels and expect retailers to be pro-active and
react, often in as little time as a matter of hours.
In addition to monitoring and communicating, social media is a powerful promotional channel. Research
suggests that social media lead conversion rates are 13% higher than traditional marketing conversion
rates. Companies themselves are reporting in increasing numbers that Twitter and Facebook directly
generate leads and sales for them.
As a result, retailers who do not carefully monitor, learn from and use social media promotion is at high
risk of lagging behind. The 2013 – 2015 period will show a very significant shift: adoption rates of social
media strategy among retailers will grow, and these companies that start to listen and to communicate
with their customers online will stand to benefit from increased visibility and positive brand recognition.
As more companies become adept at social media connectivity, those that are slower to incorporate social
media into their overall strategy may see a significant drop in sales and customer loyalty.
2|P a g e
[email protected]
What are the different social media channels and what role do they play?
Key Features (users, profile)
Short messages (140 characters or less); very
fast reaction; messages can spread very
quickly due to easy re-tweets; customers
often use twitter as a substitute for customer
service; people will often use Twitter to
subscribe to strangers like celebrities,
comedians, or pundits
Status updates; photo albums; tagging of
friends; “like” button to subscribe to pages or
show approval for individual posts; share
button to repost and spread links; used to
connect to people primarily known in real life
relationships like family and friends;
usernames tend to be peoples’ real names
Is not self-contained – interacts with all
other Google products and services; stream
that displays updates from followed users;
“circles” that allow grouping of friends; video
and audio chat functionality; “like” button
represented as “+1”; requires users to use
their real name
Focused on images; image collecting and
sharing; feed of new images from people a
user follows; images can be sorted by hobby,
event, interest, etc; can “re-pin” someone
else’s image to your board; can “like” an
image; also supports video sharing.
Microblogging; like Twitter, intended for
shorter posts; not limited by character
length; can post images, text, video, audio;
very frequently used reblog button
Targeted at professionals; contains resume;
shows job experience; has a job board type
functionality for jobseekers and employers;
people link to colleagues and business
associates; features a “like” button for
individual posts; allows users to subscribe to
company pages and other users; users can
“endorse” another user’s specific business
skill, which accumulate on a user’s profile;
features interest-based groups; requires
users to use their real names
Wordpress and other blog hosting sites
Reddit : News sharing
MySpace: Pre-Facebook, similar
Twitter is a public social media channel used by over 500
million users. Due to the large number of tweets per day
(nearly 60 million), many of which have no relevance to
retailers, it requires advanced technology to extract all the
relevant information and to exclude the noise. Twitter is
the most “Social Media analytics friendly” channel
Facebook is a social media channel that experienced
massive growth in the last decade and has over 1 billion
users – it features banner ads on the sides of the screen
but also can promote corporate content into a user’s main
feed, which they see when they login on the home page.
Because people tend to have close friends and family on
Facebook, often they will use it to solicit advice on an
upcoming purchase decision
Slower growth than Facebook but has reached 500 million
users as of 2013. Having lots of Google+ followers boosts
SEO for Google search engine. Similar to Facebook in
functionality but is more connected to a user’s other
internet usage, such as YouTube, Google Search, Gmail,
etc. G+ appeals to more mature users than Facebook but
is less formal than LinkedIn.
Pinterest users are about 70% women; popular topics
include cars, fashion, food, travel and design. Pinterest
drives sales primarily for the retail industry and has
demonstrated conversion effectiveness equal to or
greater than bigger platforms like Facebook etc.
Tumblr tends to be less about the user’s personal life and
more about hobbies and interests; some users, rather
than generate content, use the “reblog” function to collect
content they enjoyed onto their dashboard, allowing other
users to scroll through and see their collection of
reblogged items, which may or may not be of their own
creation. This platform is often used to talk about pop
culture and entertainment, but also covers things like
politics and activism – things can spread very fast if they
evoke strong emotion like anger or delight.
LinkedIn is a highly professional networking platform that
also functions as a hiring and job seeking platform. It has a
higher lead conversion rate and a user’s connections on
LinkedIn will often be very different from their
connections on platforms like Facebook or Twitter
because the emphasis is on professional and not personal
relationships. Has 225 million users, is smaller and more
targeted than Facebook.
Instagram & Flickr: Photos
DeviantArt: Artist networking
Other small niche networks catering to specific interests
3|P a g e
[email protected]
What and how can we learn from Twitter?
First we begin with a brief overview of terminology specific to Twitter.
Tweet: Twitter is a platform for publishing posts that are under 140 characters of text; this is also known
as “microblogging”. One of these single posts is called a Tweet. When a user Tweets something, it appears
both on their twitter account (or “timeline”) and in the subscription feed of those users who follow them.
All Tweets are publicly accessible unless a user changes their privacy settings.
Username or Handle: When a person or company signs up for Twitter, they select a username. Anywhere
that a username appears on Twitter, it will be preceded by the “@” symbol to indicate that the following
characters represent a user. This is the same for both corporate accounts and personal accounts.
Example: @HoltRenfrew is a username, Holt Renfrew is not.
When someone posts a Tweet that contains a username, as demarcated by the “@” symbol, the owner of
that account is notified by Twitter that someone has mentioned them. This is an easy way for a retailer
with a corporate account to monitor Tweets that mention or are directed at their corporate account.
It is important to point out, however, that as in the example above, “@HoltRenfrew”, which uses the “@”
symbol and references an actual Twitter account, will trigger a default Twitter notification, where the
plain text “Holt Renfrew” will not trigger any sort of notification.
Follower: By “following” a Twitter user (this can be anybody from a corporate Twitter account to a
celebrity to a close friend) you subscribe to any Tweets they post in the future. If somebody is your
follower, it means all your Tweets appear in a live feed on their Twitter home page or application along
with the Tweets of the other users they follow.
A user’s number of followers is a useful and easy to find metric which indicates how many people have
subscribed to see the content that user posts.
Retweet or RT: If a Twitter user encounters a Tweet by another user they find noteworthy, they may
choose to Retweet the message. This reposts it on their own timeline while identifying the original author,
and will cause it to appear in the subscription feeds of their followers.
When a person or company posts a message, Retweets can allow it to reach a significantly broader
audience by passing it through the network of its followers, such that a person that does not follow the
original author of the Tweet may be exposed to the Tweet via one or many intermediaries who felt the
message worth passing on.
Opinion Leaders: Some users have a large number of followers and are Retweeted often; those who tend
to Tweet about specific subjects in particular tend to be regarded as having greater expertise, but
influential users need only a large, devoted following to have an impact. Some examples of opinion
leaders in the retail space include celebrities, fashion magazines, and prolific blog authors.
4|P a g e
[email protected]
It is very beneficial for retailers to have good relationships with opinion leaders, as opinion leaders by
definition have a very large follower base. If an opinion leader retweets a message from a retailer it can
exponentially increase the reach of that tweet. (The number of opinion leaders following a company is a
very useful metric – but to get this metric a Social Media monitoring tool is required)
With this understanding, we can assert that if a person or company wants to achieve the broadest
possible reach with their Twitter marketing messages, they should focus on:
 Maximizing the number of their followers and
 Developing good relationships with opinion leaders, to ensure that they retweet their messages
Before we review what we can learn from Twitter, we need to discuss who we are learning from.
We can identify three large groups as it relates to Twitter activities:
Corporate (retailer)
Twitter Account
This is the twitter account of a
retailer. A retailer can use it to
share news about products,
promotions etc. A retailer can also
use it for customer service by
responding to customer concerns.
Opinion Leader
Opinion leaders are important, as
they are followed by many Twitter
users and have the power to
influence their opinions and reach
many with just one Tweet.
Casual User
Individual accounts belonging to
casual users who are not opinion
leaders but may be following
opinion leaders.
A simple view of corporate Twitter activities can be
achieved by anybody; Twitter is free to access and
participate in. Metrics like topic trending, sentiment
monitoring, and changes over time require more
advanced tools than Twitter’s default interface.
Many leading organization actually have multiple
accounts, with one dedicated specifically to customer
service. (e.g. @Username vs @Username_Support)
Social media monitoring tools can easily identify
activities of opinion leaders, as long as they mention
the company account using the @Username format. If
the “@” sign is not used and the company name just
mentioned in passing as part of discussion, more
advanced tools are required to identify when an
opinion leader is talking about the company.
Social media monitoring tools can easily identify
activities of casual users, as long as they mention the
company account using the @Username format. If the
“@” sign is not used and the company name just
mentioned in passing as part of discussion, more
advanced tools are required to identify when a casual
user is talking about the company.
5|P a g e
[email protected]
The following table summarizes some of the insights we can gain:
What can we learn?
How much
discussion of the
company is there on
social media?
What topics do the
customers or target
markets talk about?
What is the reaction
to marketing
Who are the opinion
What sentiments do
consumers express
about the company?
The simplest way is to track mentions of the
company handle using the format
@Username; these mentions will appear on
the standard Twitter profile page of an
account. Achieving a more advanced view
(capturing discussions about the company
without Twitter’s @Username flagging, or
sorting by opinion leader vs. casual user)
requires more advanced tools.
Understanding what product categories or
brands customers talk about and how those
topics change day to day, week to week or
month to month gives insight to what is
important to customers. Examination of
topics over time related to the retailer can
reflect seasonal patterns or indicate
noteworthy instances of change in a specific
time period. Identification of local trends (by
looking at where people are talking) can add
further detail.
Retailers can measure the results of their
campaign both in real time to tweak their
approach and retrospectively to help gather
metrics on success. Real time monitoring
allows retails to react fast if those results are
not aligned with their expectations.
Measuring the number of followers a user
has is valuable for indicating a possible
opinion leader, but a strong monitoring tool
also incorporates more sophisticated
grading. Some tools, like Klout, offer a less
focused social media influence score.
Consumers do not just talk about topics, but
they also often express likes or dislikes.
Twitter does not come with any inherent
measure of positive or negative sentiment,
but many social media analytics tools come
with build in sentiment analysis that uses
technologies like natural language processing
to determine where a Tweet lies on a scale of
positive, neutral, or negative.
What is the profile
of opinion leaders?
Detailed leader profiling is offered only by
very few advanced social media monitoring
tools. The impact of such an understanding
can be significant but is difficult to
What are the
activities of
Tracking the above metrics for the market of
a competing retailer allows the advantage of
learning from their mistakes or successes
without any risk. A social media monitoring
tool is needed to do this comprehensively.
In many cases users refer to companies
without using the standard twitter handle
(“@Username”). Filtering out the relevant
mentions can be very challenging; for
example, the recording artist “Gucci Mane”
has nothing to do with the Gucci Brand, and
“The Bay” could refer to the retailer or to
the San Francisco Bay Area.
Most social media analytics tools do not
provide a broad list of topics and topics
categories. Having a predefined (and, ideally,
easy to customize) list can make
identification of trending topics fast and
Customers react very quickly to marketing
campaigns. The reaction (or lack of it) can
allow companies to adjust their campaign as
soon as hours after it has started, which
helps to ensure an optimal outcome.
Opinion leaders are an important part of the
Twitter ecosystem and how a message
spreads. Companies can benefit significantly
by having good relationships with a large
number of opinion leaders
Sentiment analysis can add great additional
insights to topics and topics trends, offering
a way to compare overall perception
between a company and its competitors,
different brands, and different products.
Retailers may want to identify opinion
leaders with negative sentiment, as they can
easily spread a dislike of a company to their
followers. Targeted communication can lead
to a positive boost in sentiment.
Building a relationship with opinion leaders
is made easier by understanding what the
opinion leader is interested in or passionate
about. Understanding, for example, that a
fashion commentator is also interested in
golf and that she has two teenaged
daughters can enable a better connection.
The learning gained about competitors can
lead to clearly measurable improvements or
6|P a g e
[email protected]
What are some of the challenges related to using social media analytics?
Tweets to the company username vs. all tweets: While some of the social activity related to a retailer
makes direct reference to the Twitter account of the retailer, the majority of discussions about a company
must be actively collected, not passively observed. Most comments about a company, product, or service
do not reference the Twitter handle (“@Username” format) – so basic name searches, which look for the
“@” sign, fail to capture them. Research indicates that up to 97% of retailer related tweets are posted
without using the “@” sign, which means most social media monitoring tools are missing up to 97% of the
relevant discussion. (Source)
Qualification of tweets: Twitter and social media in general is a very “noisy” environment. There are many
discussions going on but only a small percentage are relevant or meaningful. Many social media tools try
to overcome the limitation of the missing “@” sign by allowing for text search. Users could search all
tweets to collect all mentions of “The Bay”. These search results tend to return a very low percentage of
relevant activity. Only the most advanced tools in this space have the ability to track all company mentions
in all formats without also including “dirty data”, i.e. irrelevant or meaningless posts that happened to
include the right keyword.
Specific topics (pre-defined) vs. the need to identify topics at each time: Social media provides millions
and millions of data points. Presenting those data points in a way which is easy to read for a retailer can
be a challenge. Tools which provide a predefined (and customizable) list of topics make it easy to see
quickly what brands, apparel types etc. are trending without the need to do one by one topic searches.
Person vs. account: Social media users almost always have multiple accounts across various channels.
They may have an account on Twitter, one on Facebook, two on Tumblr, and a blog on Wordpress. To
understand the person (matching and combining the comments and activities across social platforms) can
be challenging. It is not always evident that the different accounts belong to the same person, and most
monitoring tools are unable to account for this. Looking at each separate account as a self-contained
individual is inaccurate in many ways: it can skew the numbers of sentiment if one person, represented by
three accounts, has the impact on the data of three people; it makes for incomplete datasets such that the
tool provides three semi-completed profiles rather than one consolidated profile.
Profile vs. lack of profile: Social media allows for profiling of the customer / user / opinion leader. Profiling
enables retailers to better connect by understanding a user’s preferences and passions. Profiling can be a
very powerful component of a social media tool, but it lacks accuracy without the view of the person (not
just the view of the account) as mentioned earlier. Profiling can also provide interesting insights that can
lead to smarter advertising on traditional channels, for example what TV shows a particular target market
talks about the most on social media.
Transparency in methodology: Social Media monitoring is not an exact science. There are no clear
definitions to many aspects of it, and the various tools use their own approaches. One example of this is
the way an opinion leader is identified. Different tools use different ways to determine who is an opinion
leader, and how important that opinion leader is. Before any analysis is conducted it is important to make
sure that the retailer knows what the data means and how it was determined. Are Klout scores used to
determine who a leader is? How do those scores work? Are other formulas / scores applied? What are
those formulas? How are they calculated? Having a clear view on such aspects allows for better insights
and therefore better business decisions.
7|P a g e
[email protected]
Changes in Social Media Channels: Social Media is evolving very rapidly. New social media channels
emerge; old ones decline or introduce controls (related to third party analytics). Tools that focus only on
the current set of Social Media channels can quickly become obsolete.
How is Social Cue™ solving the Social Media Analytics Challenges?
Social Cue™ was created with a vision different from 1 generation tools. The vision for Social Cue™ was to
provide a tool that answers retailers’ questions about their customers in a simple, easy to understand and
accurate way. Social Cue™ was built to speak the same language as the retailers using it.
Social Cue™ overcomes the challenges faced by first generation social media tools with the targeted goal of
focusing on what retailers need out of the tool and not what metrics around social media are easiest to gather.
This is a tool build for business; after only a brief introduction, business users have found they are able to not only
navigate it easily but get information that is immediately advantageous to the company.
Social media is only useful if it is easy to use and if the data is accurate: this is why Social Cue™ was created.
Social Cue™ is a Next Generation tool using patented machine learning and contextual pattern matching
technologies that offers the following advanced features:
Tweets referencing the company username vs. all tweets: Social Cue™ captures all tweets about a
retailer. Social Cue™ tracks every mention of the retailer, whether it references the company account
(@Username) or not.
Qualification of tweets: Social Cue™ uses a sophisticated qualification engine that “knows” the difference
between “#TheBay” when mentioned in the context of the San Francisco Bay area and “#TheBay” in the
context of Canadian department stores or “Gucci” when mentioned as a fashion brand vs. “Gucci Mane”
the recording artist. As a result, the tweets analyzed are relevant and accurate.
Specific topics (pre-defined): Social Cue™ has a very broad list of predefined topics. The default list
provides insights to what brands, apparel types, jewelry type etc. people mention. The list is just a starting
point – used to create a list which exactly reflects the product categories (and related products), brands
etc. the retailer using the tool would be interested in monitoring. That customized list (which can be
changed very easily any time) ensures that the retailer can see discussion in and around its more
important topics at a glance.
Person vs. account: Social Cue™ compares the different accounts (be it on Twitter, Facebook, blogs or
other social media sites) and profiles them in great detail. Using this profile, Social Cue™ can predict if a
Twitter account and Facebook account belong to the same person by matching a wide variety of complex
variables. As a result the tool provides a true view on the number of people interested in a retailer and a
consolidated profile of those people.
Profile vs. lack of profile: Connecting to opinion leaders is very important in social media. Social Cue™
profiles opinion leaders in a level of detail that allow for a view of not just which companies do they talk
about, but also what in general is of interest to them. Do they like to make pasta? Or do they have two
8|P a g e
[email protected]
amazing grandchildren they care about? Knowing such personal details allows for connecting on a much
deeper level.
Transparency vs. not clearly communicated methodology: Social Cue™ is 100% transparent in its
methodologies. Clients can learn how scoring and profiling is done to ensure that any analytics they collect
and decisions they therefore make are based on a solid understanding of how the tool works.
Changes in Social Media Channels: Social Cue™ is built on the solid foundations of a powerful Big Data
Insights engine that InfoTrellis has branded as AllSight. The AllSight platform is the driving force behind the
data quality, relevance filtering and contextual pattern matching of Social Cue™. AllSight is capable of
gathering and qualifying insights from any current or future Social Media channel. Social Cue™ is futureproof – ready and able to adapt to any emerging and receding Social Media phenomenon.
Graphical Summary
9|P a g e