Lingua Inglese II

COMMUNICATION

Reputation management: badmouthing can be defeated:

if you give people the opportunity of communicating with

you, you can share your side of the story in a polite,

professional way. At the same time, if someone is praising

your business then you can thank him and draw attention

to his kind words

Crisis management: e.g. Philadelphia Starbucks store

where two black men got arrested and the hashtag

#BoycottStarbucks was created becoming then viral.

Starbucks publicly apologized

Costumer services and support in real time

TARGET

—> Business can learn form social media what you like and

then make profit of it <—

Social ads are inexpensive way to promote a business

and distribute content 48

Retargeting: when costumers abandon cart chase them

with tools like Facebook Pixel (= a tool which is used for

online shopping websites by businesses; through it,

businesses track what you look for on the internet and

transmit the information to the social media where you

have an account, so that the social media will show you

an advertisement of what you looked for, because they

identified you as potential target and so they chase you)

Proof of ROI (Return On Investment) through analytic

tools (e.g. Google Analytics or Auditel) which tell you

whether that advertising/ads was clicked on or not.

2. Social media, business and grammar

What do social media and grammar have to do with each

other? There’s something called ‘semantic web’ (semantics

is the discipline that studies the meaning of words), an

extension of the internet which tries to make sense of the

internet. The amount of information available online is

huge.

Semantic web is:

• A mesh of data associated in such a way that they can

easily be processed by machines instead of human

operators

• An extended version of the existing World Wide Web, it

represents an effective means of data representation in

the form of a globally linked database

• Conversion of the presently available Web of

unstructured documents to a web of information/data

17-04-19

Semantic web transforms unstructured data (random

information) to a corpus (pl. corpora) = a collection of

written or spoken material stored usually on a computer or

online and used to find out how language is used.

Corpora and annotation

• Unannotated corpus (=we collect all of the texts we need)

- simple plain text or raw text

- the linguistic information is implicit, by looking at the

49

corpus we might not know the relevant phenomema

e.g. no explicit representation of a present as a noun

• Annotated corpus

- no longer just text

- real repository of linguistic information. The relevant

linguistic information is now explicit

e.g. present as a noun, adjective or verb

Corpus annotation exam!!!

What is it?

- The process of adding interpretive linguistic information

to an electronic corpus of spoken and/or written

language data

- Broadly, it also refers to the results of the annotation

process

How are corpora relevant to our lives?

Hashtag

Hashtags are semantico-syntactic constructs used across

various social networking and microblogging platforms to

enable users to start a topic-specific discussion or classify a

post into a desired category. You annotate your own

content for other people. You can’t put space or commas in

between.

Hashtags, but also everything we say online, is transformed

in corpora and analyses by businesses.

—> Semantic analysis of hashtags could therefore help

understanding and extracting important information from

microblog posts. This could be manually, automatically, or

semi-automatically but automatic semantic annotation

methods are required, in order to make sense of the

millions of messages posted daily on facebook, twitter,

likedIn,…

Why doing semantic analysis?

• Extracting useful information from large bodies of

unstructured data. Extensive amount of data from almost

every consumer are useful, but also enormous amount of

work managed with semantic analysis 50

• Costumer service and search engines - semantic analysis

finds relations between problems and solutions, and

automatizes costar service + analysis of search engine

searches - deliver relevant content to its users

• Uncover eating in text through hashtags but also broader

text analysis. Nowadays it is able to ‘clear noise’

(understand despite spelling mistakes, abbreviations and

mixed languages)

e.g. talk to books, semantris

Sentiment analysis (SA or pinion mining) exam!!!

A feature of text analysis for information regarding how

costumers are reacting to their products and services.

Sentiments refer to human attitudes, opinions, and

emotions, which are qualitative objects. To identify the

underlying tone of the expression, SA uses natural

language processing (NLP).

Natural language processing (NLP) is the ability of a

computer program to understand human language as it is

spoken. NLP is a component of artificial intelligence (AI)

Reasons why you should not ignore sentiment analysis

Reason 1 Reason 2

Reason 3

| | |

Drives proactive measure of your improve

product

business solutions marketing campaigns quality

Sentimental analysis doesn’t go very subtle, it polarizes the

opinions between positive and negative (or neutral) —>

POLARITY, a metric to a piece of text - how positive and

negative - value ranging form +1 (extremely positive) to -1

(extremely negative).

The simplest algorithm goes by the dictionary e.g. ‘all you

need is love’ - positive; ‘I still haven’t found what I’m

looking for’ - negative.

Advanced algorithms do machine learning = capture

nuances e.g. bloody may be positive (bloody excellent).

51

These algorithms learn from large data sets, often already

graded by actual people as to what constitutes positive and

negative. You can teach the machine not only to identify

the single word, but also what stays before and after it and

combinations of words that could be relevant.

Positive Negative

Costumer service

By looking at reviews you may

Timely Don’t fix

Nice issue

find words related to positive

Helpful Rude

opinions

or to negative ones

Machine learning is an application of AI that provides

systems the ability to automatically learn and improve from

experience without being explicitly programmed. Machine

learning focuses on the development of computer

programs that can access data and use it learn from

themselves

3. Business writing in new media

Create and apply brand standards 52

• Check the mission statement: this should outline what

you hope to accomplish trough your social media strategy

• Audience: try to study and find out data about who your

audience members are and what they are looking for

from social media content. ‘Who are you writing for?’

• Brand voice: describe or find out whether the company is

a casual communicating one or a professional one …

• Tone: it depends on the brand voice. This will help

determine what types of content to post, e.g. helpful.

Authoritative, humorous,…

• Branding: determine any important brand elements, e.g.

spellings, abbreviations

• Message: guidelines concerning which types of

messaging to post on specific platforms, e.g. twitter,

linkedIn and facebook

—> Shareability of a post:

- Openable from mobile phones (mobile-friendly)

- Catchy titles

- Visible share buttons

- Hashtags and multimedia

—> Speak a language that your audience is familiar with,

so study, read, survey them

Social media writing tips

1. TONE: casual, brief, adjusted. Even if the company has

an authoritative language, you should soften it.

- Social media accounts don’t have to be replica

websites, coherent in content but in a more casual tone

- Social media platform needs specific adjustments

e.g. The Agorapulse team has compiled a list of the best

social media writing tips. Click below to read them

(linkedIn); need to improve your social media writing

skills? Check out our hot tips (facebook, which is more

about sharing).

Overall, avoid lengthy sentences and long words (e.g.

optimal FB length, 40 words max).

Avoid colorless ‘business-speak’, try to give some

fascination.

Teaser captions are short and applicable 53

15.CTA Call To Action

- you shouldn’t be constantly promotional

- elicit responses (watch a video, share comment, etc.)

- you need to be punchy and direct e.g. ‘watch now’ or

‘read more’

- ask short question or make an impactful statement to

invite people’s attention

- don’t confuse by asking for too many actions, one

thing per post

16.Keep it personal

- use pronouns such as ‘we’, ‘I’, ‘me’ and ‘you’ as

opposed to ‘the company’, ‘the client’ or ‘the team’

(which are though fine on linkedIn) e.g. post by Barclays

UK on facebook: we’ve created 5 top tips to help your

finances - we instead of Barclays UK

- do not lecture the audience, because most people

come across lectures everyday

- write as though you are talking personally to each

follower, do not be too general, to create a sort of

dialogue

- include data in your posts if it’s needed e.g. 85% of

our costumers surveyed love this new shampoo

- tailor your language to your audience age, gender,

interests,…

17.Great responses

- understand your brand voice and how you will respond

to fans on social e.g. cheeky brand that likes to play

with fans? Straight-to-the-point brand that answers

questions without chatting? - Post by Clayton Hove on

twitter and answer by Official smart USA

- same style as brand - the audience enjoys the same

brand experience throughout

- friendly, polite and informative

- pass the client to your support team if necessary but

make sure it is communicated properly

29-04-19/01-05-19

No lezione

54

06-05-19

3. Sample business writing online 1: trip advisor

- “There was poo in the kettle”

- “haunted”

- “Found a dead cockroach in my gin and tonic”

- “I’d rather sleep on the sleep”

- “Beach is too sandy. But clear water”

Some reviews can be justified, other are absurd

Why respond to a negative review?

Bad reviews may deter travelers from booking the

hotel/restaurant. By responding:

• You show you’re listening and you care

• You change perceptions

• You reassure other travelers that the same thing won’t

happe