Uivei deg

CorsodiLaureaMagistralein IngeneriaInformati a

Analysis and implementation of

so ial networking methodologies

in a quality management system

Relatore: Ch.mo Prof. MatteoBerto o

Correlatore: Ing. Mauro Fran hin

1 Introdu tion 1

2 Part I - Customer Satisfa tion& So ial Networks 3

2.1 So ialNetwork . . . 3

2.1.1 Fa ebook . . . 4

2.1.2 Google plus . . . 8

2.1.3 Twitter . . . 11

2.2 ITtoolsfor ustomer satisfa tion mining . . . 14

2.2.1 Chat . . . 15 2.2.2 Ti ketTra ker . . . 18 2.2.3 Web Area . . . 19 2.2.4 Dire tInput. . . 20 2.2.5 Twitter . . . 21 2.2.6 Forum . . . 24 2.2.7 Systemdesign . . . 26

2.3 Amessaging tool forte hni al or ommer ial support . . . 33

2.3.1 Obje tives . . . 33

2.3.2 Components. . . 34

3 Part II - Te hnologies 45 3.1 jQueryand jQueryUI . . . 46

3.2 jQueryMobile. . . 47

3.3 Mantis BugTra ker . . . 48

3.3.1 Authorizationand A essLevels . . . 49

3.6 Sma k API . . . 71

3.7 Java Servlet . . . 72

3.8 JSP . . . 75

4 Capter III - Implementation 77

4.1 Te hni al Analysis . . . 90

4.1.1 Communi ationbetween ustomer appand Mantis . . . . 90

4.1.2 Communi ationbetween analysistooland Mantis . . . 97

Bibliograa 99

List of Tables 100

Introdu tion

The ompanywherethethesisworkwasdevelopedisMidaSolutionsS.r.l. . The

ompany was founded in 2004, by a team of tele ommuni ation experts, with

the missionto provide value added innovative te hnologies for ommuni ation.

Midafo ushasbeensin ethebeginning te hnology andpeople. MidaSolutions

providesunique expertise anda ompletesuite of Voi e Appli ationsand Value

AddedServi es,withthegoaltodenitelyimprovetheTelephonyInfrastru ture

fun tionalities.

InMar h2012,DNVhaso ially ertiedMidaSolutionsQualitySystemin

a ordan e withISO9001. The erti ation ofthemanagement system

demon-stratesthe ontinuouseortofthe ompanyina tivatingpro essesof ontinuous

improvement, aimingat developing thebusiness,inalogi of sustainability.

For a ompanyinsert anew produ t/servi e into themarketis a signi ant

burdenfromboththeorganizationalandnan ialaspe ts. Thisphase,however,

isonlythe endofavery omplex pro ess. Ingeneral itispossible tode ompose

this pro essina su essionofstages, the hoi e of whi h dependson many

fa -tors. Withoutgoinginto detail,we an saythat,forthesu essofthe ompany,

it is ne essary that ea h of the a tivities identied is arried out in order to

maximize theprobabilityof su essand,simultaneously,redu etimeand osts.

Imagine,forexample,tostru turethepro essinve-step: marketing,design,

manufa ture,produ tionandsale. Editingaproje tinvolvesa ostthatdepends

ontheextentof orre tionneededandthetimewhenthisde isionistaken,that

is,dependingonthestagerea hed inthepro essforthe onstru tionofthenew

produ t: theamountrequiredfora hangemadeintheearlystagesofthedesign

issigni antly lessthan thatrequired for themodi ationof aprodu t already

on sale.

Thepro essofinsertinganewprodu tinthemarketmusttakeinto a ount

that the ompany, at every stage, takes into onsideration the needs of the

ustomer and a hieving what he expe ts. In on lusion, this an be translated

in bring the ustomer's voi e and letit guide you throughout the development

pro ess, fromdesign to produ tionof the good or servi edelivery.

With the oming of Web 2.0, in parti ular through to the spread of so ial

networks, the level of intera tion between web and user in reased signi antly.

Observing this in rease, ompanies have been able to use these new tools to

establishdire t ommuni ationwithits ustomersinorderto have instant

feed-ba k on every produ t or servi e, and thus be able to evaluate and possibly

modify their strategies a ording to the level of satisfa tion manifested. Su h

tools arein reasinglytakingmore signi ant importan e be ause theyhave

be- ome thenew way throughwhi h thevoi eof the ustomer an be olle ted by

the ompany.

Buthowtousethemtoextra tthelevelof ustomersatisfa tion? Thisthesis

Part I - Customer Satisfa tion &

So ial Networks

2.1 So ial Network

If we wereto provide a trivialdenition of the phenomenon of so ial networks,

we an say thata so ial network is a onne tion between people with dierent

relationships: work, friendship, family. The Internet version of so ial networks

amplies the on eptofsharing and parti ipation anddevelops one ofthemost

advan edforms of ommuni ation based on thedesign of relational maps. The

onstru tion ofthese relationships arestru turedinthemanner oftheso- alled

ƒweb2.0‚,resultofthedesignofweb sitesandappli ationsthatputthe ontent

generated bythe useror not, inthe handsofthe onsumer.

Greatly simplifying the meaning of so ial networks, we an imagine it like

a omplex platform where you an reate a network of people withwhom you

arein onta t(atdierent levels) sharing information and ideas withthemand

followingthosetheyenter. Inpra ti eitispossibletohaveaprole(a ardwith

information about who we are), you an insert messages (but also multimedia

ontentlikephotos, videosandlinkstowebpages)andread, ommentandshare

whatotherusers publish. You anusually tellthesystemwhohasorhasn't the

ability to read our messages and information thatwe publish. In on lusion,a

so ial network is a powerful tool to keep intou h withpeople, to learn and to

inform.

One ofthereasonsthat drivesa ompany to useso ialnetworks istheneed

to listen whatpeoplesayabout the ompanyand see whi h is thebrand image

among ustomersor users ingeneral.

The hoi eisneverspeakwhetherornotaspe i brand,produ torservi e,

buttoparti ipate ornotina onversation. Oneofthebiggestmistakesmadeby

itsbusinessshouldbetranslatedastheabilitytolistentowhatotherssayabout

the ompany,what aretheper eptions thatothers have of itsmarket and what

ompetitors do.

In this hapter we will analyze the hara teristi s that have made famous

the main so ial networks inorder to identify and analyzethe featuresthat an

beusedfor theextra tionofthelevelof ustomerssatisfa tion. Thebestknown

and most used so ial network is Fa ebook, followed by the no less important

Twitter andGoogle+.

2.1.1 Fa ebook

Fa ebook is a so ial network laun hed in 2004, owned and managed by the

orporation Fa ebook, In . The site is free and takes prot primarily from

personalizedadvertising for ea h user, oering produ tsand servi es whi h are

oherent withthe a tivitiesperformedby the userwithin theso ial network.

Theinteresting aspe tsof thisso ial network arethefollowing:

ˆ so ial marketing on the platform. One ofthemain strengths ofthe so ial

networks are thetools of so ial marketing that allow planning of real

ad-vertising ampaignstargetedto ertain ategories of ustomers. Themost

important of thesetools isFa ebookAds.

ˆ analysisof ampaignresults. Inadditiontomakingpossibletherealization

of advertising ampaigns, Fa ebook provides the tools ne essary for the

analysisof the results ofthese ampaigns.

ˆ integration with external sites. You an use your Fa ebook a ount to

intera t withwebsites thathaveintegrated insideitselftheso ialnetwork.

This platformis alledFa ebookfor Websites.

ˆ openness to developers. Fa ebook has made available to the developers

theAPIsthat allow the reationof appli ations.

Fa ebook Ads

ƒWith Fa ebook Ads, users an learn about brands and businesses through

trusted referrals fromtheir friends on Fa ebook. Advertisers an onne t with

users by reating a presen e on Fa ebook and targetingthe exa t people they

want‚

The presen eonFa ebookisastrategi element for ompaniesthat want to

exploit the full potential of Web 2.0 for their business: themost popular so ial

network now olle ts a pool of over one billion users worldwide with a growing

trend. The reation of o ial pages for the ompany and dis ussiongroups on

spe i topi s that a hieve indire tly a larger number of users, are among the

main driversof whatis alledso ial marketing.

Fora ompanybepresentinFa ebookhasanumberofadvantages,themain

onesare:

ˆ in reaseknowledgeof itsbrand

ˆ update the lients on produ ts, promotions and events

ˆ promotetheir business indire tlythrough aggregationthemes

ˆ getfeedba kfrom thefansto knowtheir opinionsand their needs

ˆ take advantage from viral marketing, en ouraging fans to involve their

friends

In the o ial press release of the servi e were ited some large ompanies

thathave seizedthe newopportunitiesopeningupandpromotingafanpageon

Fa ebookinorderto emphasizetheimportan eof thetoolfor business

market-ing.

The ads on Fa ebook appearalways in theadvertising spa e lo ated to the

right of the page,and aredivided into thefollowing elements:

1. Bodyof thead, onsistingoftitle, image and brief des ription.

2. Thenames of the friendswhoarealready fansofthepage advertised.

3. Link to the page to be ome a fan or external link to the website being

advertised.

Through the use of Fa ebook Ads you an dene the exa t hara teristi s

of the re ipients of the insertion, by setting up demographi or psy hographi

ˆ Gender,age, level of edu ation,sentimental situation andsexual

orientation.

ˆ Keyword, al ulated based on the prole information su h as interests,

a tivities, books,favorite movies, et .

ˆ S hool or University.

ˆ Conne tions, allows you to onta t those who are already fans of the

page, or who use a parti ular appli ation, or who has joined a ertain

event.

ˆ Friendsof onne tions,allowsyouto onta tonlythosewhoarefriends

with those who are already fans, or who use a parti ular appli ation, or

who have joined to a ertainevent.

ˆ The birthday, displays thead ontheuser'sbirthday.

ˆ Geographi alLo ation,Fa ebookdeterminesauser'slo ationbyIP

ad-dress,but you analsousetheaddressthattheuserhaspossiblyin luded

inhis prole.

One of the drawba ks of this approa h an be identied, inaddition to the

real reliabilityof the proles,even inthea tual ompleteness of these.

Afterhaving identied the hara teristi softhe re ipientsof theadvertising

message,Fa ebookisabletoprovide,inrealtime,aroughestimateofthenumber

ofusers whowillseethead. Inthis wayitispossibletorealize thea tualviews

that an be obtained with ertain lters and if too restri tive, hange them to

broaden the target audien e.

Fa ebook Insights

Fromthedenitionoftheme hanismthatregulates theads,it anbesaidthat

the obje tivesofweb advertising onFa ebook an bemainly:

1. Divert users onits website.

2. In rease the numberof fansof thepage.

Start an advertising ampaign through the so ial network is a ne essary

step, butnot su ient toensurethespreadoftheknowledgeofyour produ tor

servi e. Indeed,itis ne essaryto monitorthea hievements ofthe ampaign.

To makethis possible,Fa ebook oers ompanies astru tured analysistool

that allows you to measure the exposure, the a tions and behavior of users in

ˆ information about fans: number, gender, geographi aldistribution.

ˆ intera tions: total numberof omments, wall posts and preferen es.

ˆ intera tions for post: average number of omments, wall posts and

appre iation generated byea h ontent.

ˆ quality of the post: s ore that measures the interestof the ontentsof

the page byFa ebook users.

ˆ dis ussionpost: the numberofdis ussiontopi s reatedbytheuserson

the page.

Fa ebook Insights summarizes the general trend of the listed variables,

al-lowing alsotodeepenthe trenddataandthetypesofintera tions, andgenerate

graphsofdetailforea h itemthatyou want to analyze. Anexample ofas reen

obtained throughtheuseof Fa ebookInsightsisreported inFigure2.2.

With this powerful tool is possible to analyze in more detail the quantity

and qualityof intera tions o urred on the ompany prole,noting thelevel of

involvement exer ised bypost on the fans, and then ensures thesu ess of the

entire so ial marketing strategy adopted.

2.1.2 Google plus

Google

+

is the so ial network of the famous sear h engine laun hed June 28, 2011. It has a number of features that distinguish it and that allowed him to

grow qui kly despitethe alreadyestablished ompetition.

Figure2.3: Google pluslogo.

Themain featuresoered are:

ˆ Cir les. Theyareoneofthemostimportantinnovationsintrodu ed bythe

so ial network thatallows to sharerelevant ontent with theright people

and nd ontent inwhi h you areinterested.

ˆ

+

1button. Itisusedtoreportpubli ly a ontent appre iatedbytheuser.

ˆ Hangouts. This feature allows you to make video onferen e of up to 10

people.

ˆ Chat. It allows you to ex hange information between multiple people,in

real-time.

Belowwebrieydis ussedthe possibleusesof the ir les andthe+1button

in the orporate ontext analyzed in order to verify the usefulness of their use

to a hieve the prexed obje tives.

The ir les

ƒCir les make it easy toshare with the rightpeople‚

Themain purpose of the Google+ ir les isto manage groups of people

a - ording to the real so ial onne tions of life. It is useful to separate the family

be-thentode idetowhommakeknown ertain ontentandthosewhomkeepthem

hidden. In a business ontext this instrument an be used invarious ways, for

example:

ˆ Sharespe i ommuni ations between the onta ts thathavein ommon

a distin tfeature.

ˆ Conta tdire tlytheir onta tswithquestionsandsurveystoobtain

infor-mation on thebasis of whi h reate the ir les. Based on theanswers to

the question ¨What produ t do you use?¨You an, for example, add the

usertothe ir le ontainingalltheusersthatusethatparti ularprodu t.

ˆ Create ir les based on the role played by onta ts in the ompany (for

example exe utives, ustomers, employees). This allows you to share

in-formation and arry out dis ussions on spe i topi s of interest to users

and allow youto gain newinformation about them.

ˆ Involve onta ts asso iated to ustomers to get feedba k, intera t with

themand know thembetter.

Figure 2.4: Cir les inGoogleplus.

Analysis of fun tioning Summarizing the fun tioning of the ir les we an

state thatthis instrument takes areof showing the ontents only to users who

belongto a ertain setand onsequently hides themto all theothers.

From the implementation side this an be translated in the asso iation

listof usersidentiers, usedfortargetedshares: Whenyoude ide to share

on-tent with a parti ular ir le, you go to set thevisibility of that ontent only to

membersof the ir le. Exploiting su hstru ture isalso possible to dynami ally

lo ate the ir les of whi h a given user is a member, verifying his presen e in

ea h of them. The s heme ofsu h me hanismis showninFigure 2.5 .

Figure2.5: S heme of ir les.

+1

button

Many websites have already integrated into their pages the Google +1 button

anditisusedtosignaltotheworldtheuser'sappre iationforthatpage. Besides

in reasing the ounter asso iated to the button to in rease the index of overall

enjoyment of the page, you an share a link to your Google

+

prole, possibly a ompanied by a brief omment. The last great advantage of this button is

onstituted by theinuen e that the li ks made on it have on the position of

the pagewithin thesear hengine.

Figure2.6:

+1

button ofGoogle.

beaverypowerfultoolbe auseitallows youtohavearstfeedba konthelevel

of satisfa tiongiven by ustomersto thatparti ular produ tor servi e.

2.1.3 Twitter

Twitter is a free so ial networking and mi roblogging servi e, reated in 2006,

whi h provides users a personal page updated via text messages, alled tweets,

with a maximum length of 140 hara ters. The main feature of Twitter is the

speed,that isthereal-time natureof theinformation.

Twitter than other so ial networks, it is mu h easier at fun tionality level,

but this does not mean that a ompany an not use Twitter for your business.

Infa t, it isone of thetools that ompanies an useto bridge thegap between

the ompany and its ustomers or fans and if used intelligently an oer great

benets to the brand and its produ ts. The best known example of su essful

use is that of Dell, whi h oers its produ ts on oer (out of produ tion and

inventory) through the so ial network. The ompany announ ed that it has

earned, usingTwitter withoutanyparti ular strategy,agureequal to about $

7million. Thisisnot a parti ularly signi ant gurefor Dellbut itproves that

theTwitter audien e is a tive onthis platform.

Figure2.7: Twitter's logo.

In the following we will analyze the hara teristi s fun tionality of Twitter

onsidered tobeparti ularly interesting, namelythehashtag and thepromoted

Tweets.

Hashtag

In Twitter the use of the hashtag fa ilitates ategorization, sear h and

aggre-gation of messagesand dis ussions. The hara ter ƒ#‚before spe i keywords

allowsthe transformation ofthese words ina tive links that reate digital

om-munities gathered around a topi . In this way it is possible to olle t all

the re ent posts mentioning the same hashtag in a single page alled hashtag

page. In Figure 2.8 is shown an hashtag page generated from the hashtag ƒ#

GoodThingsInTheWorld‚.

Fora ompanyisimportantto reatespe i hashtagforthebrandname,for

theprodu tand/or for an event, to whi h must beadded theshort des ription

Figure 2.8: Example of hashtag page.

but ifitisanew hashtag,it isre ognizedasoriginaland Twitter goesto reate

a new hashtag page where they will be pla ed and maintained all tweets that

ontain the keyword.

The use of this tool an be designed to a hieve two obje tives: one is to

be understood as a me hanism for dialogue with the user, the other as a tool

to en ourage so ial ollaboration to deliver value. From the fusion of the two

obje tivesispossibletoimplementanee tive strategythat ouldoptimize and

integratetheinternalpro essofti ketmanagement,respondingtotherst lient

requesting information and, at the same time, ommuni ate to all others who

have the same problem. In this way, also users may provide information or

propose theirownsolutions, thus to integrate thework of ustomer servi e.

Promoted Tweets

One of the main problems for those whopromote a produ tor servi ethrough

so ial networks is the speed with whi h the post on the boards of Fa ebook,

and even morein theTweet, run away. Thisfor es us to frequently updatethe

ontent soasnot to fallinto invisibilitywithin afewse onds andmaytherefore

be onsideredspammers.

ofasear h,inthetimelineofthepromoterorinthatofusersandaboveallremain

visiblefor dayswithout being hiddenby themessagesofother users.

Another interestingfeatureistheabilitytoidentifythere ipientsofits

Pro-moted Tweets bygeographi al riteria (Italy,Fran e, Japan, et ..) or by se tor

(auto,nan e, politi s,travel, et ..). AnexampleofaPromoted Tweetisshown

inFigure2.9.

Figure2.9: Example of Promoted Tweet.

Not being banner but tweetin all respe ts, to thePromoted Tweets an be

answered,they anbeaddedasafavorite,andaboveallyou anretweettoother,

2.2 IT tools for ustomer satisfa tion mining

ƒA satised ustomer is the best businessstrategy of all.‚

For a ompany it's important to onstantly apture the per eptions that

people have of the brand and its produ ts sin e this translates into in reased

prots andredu ed osts. Buthowto olle tthis information? Nowadays there

are various te hnologies that ombine together help businesses to olle t the

moods of their ustomers. In this thesis work we will try to identify a set of

tools to a omplishthis task,and amongall thesewill beimplemented one.

Thenalgoalistoprovideanestimationofthelevelof ustomersatisfa tion

abouttheprodu tssoldbythe ompany. Todothisitisne essarypro eedintwo

phases: therstisto olle tdataonhowtheuserintera tswiththeinstruments

atitsdisposal,andthese ondmustpro essthedatapreviously olle tedinorder

to extra t the estimatedvalueof satisfa tion.

The traditional te hniques thatallow therealization of these steps, su h as

llinginquestionnaires, arenot enoughto olle tmeaningfuldatabe ause they

are onsidered too invasive and the user usually tends to ignore them or give

unreliable answers.

The need therefore arises to design tools that are able to extra t the

ex-pressed sensations by the analysis of the a tionsperformedby theuser. These

instruments must play their role while providing an additional servi e to the

ustomer,thus fa ilitating theintera tion between ustomerand ompany. The

mainsour eofinformationisexa tlythisintera tionbetween ustomerand

om-pany, be ause the manner in whi hit takespla e and thetools usedarepoints

inwhi h an happen the gathering ofinterest data.

The identi ation of the most suitable tools, requires an initial in-depth

analysis of the intera tion that orrelates the ustomer to the ompany. The

lient intera ts through representations of produ ts (pur hased or potentially

pur hasable) provided by the ompany through various hannels. It is

there-fore important to identify the modes of intera tion that are tra eable. Tools

that enable the olle tion of per eptions about ustomers need to reate new

ommuni ation hannels parallelto theexistingones.

The hannel more easily tra eable is the ompany website, whi h plays the

role of interfa e between the two entities. Through it, the ustomer an see

the onlinemanuals, the page of Frequently AskedQuestions(FAQ) andbrowse

the list of produ ts. With these simple a tions, using a spe ial omponent for

tra king the paths followed, it is possible to extra t a rst set of information.

This information an give us an estimate on the produ ts that most ae ting

theusersandontheemotional stateofthe ustomer. Infa t,ifausernavigates

Figure2.10: Intera tions between ustomer and ompany.

mannerandprovideindire tper eptions,ieofwhi hhasanestimatedvalueand

thereforeun ertain.

A more dynami way to obtain information onsists in using te hniques of

so ial networking, whi harea olle tionoftools thatallowamore dire t

inter-a tion between ustomer and ompany. Some of these tools appear among the

hara teristi features of so ial networks analyzed above, su h as the "I Like"

button, the systemfor managing omments, hat,post, et . The data olle ted

withtheuseofthesete hnologiesaremoreexpli itandallowtogivemore

truth-fulsatisfa tionestimates,sin etheyexpressfeelingsdire tly. Itis leartheneed

to reate alternative ommuni ation hannels that an host thesetools.

The immediate s enario that emerges is the possibility to realize a system

apableofprovidingas ore obtainedbyextra ted per eptions,weighed

a ord-ing to the used tool: the s ores obtained from instruments that provide dire t

per eptions will have a greater weight than those that provide indire t

per ep-tions.

The next step is to identify a set of tools that are able to olle t the

in-formation of interestin analternative waythan traditionalmethods using asa

referen e model the hara teristi s identied in the analysis phase of the most

popularso ial networks. Ea h toolis des ribed through theuse of at leasttwo

summarytables:

ˆ the rsttableisthe tableofa ess,whi h isusedto spe ifythe ategories

ofusers to whom the toolis addressed;

ˆ the other tables spe ify the a tions that the various types of users an

perform.

2.2.1 Chat

One of the rst tools identied to a hieve the obje tives set is the hat. Chat

isanadditional servi e,availableonlyto registered ustomers, halfwaybetween

Customerswhoneedto ommuni atewiththe ompany anusethistool. In

ordertoavoidahighworkloadforthe ompany,itwasde idedtooertheservi e

only to the ategory of ustomersregisteredat the ompany(forexample, only

thosewhohavealreadypur hasedaprodu t)andforalimitednumber oftimes.

Through this instrument ustomers an open hat after hoosing thereason for

whi h they de ided to use it and the produ t of whi h they wish to dis uss.

Subsequently, the hat is opened and the onversation begins. With the aim

to in rease ee tiveness and speed of ommuni ation, have been introdu ed a

numberoffeatures,forexamplethepossibilitytouploadimageslikes reenshots

of error messages.

InFigure2.11isshownthestoryboard ofthe ustomer hat. The basi idea

is to provide an intuitive and easy to usetool to promote a tive parti ipation.

Thisimpliesamaximumredu tionofthenumberofs reensandbuttonsinorder

to avoid unne essary steps. As shown in the gure, the user need only sele t

the motivation and the produ t he wants to dis uss, and inthe next s reen he

an immediately ommuni ate withtheemployee. Witha few li ksthe hat is

started.

Figure 2.11: Storyboard of the ustomer hat.

Eventheemployeesofthe ompanymusthavea hatthroughwhi hrespond

to various ustomers. They need for additional features to keep tra k of hat

In Figure 2.12 is shown the storyboard of the employee hat. This hat is

slightly more omplex than the ounterparty targeted at ustomers be ause it

mustmanagemultiplesessionsandallow lassi ationofea hofthem. Therst

s reen shows thelist of ustomers registered to the servi e, divided into online

users and oine users. When a user starts a new hat session, his appli ation

sends a noti ation that the employee side appli ation is apable of handling.

The employee li k on the button with the name of the user and opens the

se ond s reen inwhi h begins the ex hange of messages. When the hat ends,

the employee press the button to resolve it. This a tion opens the evaluation

s reen where is shown the dis ussion just ended and the possible votes to be

assigned. Inthe design phase, hasalsobeen in luded theabilityto assignmore

votes withinthe same hat session, be ause while hatting you an ta kle more

than one issue. Finally, the last s reen shows the ability to onsult the hat

already losed in order to save time, proposing solutions of problems already

solved.

Table 2.1: A esstable to hat.

UserType A ess

NotRegistered NO

Registered YES

Employee YES

Table 2.2: A tionsof the hat availableto ustomers.

A tion Des ription

Motivate Givesa reasonfor openingthe hat

Find Opensanew hat session

Show Loadsand sendanimage

Resolve Closes the hat session

Express Writesinthe hat

Vote Providesavoteonhis levelofsatisfa tion

Table 2.3: A tionsofthe hat available to employees.

A tion Des ription

Open Opens anew hat session withthe ustomer

Show Loads andsendmeaningful data(e.g. diagrams,images)

Resolve Closes the hat session

Reply Writestheanswerinthe hat

Consult Views previous hats

Modify Changesprevious hat

Vote Provides avote onlevelof ustomer satisfa tion

2.2.2 Ti ket Tra ker

One of the tools a hievable in a business ontext is the ti ket tra ker. The

basi idea is to en ourage and fa ilitate the intera tion between ustomer and

ompanytakingasa ben hmarkthe immedia yof Twitter.

Theti kettra kerthus allows youto sendreports of various kinds,in

refer-en eto aspe i produ t,inafeweasy steps. Thisinformation isstored inthe

form of Mantis bugtra ker ti ket (also alled issue). Through this me hanism,

the lient is a tively involved and ontributes to the improvement of the

prod-u tsdevelopedandtheservi esprovided. Su has hat,eventheti kettra keris

providedonlytoregisteredusers. Theappli ationisdesignedformobiledevi es,

allowing users to maketheir reports at any time fromtheir mobilephones.

The ompany ounterparty of the appli ation allows to manage this

infor-mations andassignthemassessments,whi h willbeanalyzedlater bya spe i

ˆ viewthenew messagesre eived (divided by ategory);

ˆ sele tthem one at atime;

ˆ reply(ifne essary);

ˆ givea voteon the levelof ustomer satisfa tion per eived.

Table2.4: A esstableto ti kettra ker.

User Type A ess

NotRegistered NO

Registered YES

Employee YES

Table2.5: A tionsofti kettra keravailable to ustomers.

A tion Des ription

Motivate Givesa reasonfor openingthenew report

Express Writesthereport

Consult Viewsprevious reports

Table2.6: A tionsoftheti kettra ker availableto employees.

A tion Des ription

Reply Writes thereporting reply

Consult Views previousreportings

Vote Providesa voteon levelof ustomer satisfa tion

2.2.3 Web Area

A ru ial tool for the tra king of user a tions is ertainly the ompany web

area. Through it, theuser an knowthe produ tsoered, solve basi problems

and stay informed about the news. The web area is onsulted by all types of

ustomers, iethosea tual andpotential and thus brings inherently a verylarge

information ontent.

Inadditiontothekindsoffavoriteprodu tsispossibletore onstru tthe

nav-igationpathsmadebyindividualusers,andfromthesetrytomakeassumptions

about their emotional state: afast navigation may involve frustration, be ause

poolof informationobtained fromthis instrument delivers pre ise,but not easy

to analyze, indi ations. In fa t, with only this information, you an get the

simplied assumptions, that are hard to verify. To fa ilitate the phase of data

pro essing, itisusefultoexploitthestru tureofthewebarea,iebydividingthe

Company'swebsiteinthreemain parts: rootnode,intermediatenodesand end

nodes. With this further subdivision, statements an be made more pre ise: if

theuserremainsstationary onaspe i terminal nodeandthenleavesthesite,

thenmostlikely willhehasfound whathesought,onthe ontraryifheremains

stationaryon anintermediatenode foralongtimeandthenleavesthesite,this

may meanthathe didnot ndwhat he waslooking for.

Therealpowerofthisinstrument howeverisinits ompleteness,infa titis

able to provide also those omponents for dire tevaluations alreadymentioned

above. Theuseraswellasbrowsethroughthewebsite, analsomakejudgments

through the additionof ommentsto thepost, li ks onthe"+1"at anews, he

shares page asso iated to a produ t in his Fa ebook a ount, et . Due to its

ompleteness,you an overthe majorityof feelingsexpressedbytheuser, thus

notjustlimitedtoestablishonlyinterestorindieren etowardspe i produ ts.

Table 2.7: A esstableto web area.

UserType A ess

NotRegistered YES

Registered YES

Employee NO

Table2.8: A tionsofwebarea available to ustomers.

A tion Des ription

Sear h Browse inweb site

Leave Leavesthe web site

Vote Intera tswith votingme hanisms

Motivate Givesthe motivation for thenewreporting

Express Writesthe reporting

Consult Readsweb page ontents, downloads do umentation

2.2.4 Dire t Input

Thedire tinsertionofsignalingsisusedto arryoutdire tlyanytypeofreport.

The instrument hasbeen designed asintegration ofthewebsite. Thebasi idea

is to provide a tool that enables in a few simple steps the insertion of votes

theaddition ofafeature,hewillhavetopressthespe i buttontoopenaform

where he an:

ˆ sele tthereason for reporting;

ˆ writethe ontents ofreporting;

ˆ submitthereporting;

ˆ votethesele ted produ t.

In this way the details of reporting, su h as the name of the produ t, will

be automati ally lled by the appli ation, based on the terminal node from

whi h is arried out the insertion. For messages of a generi nature, the user

must simply li k the buttonasso iated to thefeature on thehomepage of the

ompany website.

The message an be forwarded to the ompany, whi h will use a spe ial

softwareor simply emailto re eive noti ation. Signalings by dire tinput an

be imagined as omments already lassied based on the ause that generated

them.

Figure 2.13 shows the storyboard for dire t reports. On e the user arrives

at the desideredprodu tpage,he li ksthe buttonto thedire treportingsand

opensapopupwindow withwhi hhe an intera t intuitively.

Table 2.9: A esstable for dire t input.

User Type A ess

NotRegistered NO

Registered YES

Employee YES

Table2.10: A tionsofdire t inputavailable to ustomers.

A tion Des ription

Vote Givesa voteon produ t

Motivate Givesthe motivation for thenewreporting

Report Writesthereporting

Consult Viewsprevious reportings

2.2.5 Twitter

om-Figure 2.13: Storyboard for dire treportings tool.

related to produ ts in the produ ts website page and then, after a possible

l-tering operation, publish them in the Twitter ompany a ount automati ally

using spe ialshashtag.

Thehashtag me hanism,infa t,providesaverybasi ,butatthesametime

ee tive, lassi ation be ause all the omments arein hronologi al order and

thenyou anrebuildentire onversations. Inthetransitionbetweenthe ompany

websiteandthe publi ationonTwitter we anevaluate the ommentsonthey

or save them for later analysis. In this ase it is ne essaryto reate a suitable

softwaretoolfortheemployee,whi hisabletotakethe ommentsfromthesite,

view them, lassifythem and eventuallypostthem onTwitter.

So the integrationof thetwo hannels,that relatedto the ompany website

and that on Twitter, through an appli ation bridge enables theanalysis of the

omments. The main strength of the instrument lies in the simpli ity with

whi h the user an intera t withit. Con eptually, the me hanism for inserting

omments sostru tured, it is halfway between a forum anda hat, asit allows

users tointera t withea h other,butat thesame timethe ompanyhas ontrol

an seethe omments posted.

The implementation of the tool on the ustomer side is really quite simple,

sin eit onsistsinthesimpleaddition ofaspe ialtextareanexttotheprodu t

des ription.

Turnsout to be mu h more omplex thestru ture of the ompany software

for managing omments. Hereinafter will dis uss the storyboard shown in

Fig-ure 2.14 When theappli ation starts, a s reen appears thatsummarizes all the

new ommentstobeanalyzed,groupedbyprodu t. Theusersele tstheprodu t

for whi hwantstomanage ommentsandtheappli ation opensthedetailpage

asso iatedwiththat produ t. Thispage lists four maininformation:

1. the produ tname;

2. the produ tdes ription;

3. the listof approved omments withother information su h asauthorand

dateof submission;

4. thelistof ommentstobeapprovedwithotherinformationsu hasauthor

and dateofsubmission.

For the management of the new omments, the appli ation provides three

buttons:

ˆ Delete. Pressing thisbutton willopen apopupwindow inwhi h you are

prompted for onrmation before deletingthemessage.

ˆ Reply. Thebuttonallowstoreplyprivatelytothe omment. Youwillsee

the appropiatepage, whi h ontains the omment andthereporter name.

Theanswer an besent through the orrespondingbutton.

ˆ Approve. Whenthebuttonispressed,theappli ationloadsapagewhere

you an assign a ategory and a vote to the omment. Pressing another

buttonthe omment is posted to ompanyTwitter a ount.

Table 2.11: A esstable for Twitter integrationtool.

User Type A ess

NotRegistered NO

Registered YES

Figure2.14: Storyboard ofthebridge appli ation.

Table 2.12: A tionsof Twitter integration toolavailableto ustomers.

A tion Des ription

Report Writesthereport

Consult Viewsprevious reports

2.2.6 Forum

Table 2.13: A tionsof Twitter integration toolavailable to employees.

A tion Des ription

Reply Writesthereporting reply

Consult Views previousreportings

Approve Approves omment

Post Post omments onTwitter a ount

Classify Assigns ategoryto omments

Vote Provides avoteon levelof ustomersatisfa tion

forum allows users to sendpubli and private messageswiththe helpof one or

more moderators. In this way a solution an be shared among several people,

respondingwithasingle post toall thosewho areinthesame situation.

Thisapproa h bringstogethermultipleentitiesgoingto reatea ommunity

of people who share interests. In our ase, we an get groups of people united

bythe produ tspur hased. Onlyregistered ustomersarea tiveparti ipantsin

adis ussion, everyone else an seethethreads. Withthislogi of ollaboration,

users an nd solutions independently and thus indire tly help the ompany in

its a tivityofsupport ustomers.

On the other hand this approa h is risky be ause small problems an be

expanded, spreading dis ontent in many ustomers and so the ompany image

an be damaged. A solution to this problem is to lter the messages left by

users: before being published the message must be approved by a moderator.

Thisinvolves a osttothe ompanybe auseanemployee hastousesomeof his

timeinreading theindividualmessages.

Table 2.14: A esstable for Twitter integrationtool.

User Type A ess

NotRegistered NO

Registered YES

Employee YES

Table 2.15: A tionsof forumtoolavailableto ustomers.

A tion Des ription

Report Writes the omment

Consult Viewsprevious dis ussions

Motivate Givesthemotivationfor thenew dis ussion

Table2.16: A tionsofforum toolavailableto employees.

A tion Des ription

Reply Joinsin thedis ussion

Consult Viewsprevious dis ussions

Approve Approvesdis ussion/ omment

Post Adds approved omments to thedis ussion

Modify Edits omment

Classify Assigns ategory to omments

Vote Providesavoteonlevel of ustomer satisfa tion

Figure2.15: Channelsfound.

2.2.7 System design

Before starting the design phase it is ne essary to make a areful analysis to

identify the dierent a tors intera ting inthe systemandtheir role init.

Intuitively, one an say that the entire appli ation works around the

us-tomer. Infa the isthe subje tfrom whi hwe obtain information insomeway.

On an abstra t level, it is possible to imagine the user as the sour e of input

data tobeanalyzed.

the information to be obtained is the level of ustomer satisfa tion, whi h an

be manifested through emotional states that the lient expresses dire tly and

indire tly. For these reasons it was de ided to reate a spe i ategory alled

per eption. Examples of values for the ategory are: satisfa tion, anger,

ap-pre iation,interest, and soon.

Per eptions an beobtaineddire tly,through evaluations thatusersprovide

onaspe i produ t,orindire tly,throughtheanalysisofhowtheyusethetools

available to them. Customerfeedba k an not always be a hieved through the

useoftoolssu hasmandatorysurveys,be ausethispro edureis onsideredtoo

invasive by the user who generally tends to avoid it. It is therefore important

to look for alternative methods to extrapolate judgments about produ ts and

servi es. One of the proposed solutions is to analyze the a tions that the user

performs with the tools provided. Thinking users who browse inthe ompany

website, you an tra e the paths that they arry out and make assumptions

about their emotional state. From this reasoning it is lear that users intera t

with the tools through a tions. Ea h tool has its own set of a tions, however,

mostofthemare ommontoseveralinstruments. Hen ethedenitionofa tion:

doingthesamethingwithdierenttools. Forexample,"vote"inthe hatmeans

toassignanumeri al valuetouseremotionalstate,whilethesamea tiononthe

ompanywebsitemeanspressthe"Ilike"buttonin orresponden eofaprodu t

or fun tionality.

Finally, the last big ategory is represented by the ontext of navigation,

whi hprovidesusmoreinformation about thereasonsthatled ustomersto use

a tool. A user who leaves a omment after navigating repeatedly between the

same pages of a spe i produ t tells us that the user will most probably be

angry.

Figure 2.16shows s hemati allythe ategories des ribedabove.

USER

ACTION

PERCEPTION

NAVIGATION CONTEXT

Figure 2.16: Main ategories of thesystem.

Re alling that the purpose of the system is the al ulation of the level of

satisfa tion, we an saythat per eptions are loselylinked to a tions, sin e the

performeda tion ree tsthemoodof the personwhoexe utes it. In thissense,

thea tionsarethe oreofthe pro essofgatheringinformation,infa t,through

them, you an extra t theassumptionson whi hassert onsiderations.

A tions,however, varyinnumberand typologydependingon thetoolused.

So there is also a relationship between the a tion and the tool thatenabled us

to generateit, whi hallowstoknowthe sour eofa tion,andthereforeprovides

important additionalinformation. Ingeneral,mostofthetoolsallowtheuserto

provide expli itinformation about hislevelofsatisfa tion(forexample,through

the "Ilike" button, popup for voting,et .). There is, however, another type of

informationinadditiontothedire tone,thatis,theintrinsi information. From

the analysis of the ontext in whi h a tool is used, you an perform dierent

onsiderations: there is a dieren e between a ustomer that opens dire tly

a hat and a ustomer that opens it after onsulting several times the web

do umentation. From this observation it is possible to identify the need for

a further relationshipbetween theinstrument and its ontext of use.

As stated above, dierent tools have dierent a tions, but it is also true

that not all users have a ess to the same tools. The subdivision of the users

leads to the denition of the two main ategories of users already mentioned

(registered andunregistered). Theuserplaystheroleofthemaina torbe ause

he uses the tool, whi h allows to perform a tions from whi h we extra t the

desired information. On theother hand the ustomer an usea tooleven after

examiningaprodu tofinterest. Infa t,most ofthefeaturesoeredisbasedon

theintera tionbetweenthe ustomerandtheprodu t,whi hindire tlygenerates

a tions. Withthisspe i ation we an reatetwo parallel pathswhere theuser

an intera twiththe toolintwo ways: dire tlyorthroughtherepresentation of

the produ t.

Thenals hemeisdes ribedinFigure 2.17 , Table2.17 andTable2.19 .

The entity USER represents the set of all possible users of the tools

previ-ously dis ussed. Registered ustomers, potential ustomers and ultimately the

employeesbelongtoit. Thepurposeofthisentityistheidenti ationoftheuser

that intera ts with the system. It is important to know what a tions

individ-ual ustomersperformbe ause,throughappropriate software, you an gettheir

proles. This isthe basisof targetedadvertising. Iftheuser isnot a registered

user, will onlyhave a ess toa limited numberof tools. Among all thesetools,

the mostimportant is the ompany'swebsite. Inthe ategory of non-registered

users, there arealso potential ustomers,then you shouldpayparti ular

USER

TOOL

ACTION

NAVIGATION CONTEXT

_PERCEPTION

USE

BROWSE

ARGUE

EXECUTE

PRODUCT

EXAMINE

PROVIDE

Figure2.17: Finals heme ofthesystem.

The user an intera t with thesystem through the dire tor indire t useof

the instruments. Indire t use means perform a tions without passing through

produ ts representations. For example, the start of the hat an be done by

sele ting a pur hased produ t or by starting a onversation to make general

proposals. The user an also use a produ t, whi h in turn allows the use of

one of the dis ussed tool. Thisrelationship between user and produ tis alled

examine.

The entity PRODUCT allows you to olle t all the information related to

ea hprodu t. Itplays akeyrolebe ausetheassessmentsfortheprodu tsallow

for the identi ation of weaknessesand thenthe hangingof business pro esses

that manage them. Inmost ases,the evaluationis almost always attributable

to aper eptionarising fromuse ofa produ t.

If the representation of the produ t provides a tool then it is possible to

automati ally re onstru t theasso iation between tool and user that has used

it, through the utilization of the relationship provide. Instead the dire t link

between USERand TOOL ismadebytherelationship use.

The entity TOOL is the enter of the diagram be ause it is the missing

information. The information ontent is very wide, sin e the appli ations are

theoreti ally ableto tra k ea huser's li k. Thesize of theinformation ontent

dependsonhowmu htheuseruses thetool, sowehaveto provide addedvalue

to user's lifethrough the utilization oftheinstrument.

The information you intend to get is not only the expli itone, provided by

the tools themselves, but also the most intrinsi , not immediate, but it turns

out to be very important. In general, you an obtain information, analyzing

the routes taken before the use of a tool and from where did it ome. The

tra king of a tivities arried out by the user before starting a spe i tool is

made possible by the NAVIGATION CONTEXT entity. Through it is then

possible to dedu e additional information. The relationship that positions the

toolwithin the navigation ontext is alledpla e.

Among the most interesting relationships appears the relationship exe ute

whi h onne ts TOOL with ACTION. The latterrepresentsthe sets of a tions

listed in the tables in previous se tions. Ea h tool has its own a tions, ea h of

whi hallowsthere onstru tionoftheusernavigationwithinindividualsessions.

Ee ts of a tionsallow to dedu e per eptions. For this reason we have reated

a relationship alledargue.

Attheendofthepro ess,fromthevaluesstoredintheentityPERCEPTION

ispossibletoextra tanestimateofthelevelof ustomersatisfa tion. The

appli- ation, thatdeals theanalysisof the olle teddata, asso iatestheindividuated

per eptions to numeri al valuesand fromthese itispossibleto perform various

Table 2.17: Mainentities ofthesystem.

Entity Des ription

USER Havebeenidentiedthreemain ategoriesofusersthat

an intera t withthesystem:

ˆ Registered ustomer. Customerofwhi hyou

know the data and you want to keep tra k of

hisa tions. Generallybelongtothis ategoryall

those ustomers who have already pur hased a

produ t. This type of ustomer, in addition to

a ess to non-registered ustomerstools, has at

itsdisposalasetofunique tools,whi h areused

asanadditional inputsour e for theassessment

ofthe levelof satisfa tion.

ˆ Non-registered ustomer. Usersofwhi hyou

donot know spe i information about the

ur-rent a ess. Generally, potential ustomers

be-longtothis ategory. Thistype of ustomer has

a esstoalimitedsetoffeatures. Informationis

extra ted through tra king of ustomer a tions.

The s ore thus obtained must have less weight

ompared to thatobtained inanexpli it way.

ˆ Employee. The ompany'semployeehasa ess

tospe iallymodied versionsof ustomerstools

thatallow him toexpress per eptionsdire tly.

ACTION Inthis ategorythereareoperationsthatallowyouto

run thesameintera tion withdierent tools.

PERCEPTION Itrepresentstheemotionalstate oftheuser. It an be

obtained dire tly (by lling in the appropriate elds

byuser orbyanemployee who intera tswithhim) or

estimated(analyzing thetra king of usera tions).

NAV.CONTEXT An important sour e of additional information useful

to strengthen the analysisof the tra king is given by

thenavigation ontext understoodasattitudes,paths

anda tionstakenbytheuserbeforearriving ata

spe- i page or before laun hing atool.

TOOL Setoftoolsusedbytheusertointera t withthe

om-pany.

Table 2.19: Relationships between entities.

Relationship Des ription

argue The a tionsare the ore of thesystemof information

gathering as through them you an build hypotheses

on lient's per eptions about a spe i

produ t/ser-vi e. So isrequired arelationshipbetweentheentities

ACTION andPERCEPTION.

exe ute The a tions, however, are in their turn generated by

tools.

browse The sear hed information isnot only theexpli it

pro-vided by the tools themselves, but also the most

in-trinsi ,not immediate. Ingeneral,you anobtain

im-portant information,analyzing thepaths takenbefore

the use of a tool and from where did it ome to use

thetool.

use A tions are made possible through the use of spe ial

tools. These tools are provided to the user based on

the ategory they belong. This onne tion between

the user and the tool is des ribed in the relationship

use.

examine Mostofthefeaturesareasso iatedwithprodu ts: the

omments are left on produ ts, the hat is opened as

a result of problems/suggestions/questions regarding

produ ts, et . It is therefore important to make

ex-pli it this relationship.

provide The last important relationship is theone thatallows

2.3 A messaging tool for te hni al or ommer ial

sup-port

2.3.1 Obje tives

ƒAny tool isa weapon ifyou hold it right.‚

Following theanalysis arriedoutpreviously,itwaspossibletoidentify aset

of tools apable of perform our goals. This phase was followed by a feasibility

study with the purpose of hoosing only one of them. By analyzing one by

one these instruments we an extra t the drawba ks and weaknesses possessed

and rea h a ompromise between available resour es and expe ted results. The

followinganalysissummarizes brieythereasoningsfollowed forthesele tionof

ar hite ture to develop moresuitable to our needs.

The hat is a very interesting tool and able to provide a large information

ontent. Furthermore, it is not di ult to implement and ts perfe tly to the

ontext. From the pra ti alpoint ofview,however, itisdi ult to usebe ause

it requires the onstant presen e of at least one employee that responds to all

hat messages from ustomers. This involves a ex essive workload, di ult to

sustainbya ompany.

Thewebareaisone ofthemostee tivetools,on theone handbe ausethe

ompany website is an obligatory rossing point for all users and on the other

be auseit isable to olle tboth typesof per eption. The biggestdisadvantage

is the di ulty of integrating new fun tionality on a omplex and onstantly

hangingstru ture.

The me hanism for the dire t input of reports requires small additions on

the ompanywebsiteandthe reationofame hanismforthemanagementof

re-porting. Themaindrawba kisthe omplexityof hangingthe ompanywebsite.

For the same reason alsotheintegration withTwitter was dis arded.

Theforum, aswellashaving di ultyintegratingwiththewebsite,hasthe

drawba k, alreadymentioned,to exposethe ompanyto theriskofun ontrolled

spread ofdis ontent that ould damagethe image of the ompany.

Theinstrument hosenfortheimplementation isthereforetheti kettra ker.

Inadditionto theadvantages alreadymentioned, itenjoysintegration, through

appropriate arrangements, with Mantis. In this sense, the various reports an

bemanaged asti ket, therebyusing analready existingstru ture asa

2.3.2 Components

In 2012 smartphones sold in the world have ex eeded one billion units: one

person out seven has one, onrming the su ess of what has been alled the

fastest-spreading te hnology inhistory. Inone year thenumberofsmartphones

in use rose from 708 million in the third quarter of 2011 to 1,038 billion at

the end of September 2012. And that number is expe ted to double from now

to 2015. Consequently has risen the number of appli ations downloaded from

various Mobile stores. We an say that a great battle is taking pla e between

twodierentwaystoa esstheInternet: the"new"app- entered,andthe"old"

browser- entered. Whi h one will win? Impossible to say, but just think that,

a ording to thelatest estimates, by2013 the number of users who onne t to

the Internet viasmartphones willex eed those usinga omputer.

Thisinformationindi atesthatthemarketformobileappli ations,and

there-fore the entire mobile world, is an attra tive market to all and the presen ein

the eldof Mobile Appsis ru ial for a ompany. Butwhat arethereasons for

whi h the presen e in themobile store, su h asApple's App Store and Google

Android Market, it isadvantageous for a ompany? Listedbelow arethe main

ones.

ˆ Visibility. The number of sear hes within the Apps Store is very high

and therefore, appear among the sear h results, is good publi ity for the

ompany.

ˆ In rease the value of the brand. Have its ownappli ation isasign of

a dynami ompany, always attentive to the evolution of te hnology and

abreast withthe times.

ˆ Media overage. 75% of the planet is not yet onne ted to the

Inter-net, but with the mobile world we an rea h the users withpoor internet

onne tivity,therefore at hingevery opportunityfor ompany business.

ˆ Bring the brand dire tly in the ustomer's smartphone.

Smart-phones are instruments that are always available to users, and therefore

potentiallyalwaysa essible.

Thebenetsderivingfromthe reationofmobileappli ationsarevaried,but

an appli ation must be used to deliver added value to ustomers, not only for

advertising purposes otherwise you risk to redu e the value instead of in rease

it.

Insummary, we an saythat therules to a hieve su ess intheApps Store

ˆ Ease of use. Appli ationsshould be intuitive and not too omplex to be

used;

ˆ Completeness. The servi es must be fully fun tional and in ontinuous

improvement;

As a result of these onsiderations, it was de ided to implement the ti ket

tra ker as an appli ation for mobile devi es, using the PhoneGap framework.

Theti kettra kerrealizesdire t ommuni ationbetween ustomerand ompany,

making itan a tive playerinthe business life: he an praise thework, propose

solutions and riti ize the hoi es that are not shared. All in a few simple

steps. Infa t,the referen emodelisthespeedofinformation that hara terizes

Twitter, whi hallows to reate asort ofasyn hronous hat.

The designed systemismade upof threemain omponents:

ˆ ustomersappli ation;

ˆ employeesappli ation;

ˆ appli ation for the analysisof olle ted data.

Ea h of them hasits own hara teristi sthat an perform severaltasks

un-derits jurisdi tion, but the ommon feature that appli ations musthave is the

abilitytointerfa e withMantis. Withthisrequirement, theuserisabletomake

a report whi h isstored asa ti ketin the Mantis database. The employee

reg-ularly monitors from its appli ation thelist of new reports and manages them

individually. Intheend,alltheinformationstoredinthedatabasearepro essed

by the employee throughthethird appli ation.

Analyzingattheabstra tlevelthefun tioningofthesystem,itisinteresting

to notethatthe omponentsdonot ommuni ate dire tlywithea h other. The

individualappli ations ommuni atewithea hotheronlythroughtheuseofthe

Mantis database, whi hthus assumes a entral role inthear hite ture.

Here below is proposed a rst overview of the individual omponents that

servesto understand indetailthe te hnologi al needsofea h of them.

Customer App

Theappli ationusedbythe ustomerisstru turallyverysimple. Butwhatuser

needs for a omplishhis tasks? Surely he mustbeable to view hisreportsand

topla enewones. Toen ourage theuseoftheappli ation,prioritymustbethe

speed of use. On e established these hara teristi s it is possible to propose a

assignedhimbythe ompany. Inadditiontotraditionalusernameandpassword,

it is possible to also enter the address of the server that provides the servi es

ne essary for the appli ation. If the login is su essful, the appli ation will

display the main s reen that shows two lists: one ontaining the reports in

resolution andanother ontainingthe losed reports.

Figure2.18: Loginpage in ustomer ti kettra ker app.

The ommuni ationsbetween ustomerandemployeeinvolvetheex hangeof

multiple messagesand onlywhen both parties onsiderended the onversation,

theemployeemayendreporting. From reationto losingthereportisinthelist

of reports in resolution. Sin e the losure, thereporting is in thelist of losed

reports. Therefore the purpose of this subdivision is to provide a on eptual

ordering of data that allows user to develop a global view of the urrent state.

Inordertoobtainalsoavisualseparationofthelistswere hosendierent olors

for ea h list.

Everyelementofea hlist ontainsthebasi informationthatsummarizethe

report. Thisinformation is:

Figure2.19: Homepage in ustomer ti kettra kerapp.

ˆ Type of reporting. The purpose of this value is to provide a rst logi al

divisionof thedata;

ˆ .ReportStatus. It isused to ontrol howthesignalevolves.

The messagesaresorted a ordingtothe dateofthelast updateinorderto

keep under ontrol the reports that re eive answers. It is possible also, in the

implementationphase,addanadditionalviewwithreportdetails. Inthiss reen,

it's possible add all the informations thatdes ribe indetail the ti ket. Among

all these, the appli ation must display a spe ial se tion that shows the entire

onversation between reporter and handler. Inorder to exploit thepotential of

mobile devi es, ithas been added theability to s roll through the listof ti ket

details usingtheswipe a tion.

The last s reen is for the insertion of a new alert. Within it the user must

rstsele tthetypeofsignalingbetween threepossiblevalues: praise,suggestion

and omplaint. The se ond valueto spe ifyis the produ tfor whi h you want

make the alert. Thisvalueand the previous one are hosen by theappropriate

drop-down menu, inorder to avoidinput errors. The third and fourtheld are

respe tively title and des ription. With these four values the message an be

Figure2.20: Page for reportdetailsin ustomer ti kettra ker app.

Tofa ilitatethesear hforreportsinthelists,theappli ationprovidesalter

to sear hfor keywords withintheinformation displayed. The lter sear hesthe

inserted keyword into its bar anddisplays real-time listitemsthat ontain it.

Employee App

The employee app is mu hmore omplex be ause itmust be able to handle all

the reports made by ustomers. Again the question arises: what theemployee

needs to a omplishits tasks? The employee mustbeable to:

ˆ view newmessages(in additionto thosealready losed);

ˆ onsultthe listofreports,divided by ategory;

ˆ display thedetailpage ofthereport;

ˆ respond tothe report;

Figure2.21: Storyboard for ustomerti ket tra kerapp.

Theappli ationwillthenprovideasetofpagesthat anperformallthese

a -tions. Atthispoint,ithasthereforebeenproposedthestoryboardinFigure2.23 ,

whi h showsthemain s reens.

After the usual login page, the appli ation will automati ally download all

ti kets stored in Mantis. Then the messages aregrouped a ording to type, in

three separate lists. A further subdivision of the ti kets based on thevalue of

the state: the main page shows the number of new ti ket with the status for

ea h ategory. From this page you an a ess pages of reports and resolved in

resolution by pressingthe orresponding buttons. Thesepageshave astru ture

similarto that ofthemain page.

Sele ting a type from the list you will see the s reen shown in Figure XX

ontaining the list of messages asso iated with that parti ular type. Even in

this ase, for ea h ti ket areonly reported basi information (title,des ription,

statusand type). The listelements are sorted a ordingto the dateof thelast

update, pla ing the most re ent items at the top of thelist. The employee an

s roll throughthelistand de ide whi h signaling resolve.

After li kingthedesirediteminthelist,thepage ontainingti ketdetailsis

opened. Amongthe displayedinformation appearthename oftheprodu t, the

reporter,submissiondate, thenameofthe instrumentbywhi h itwasmade

Figure2.22: Homepage for employee ti kettra kerapp.

to newest, soyou anautomati ally build the onversations thattookpla e. In

the ase of reports that have not yet been answered, in this area is displayed

a string that noties the la k of answers. If the user wants to answer, he an

insert his reply in the text box at the bottom of the page and then press the

"Reply". Ifyou want to dire tly lose the alert,he presses the"Close" button.

In the latter ase, before updating the ti ket in Mantis, the appli ation opens

a popup window whi h asks the employee to assess the state of the ustomer.

After performing one of the two operations, the user is redire ted to the main

page that isre- reated dynami allywiththe hanges.

The pro ess for the management of onversations in resolution is similar

to that for new ones. Slightly dierent is the pro ess of onsultation of alerts

resolved,asthey an not be hanged,but only onsulted.

Analysis tool

After olle tingthedatathroughtheappli ationusedbythe ustomer,andafter

providings oresto thelevelof ustomersatisfa tionforea hreports,itis

ne es-sarytopro ess thedatainorderto extra tinformation ofinterestfor ompany.

Figure2.23: Storyboard foremployee ti kettra kerapp.

able to extra t the typesof produ ts preferred by ea h user, through the

anal-ysis ofthe reports. On the other hand,this tooldoesnot have the ambition to

perform omplex analysis, but it is meant to show how it is possible to gather

andpro essthedataandthenamongallpossibleinformation ithasbeen hosen

ˆ thetotal s oresasso iated with ea h produ t;

ˆ thebestand theworst produ t;

ˆ reports grouped byprodu tand byusers;

ˆ reports madebyindividualusers.

After xing these obje tives it is possible to onstru t a storyboard of the

appli ation. Unlike the other two appli ations, this hasnot been implemented

for mobile devi es, asit is not ne essaryto have it always available. It is used

regularly by employees to monitor the performan e of the produ ts and other

parameters. Sothe appli ation hasbeen designed asa web interfa e, onsisting

of asingle page inwhi hare gatheredall thedataof interest. The basi ideais

to makeavailabletwotabs: onerelatedto theprodu tsandtheUsers. Ea htab

onsists ofa summarytable whi h liststhemain data. Inboth ases,the table

ontains the number of reports made divided by type. It also needs a spe ial

area where to show all omments divided on the basis of the main entities of

the belonging tab, for example in the produ ts tab will be reported all of the

omments assigned to ea h produ t divided into three sets, one for ea h type

signaling (praise,suggestion and omplaint). Theproposedstoryboardisshown

inFigure 2.24 .

The instrument able to immediately understand the timetrend of any

phe-nomenon is the plot. Therefore it was de ided to provide the appli ation the

ability to dynami ally reate harts. In this way,bysele ting any produ t, the

employeeseesthetimetrendofthevotesallo atedtoit,allowinghimtohighlight

any riti alissuesanddeterminewhi h ompanypro essesneedadjustment. An

example of hart for the representation of the temporal evolution of a produ t

isshown inFigure2.25.

Part II - Te hnologies

Afterdeningtheobje tives,theanalysismovestothedevelopmentenvironment

ne essary to a hieve them. The development environment onsists of a set of

suitabletoolsfor ourneeds,whi hwere hosentomaketheappli ationthemost

userfriendly possible.

Before writing ode, itis ne essaryto olle t some librariesand some tools.

Theseare:

ˆ jQuery : ThejQuerylibrarymakesdealingwithHTMLandCSSabreeze,

and it is also extremely handy for manipulating XML, and therefore,

XMPP stanzas.

ˆ jQuery UI:The jQueryUIis a urated setof userinterfa e intera tions,

ee ts,widgets,andthemesbuiltontopofthejQueryJavaS riptLibrary.

ˆ jQuery Mobile: A unied, HTML5-based user interfa e system for all

popularmobiledevi eplatforms,builtonthero k-solidjQueryandjQuery

UIfoundation. Itslightweight odeisbuiltwithprogressiveenhan ement,

and hasa exible,easily themeable design.

ˆ Mantis: MantisBTisa freepopularweb-based bugtra king system. It is

written inthePHP s ripting languageand workswithMySQL, MSSQL,

and PostgreSQL databases and a webserver. It is used for storage and

management of reports.

ˆ PhoneGap: PhoneGap is a free and open sour e framework that allows

you to reate mobile appsusing standardized web APIs for theplatforms

you areabout.

ˆ XMPP: Extensible Messaging and Presen e Proto ol is a

ommuni a-tionsproto olformessage-orientedmiddlewarebasedonXML(Extensible

open-information, and onta t listmaintenan e. Designed to be extensible,the

proto ol has also been used for publish-subs ribe systems, signalling for

VoIP,video, letransfer, gaming,Internet of Things appli ationssu h as

thesmartgrid, and So ial networkingservi es.

ˆ Sma k: Sma k is an Open Sour e XMPP lient library for instant

mes-saging and presen e. A pure Java library, it an be embedded into your

appli ations to reate anythingfrom afull XMPP lient to simple XMPP

integrations su h as sending noti ation messages and presen e-enabling

devi es.

ˆ Servlet andJSP:AservletisaJavaprogramminglanguage lassusedto

extend the apabilitiesofaserver. JavaServerPages(JSP)isate hnology

that helps software developers reate dynami ally generated web pages

based onHTML, XML,or other do ument types. The dieren e between

servlets andJSP isthatservletstypi ally embed HTMLinsideJava ode,

while JSPsembed Java odeinHTML.

This hapteranalyzes indetailthese instruments.

3.1 jQuery and jQuery UI

The jQueryandjQueryUIlibraries areavailable fromhttp://jquery. omand

http://ui.jquery. om respe tively. jQuery, like many JavaS ript libraries, is

availableinnormalandminiedversion, thatisa ompressed versionofthele

obtained through the operation of mini ation. Mini ation is the pra ti e of

removing unne essary hara tersfrom ode toredu eits sizetherebyimproving

loadtimes. When odeisminiedall ommentsareremoved,aswellasunneeded

white spa e hara ters (spa e,newline,and tab). Inthe aseofJavaS ript, this

improves response time performan e be ause the size of the downloaded le is

redu ed.

Google has made available many libraries through its AJAX Library API,

and among them also appear jQuery and jQuery UI. So, if the appli ation is

designed torun ondevi es onstantly onne ted to internet you donot need to

download andstore thelibraries, but just onne t to one of theGoogle'ssuper

fast servers.

Alternatively, it is possible to download the libraries and in lude them in

your HTML page. The minied versions an signi antly redu e the spa e

re-quired for storing, in this way the in lusion does not ae t the appli ation's

performan es. To in lude jQuery, jQuery UI and the UI theme CSS in HTML