Data driven management in Industry 4.0:
a method to measure data productivity
Giovanni Miragliotta*, Andrea Sianesi*, Elisa Convertini*, Rossella Distante*
*Department of Management Engineering, Politecnico di Milano, Via Lambruschini 4/B; 20156 Milano (Tel: +39 02 2399 278;e-mail: giovanni.miragliotta@polimi.it) *Department of Management Engineering, Politecnico di Milano, Via Lambruschini 4/B; 20156 Milano (e-mail: andrea.sianesi@polimi.it ) *Department of Management Engineering, Politecnico di Milano, Via Lambruschini 4/B; 20156 Milano (e-mail: elisa.convertini@polimi.it ) *Department of Management Engineering, Politecnico di Milano, Via Lambruschini 4/B; 20156 Milano(e-mail: rossella.distante@mail.polimi.it )
Abstract: In the early 1900s, together with the birth of mass production, modern managerial approaches were conceived, under the motto “you can’t manage what you don’t measure”. Since then, operations managers throughout the world had been getting used to measure the productivity of materials, machines and workers to control and improve their own businesses. Nowadays, in the Industry 4.0 era, the emphasis is shifting toward data, under the new motto “data is the new oil”. Despite many managers pledging allegiance to the principles of data driven decision making, still no comprehensive approach exists to measure how good a company is at exploiting the potential of its own information assets; in other words, no “data productivity” measure exists. In this paper, we present a first method to define and measure data productivity. Relying on a comprehensive literature review, and inspired by the traditional OEE framework, this new method brings some innovative perspectives. First, data productivity is broken into data availability, quality and performance of the decision-making process using those data. Second, it includes both technical and organizational factors, helping companies to evaluate their current level of productivity, and actions to improve it. The model has been tested through three cases studies and it results as effectively implementable. The results obtained from its application reflect the expectations of companies‘managers accelerating the cultural shift needed to fully express the potential of Industry 4.0. Keywords: Data productivity, Performance measurement, data driven decision making, Industry 4.0, Information Management.
1. INTRODUCTION
The industrialization phenomena that characterized different historical centuries, has been strongly influenced by technological progresses called Industrial Revolutions. These revolutions have always drive business to drastic increase in productivity. The first was triggered by the introduction of the steam power, in the middle of the eighteenth century; the second started conventionally in 1870 with the establishment of electricity, chemical products and crude oil, together with the changing concept of mass production; the third, one century later, referred to the effects provided by the increasing usage of electronics and IT in manufacturing industry; finally, in 2011, the term Industrie 4.0 was introduced at the Hannover Messe in Germany.
Industry 4.0 is “a vision of the future of Industry and Manufacturing in which Information Technologies are going to boost competitiveness and efficiency by interconnecting every resource (data, people and machinery) in the Value Chain” (Politecnico di Milano 2017).
The interconnection between the resources is the cardinal element of this revolution and the exchange of data is becoming the new flow to manage with the same relevance of
the management of the materials and products one. Indeed, digitalization, intelligence and connection are the three pillars of the Industry 4.0 that poses the data as the principal actor of this new paradigm.
According to the research reported by the Industry 4.0 Observatory of Politecnico di Milano in 2017, companies currently have acquired knowledge and mastery of the Industry 4.0 paradigm. However, this knowledge is still linked to the theoretical concepts on the importance of data and there are currently no tools that measure their productivity).
In this scenario, companies’ attention is more and more moving towards data and towards the information asset that they own. Indeed, companies are aware of the importance of the data and their potentiality, but currently they have not metrics to measure the efficiency and the value that their information asset is giving or can give to them. In addition, before the current data-revolution, the value of their information asset has been not proper valuated, because of the difficulties of set a proper measure and the limited attention of top managers to this asset. This research has the aim to provide a structure measure of data productivity, answering to the questions: What’s mean that a data is
productive? How much a data can produce value for a company?
The structure of the paper is the following: in the first part, all the topics on which the research is based are presented; afterwards the definition of data productivity is introduced and its related assumptions; then the procedure to calculate the data productivity measure is presented; finally, there are the main insights of the model application.
2. LITTERATURE REVIEW 2.1 Industry 4.0 and Data
Data is one of the main pillars of Industry 4.0, the 4th generation of manufacturing that uses concepts such as cyber-physical systems, virtual copies of real equipment and processes, and decentralized decision making to create a smart factory or “Factory 4.0”.
Within the Industry 4.0 world, big data impact can be summarized in six Cs: connection (sensors and networks); cloud (computing and on-demand); cyber- (model and memory); content/Context (meaning and correlation); community (sharing and collaboration); customization (personalization and value). In this new paradigm, a key role is played by the Industrial Internet of Things (IoT). IoT allows companies to capture data about process and products more quickly, to have global visibility on the overall supply chain, to work with more efficient and intelligence operations that, thanks to autonomous data collection and analysis, allows on-the-fly decision making (IDC Digital Universe 2014). The effective extraction and use of information embedded in the data have become the next frontier to drive innovation, competitiveness, and growth in manufacturing, as highlighted by McKinsey in a series of studies see McKinsey, 2011, 2015.
2.2 Data versus Information
“Data can be defined as a symbol, sign, or raw fact”, see Mingers, 2006. Usually in literature the concept of data is associated or put in contrast with the information one. So far, two different definition of information are present:
1. information is “data that has been processed in some way to make it useful [..] information can be objectively define relatively to a particular task or decision”, see Mingers, 1996. This definition of information implies that the concept of data is objective.
2. “Information equals data plus meaning”, see Checkland et al., 2000. This definition implies that information is subjective, from a set of data different information can be created. Information is subjective because depends on the receiver and on the context in which the message is conveyed see Jumarie, 1990. Information cannot be viewed as an independent entity because has attributes and reflects the intention of the sender and the receiver, see Oppenheim et al., 2003. Different observers may generate different information from the same data given their differing values, beliefs, and expectations, see Lewis, 1993.
2.3 Data and Information Asset
During the late 1990s theories of knowledge asset and their contribution to organisational wealth become popular together with the definition of information as an asset. The role of information as an asset was introduced in 1994 by the Hawley committee, which define information asset as: “data that is or should be documented and that has value or potential value”. This concept, in addition, treated information like traditional assets such as plant and machinery, see S.H Black et al., 1982. Information asset is defined also as “the ability to provide data and information to users with the appropriate levels of accuracy, timeliness, reliability, security, confidentiality, connectivity, access and the ability to tailor these in response to changing business needs and directions”, see Mithas et al., 2011. “Value should be assigned not only to data but also to the system allowing for its exploitation”, see Ahituv, 1989. Because Information Asset is an intangible asset it is difficult to measure, see Evans et al., 2015.
Data can be in the form of structured, semi-structured, or unstructured data. There are several data types such as Master data, that provides the most business relevant information about a product, a supplier, a customer, etc. or Transactional data, that describes the event that happens in a moment referring to one or more master data element. These two types of data together with Tacit Knowledge can be consider the Information Asset of a company. Tacit Knowledge is hidden in human brains and refers to communication among people who share their knowledge and observations, in both formal and informal ways.
2.4 Measuring productivity in a manufacturing company
Productivity is commonly defined as a ratio of a volume measure of output to a volume measure of input use. While there is no disagreement on this general notion, productivity literature and its various applications proposes different measure of productivity, see OECD Manual, 2001.
Looking at the Efficiency aspect of productivity, a production process is operating in full efficiency, in an engineering sense, if it achieved the maximum amount of output that is physically achievable with current technology, and given a fixed amount of inputs, see Diewert E. W. et al., 1999. Manufacturing companies are often interested in measure productivity in terms of efficiency. The Overall Equipment Effectiveness (OEE) measure gives a comprehensive response to this need, underling different losses of the manufacturing process under analysis. Measuring OEE is, indeed, a manufacturing best practice that provides insights on where it is necessary to act to improve the productivity. The most significant way to calculate the OEE index is based on three factors: Availability, Performance, and Quality.
2.5 Data value and Productivity
In literature there are not researches that provides to managers and academics a metrics or a model that allow them to measure the Data productivity. The main insights coming from the literature on the theme are:
1) some companies are aware of the potentially of the data they own, but are not able to exploit them see Ladley, 2010;
2) pro 3) imp acc inf acc attr Tay Ear tog ass con non of 200 pro Str fin to m of t Th tog inte to-pro Th we com hav firs app to fun mo cor fro com 4.1 To pre res data-driven oductivity than companies proving the in countant voic formation asse counting, so ributed to info ylor, 1986. rly studies o gether by Ahit sess this value ncluding that ne of them giv a company, 04) defined oductivity: the rassmann, 20 ancial indicat measure the p their aggregat e model wa gether with a erviews with face interview ovide punctua is provides t ere very usefu
mpanies selec ve already imp st phase of m proached. In p measure eac ndamental for odel, in terms rrectly measur om the succes mpanies objec 1 Data produc correctly ad eliminary step search analysis companies n other compa have, very nformation as ce. Taylor re et, a concept t that will alw ormation for f on the value tuv in 1989. H e (quantitative each method ve a unique in see Aihtuv, a quantitative e Information 004. Howeve
tors and accou productivity o ted nature. 3. METH s built up b an empirical companies’ m ws were usef al responses, o the researc ful to design cted belonging plemented pro model design, particular, an ch characteri r its correct ap s of its meas re the Inform ssful applicat ct of the first e 4. DATA PR ctivity and clos
ddress the re p is the def s (fig.1).
better perfo anies, see McA
often, othe sset, simply b efers to the p
that does not ways result in financial repo of informatio He considered , normative, r d presents di ndicator of th 1989. Strassm e measure o n Productivity r, this meas unting voices f a single data HODOLOGY by joining pe l research p managers, as c ful not to forc expressing t ch original po a comprehe g to manufact ojects on Indu the issue of analysis of d istics of the pplication. Th surability and mation Asset p tion of the m empirical phas RODUCTIVIT sed themes esearch quest finition of th orm in term Afee et al., 20 er priorities because it is n potential valu exist in tradit n no value b rting purpose on are consid d 4 different w ealistic, perce ifferent limits e data produc mann (Strassm f the Inform y Index (IPI) sure is based that do not p a element, bec ersonal resea hases from case studies. F ce interviewe their ideas fr oints of view nsive model. turing industry ustry 4.0. Afte measurability different possi model has he validation o d its capabili roductivity, c model to the se. TY tions, an esse he borders of ms of 12; than not an ue of tional being s, see dered way to eived) s and ctivity mann mation ), see d on permit cause arches three Face-ees to freely. w that . The y and er the y was ibility been of the ity to comes three ential f this Fig. Mas appli mana infra subs comp Data syste data high The infor and cons proc Dret know indiv proc judg Know the orga mana respo Data comp data shou regar incid bugs Data parti and main inno prop 4.2 D The unde Infor gap struc the t of prod 4.3 D The and for th elem Look aside 1. Data produ ter data ma ications, Info agement tools astructures th equent share plete master d a, properly m em allow com
asset for all h-quality inform distinction rmation, and d academics. A siders: data a essed data; kn tske, 1981, M wledge is defin viduals: it is edures, conce ments. wledge manag collective kn anization com agement is p onsiveness, se a protection panies since present in co uld be proper rds whether a dents of diffe s, hacking, fra a maintenanc icular, softwar more perform ntenance. Ind vations, syste perly, see Spru Data producti lack of a m erstand their rmation asset present in the ctured researc theme under a the “Data ductivity, decis Data producti aim of the m Information A he analysis. T ment of the mo king at the obj e from the re uctivity’s bord anagement (M ormation Man s to implemen hat support t ed use accur data”, see Losh
anaged and m mpanies to re
their applicat mation, see Lo
among the data has often As previously as raw numbe nowledge as a Machlup et al. ned as inform personalized epts, interpreta gement refers nowledge in mpete, see Vo purported to ee Hackbarth, theme is p the origin of ompanies are rly protected and how the d erent nature: ud or theft. e theme is re re solutions an mant if they deed, to m ems should be uit et al., 2015 vity definition easurement to current lev and of data, e literature. To h was essenti analysis and to Productivity sion-making p
vity and mode
model is to me Asset related The data pillar
del.
ject of the ana search work c ders of analysi MDM) incorp nagement me nt the policies the capture, rate, timely, hin, 2009. monitored, in ely on a unifi tions, thanks t oshin, 2009. e concepts been address presented, a ers and facts authenticated , 1983, Vanc mation possess information ations, ideas, o s to identifyin an organizat on Krogh, 1 increase inn 1998, Alavi e present in li f the data itse often sensiti d. Technical data is secured components elated to phy nd database c are subjecte make use of e up to date an . n tool by which vel in the heart of the I o correctly ad ial to define t o define the th model” d process. el’s assumptio
easure the pro to a specific r it is presente alysis, this res conducted sin is representati porates busin ethods, and d , procedures, integration, consistent, n the master d fied and coher to consistent of knowled ed by research commonly v s; information information, ce, 1997. Inde ed in the mind related to fa observations, ng and leverag tion to help 998. Knowle novativeness et al., 2001. iterature and elf. Indeed, si ive, their con security of d d against poss failure, softw ysical storage can become fa ed to the pro f the techn nd be maintai h companies management Industry 4.0, ddress this gap the boundarie
hree main pil eveloped: d ons oductivity of d process selec ed as the prim search does a s nce now, look ion ness data and and and data rent and dge, hers view n as see eed, d of acts, and ging the edge and in ince ntent data sible ware . In aster oper nical ined can of is a p, a s of lars data, data cted mary step king
bac into Da pos for pec but In ma on poo con of not act cor Th dat tim dec Pro mo dif Th Pro effi of be tha Th wa des tha Th ma is p on con als cap Th bet cha pat OE dev cha pro Th log ava Inf Lik Da in t 𝑫𝒂 𝑫𝒂 ck at the data o knowledge. ata are now pr ssible to comb r different pur
culiar informa t it is the data
the model pro aking process. a specific pro or performanc ntext. In other the Informati t to understan tions are need rrectly improv erefore, after ta of interest me also the to cision are sele oductivity of odel to more t fferent areas o e starting p oductivity m ficiency metric productivity, applied to the at considers erefore, the m as: how the
signed as a p at occur during e OEE metric achinery/proce possible to an the contrary t nsidered. Loo o the smaller pture in an agg e structure fo tween OEE aracteristics c th of normativ EE industry s vised assum aracteristics ocess) follows at means, the gic of the OE ailability, qu formation Ass ke the OEE, t ata Availability the following 𝒂𝒕𝒂 𝑷𝒓𝒐𝒅𝒖𝒄𝒕 𝒂𝒕𝒂 𝑸𝒖𝒂𝒍𝒊𝒕𝒚
per se, not tr resent in facto bine, interpret rposes. For thi ation, a pecu itself. oposed the fo . It is possible ocess allows th ce of the Inf r words, a me ion Asset prod
d easily the re ded to be tak ve.
the selection for the proces ools that help ected. For this the company than one decis f applications point for th odel, is the cs in manufact express by th e productive f data the pr main question Information productive res g the decision cs, is often co ess/produced nalyse the cau
this is almost oking to a spe r losses and t gregate analys llowed by thi factors a coming from ve model. Ind standard) as ming that a (the one re s a scheme si e model was EE metrics ba uality and p set characterist
the Data Pro ty, Data quali formula (1): 𝒕𝒊𝒗𝒊𝒕𝒚 𝑰𝒏𝒅𝒆𝒙 ∗ 𝑫𝑺𝑺 𝑷𝒆𝒓𝒇𝒐 ransformed in ories in a huge t them in seve is, the entity o liar transform ocus it is on a e, indeed to n he identificati formation Ass trics on the ov ductivity wou esult obtained ken and the re
of a process ss are identifie the decision , to assess the y it is neces sion-making p s. e developme e parallelism turing. This ef he concept of field but also rimary feed that guided t Asset of a source highlig -making proc omputed in rel product. In th uses of low ef impossible if ecific element their causes th sis. is research is and the In the literature deed, taking th starting poin also the In lated to the imilar to the designed by ased on the th performance tics coming fr ductivity Inde ity, DSS perfo 𝒙 𝑫𝒂𝒕𝒂 𝑨𝒗 𝒐𝒓𝒎𝒂𝒏𝒄𝒆 nto information e quantity and eral different of analysis is mation of the a specific deci
otice that focu on of the caus set in that spe
verall perform uld drive man d and which ty elated prioriti to analyse, a ed and at the maker to tak e Information A ssary to apply process, looki ent of the m with the fficiency defin f output/input to the digital of all proce the model bui
company ca ghting the “lo
ess.
lation to a spe his way, inde fficiency in de
the overall pl allows to ide hat are difficu
a pure compa nformation A e, by adopting he OEE factor nt, the model nformation A e decision-ma OEE metrics applying the hree main fac and a serie rom the literat ex is compose ormance, as sh 𝒂𝒊𝒍𝒂𝒃𝒊𝒍𝒊𝒕𝒚 ∗ (1) n nor d it is ways not a data, ision-cusing ses of ecific mance nagers ype of ies to all the same ke the Asset y the ing at Data OEE nition t, can l one, esses. ilding an be osses” ecific eed, it etails, lant is entify ult to arison Asset g the rs (by l was Asset aking s one. same ctors: es of ture. ed by hown ∗ ) The data by t inter decis actua on t exter this stora poor impl Fig. The “com ones colle respo the relia slow data inclu attrib of co Fig. The how the d data by a Data Availab at the disposa the “interestin resting data a sion-making p ally available the ones actu rnal or interna first loss of d age infrastruct r management lementation of 2. Data produ Data Quality mpleted and c s that are c ection are co ond as rapidly process, allo able data. The w data collecti , the complete ude wrong ele butes. Each da ompleteness a
3. Data produ DSS Perform the data are m decision-make refers to the algorithms. In bility factor h al of the decis ng data” and are the ones process, not c . Of these dat ual available
al company’s data is the ab ture (e.g. sen t, that fails in f these solutio uctivity’s Avai y factor selec orrect”. Form haracterized onsidered. In y as required owing the de e data that are
ion are report e and correct ements and are
ata can be eva nd correctnes
uctivity’s Qual mance factor
managed by th er in the proc
set of the gen n this case t
has the aim to sion maker an d the “gener that are of considering if ata, a portion to the decis sources of da bsence of app nsors). This is n a proper idea ons. ilability factor cts the data t m the generate by a correc other world by the user o ecision make e not timely b ted as losses. ones refer to e composed b aluated in term s.
lity factor rep refers to the he decision su cess. In partic nerated data t the losses ar o identify all nd it is compo rated data”. interest for f they are or is selected ba sion maker fr ata. The cause propriated IT
s often related ation, design
r representatio hat are “time ed data, only ct frequency s, the ones or necessitated r to work w because of a
From the tim those that do by all the relev ms of percent
presentation measurement upport system
cular, “valoriz that are mana re related to the osed The the not ased from e of and d to and on ely”, the of that d by with too mely not vant tage t of and zed” aged the
we afte alg per dat tha pro “in col pla ma inte dec “ag dif ind ava dat end on rati the Th a n sof Fig In of dec dat obt the dif to ass ass pro Sup Ch dis inte inv eakness of the er the ident gorithms, an ev rcentage of lo ta. The “On ti at arrive in a ocess because ntegrated” dat
llects data co atform in orde aker. The los
egrated with cision-making ggregated” da fficulties in dicators espec ailability to m ta to look at. d, by the decis the bounded ionality of in eir cognitive l e data losses no structured ftware or tools g. 4. Data prod addition, the the company cision-making ta and Inform taining by the e company lo fferent models assess the ri sessment cons signment of a obability of oc pply Chain. T hopra and So sruption, delay ellectual prop ventory, see C e algorithms tification of valuation of th ss because of ime” data cha a due time u e of perform a are those su oming from er to ease the
ses are ident each other g process th
ata. The losse using data cially because make the deci Then the “us sion maker. T rationality co ndividuals in imitations com for this reaso decision-mak s for taking th ductivity’s Per model is con , because the g process is co mation asset va e model applic ooking at the s in literature isk of a com sidered for th a grade base ccurrence of e The risk driv odhi and are y, IT systems priety (IP), leg Chopra et al., 2 applied to th the data th heir performa f a not proper aracteristic is useful for the ming and fast upported by an different sou eir consultanc tified by the . The data hrough indic es in this case without look e of the usua ision or the p sed” data are These further d oncept that in taking decis mbined by the n are more nu king process o he decision. rformance fac ntextualized in e entity of the oncordance wi alue. This imp cation as to be e process an that provide i mpany. The S his dissertatio ed on the lev each risk drive vers proposed e divided in , forecasting gal and econom 2004.
he data. There hat are objec ance determine exploitation o related to the e decision-ma t algorithms. n infrastructur urces in a s cy for the dec data that are presented in cators are c e are related t king at sum ally restricted presence of se those used, a distinction is b ndicates the lim
sions, highlig e time availab umerous in ca or a not suppo ctor representa n the Supply C e turbulence i
ith the relevan plies that the r e contextualize nalysed. There indications on Supply Chain on is based on vel of impact er impacting i are adapted n 8 macro a & communica mics, procurem efore, ct of es the of the e data aking The re that single cision e not n the called to the mmary time everal at the based mited ghting bility. ase of orting ation Chain in the nce of result ed by e are n how n risk n the t and in the from areas: ation, ment, . A m decis meas impl good The cont decis resul and mod struc “Thi the intel revie mod meas asses facto of t intel impo desig and choic char Beca estab the m from of m in c obtai path orga comp Asse to as comp betw to im adde the p impr on t mod poss mod from that Asse emer perfo impr matrix that c sion-making p sure of the ris lies the possib d result by the results obtain extualized loo sion-making p lt and its spec
DSS Perform del developed cture of the inking first” p decision-mak lligence, desi ew. Only the del application
sure the Dat ss their level o ors of the Dat these three p lligence phase ortance of the gn one consid
the “On time ce one, instea acteristics.
ause of the no blish the good model applica m 40% to 60% managers. In a class compan ined reflect: t that compani anizational, t panies must e et, and decisio ssess the AS-I pany related t ween the same mprove the in ed. Because o
presented app rove the mode
he involveme del application ible to refin del with other m more applic the model po et productivit rged that h ormance on t rovements. combine the processes of a sk level for ea ility to evalua Data Product ned by the m oking at the an process, the ov ific factors (D mance). To cle d has been c decision-mak presented by king process gn, choice, first three ar n, because th ta and Inform of goodness fo ta Productivity phases, but e is character Data Availab ers as the mos ” and “integr ad, needs “agg
5. CONCL velty of the m dness of the q
ation to the t %, but they are
ddition, the re y is about the novelty of ies must follow technological embrace to be on-making pro IS situation of to a specific p metrics after nformation ass f the novelty lications are t el working on ent of the act n also to othe ne the charac r experts’ opi cations. Overa sits a first app y, from its a elps compan these elemen 6. REFER risk drivers a company al ach process. T ate the relevan tivity index. model applica nalysed comp verall Data P Data Availabil early present t considered to king process Simon. This composed implementati re the ones co he scope of mation Asset for taking a de ty Index have with differe rized by a m bility and the D
st relevant the rated” data ch gregated” and LUSIONS model proposed quantitative re three case stu e in line with esult obtained 85%, therefo f the measure w to improve and cultura etter manage d ocess. The aim f the Informa process, for th
some projects set managem
of the model the first ones n the modality ctors, on the er processes. cteristics inve inions or by all, it is poss proach to mea applications nies to asse nts and to ide RENCES with the m llows to obtai This consequen nce of obtainin ations need to
pany, the selec Productivity In lity, Data Qua these insights ogether with from the mo model consid by five pha on, control onsidered for the model is t productivity ecision. The th an impact on ent impact. T more predomin Data Quality; e “valorized” d haracteristics; d the “used” d d, it is difficul esult obtained udies, that ra h the expectati d by an OEE b ore the numb e and the nee
these results; al changes data, Informat m of the mode ation Asset of his a compari s that has the ment is also va
and the fact , it is possible y of applicatio extension of In addition, i estigated by insights com sible to concl asure Informat relevant insig ess their AS entify pattern main in a ntly ng a o be cted ndex ality the the odel ders ses: and the s to y to hree n all The nant the data the data lt to d by ange ions best bers eded the that tion el is f the ison aim alue that e to ons, the it is the ming lude tion ghts S-IS n of
Ahituv N. (1989). Assessing the value of information: Problems and Approaches. Proceedings of the International Conference on Information Systems (ICIS) 1989, 45.
Alavi M., Leidner D.E. (2001). Review: Knowledge management and knowledge management systems:
Conceptual foundations and research issues. MIS Quarterly, 25 (1), 107-136.
Black S.H, Marchand D. (1982). Assessing the value of information in organisations: a challenge for the 1980s. The Information Society, 1, 191-225.
Checkland P, Scholes J. (1990). Soft Systems Methodology in Action. John Wiley and Sons, New York.
Chopra, S., Sodhi M. (2004). Managing risk to avoid supply-chain breakdown. MIT Sloan Management Review, 46(1). Diewert Erwin W., Lawrence D. (1999). Measuring New Zealand’s Productivity, Treasury Working Paper, 99(5). Dretske F. I. (1981). Knowledge and the Flow of Information. MIT Press, Cambridge Mass.
Evans N., Price J. (2015). Information Asset Management capability: The Role of the CIO. Twenty-first Americas Conference on Information Systems, 2016.
Hackbarth G. (1998). The Impact of Organizational Memory on IT Systems. Proceedings of the Fourth Americas
conference on Information Systems, 588-590.
Infor publication (2015). Big Data in manufacturing: A compass for growth. Infor, New York.
Jumarie, Guy. (1990). Relative Information: Theories and Applications. Springer-Verlag Berlin Heidelberg, Berlin.
Ladley J. (2010). Making Enterprise Information
Management (EIM) Work for Business. Elsevier, Burlington.
Lewis, P. J. (1993). Linking soft systems methodology with data-focused information systems development. Information Systems Journal, July, 169-186.
Loshin D. (2009). Master Data Management. Elsevier, Burlington.
Machlup F., Mansfield U. (1983). The Study of Information: Interdisciplinary Messages. Wiley-Interscience, New York.
McAfee A., Brynjolfsson E. (2012). Big data: The
Management Revolution. Harvard Business Review, 90 (10), 1-9.
McKinsey Global Institute (2015). The Internet of Things: Mapping the Value Beyond the Hype. McKinsey & Company, New York.
McKinsey Global Institute (2011). Big data: The next frontier for innovation, competition, and productivity. McKinsey & Company, New York.
Mingers J. (2006). Realizing Systems Thinking: Knowledge and Action in Management Science. Springer, Berlin.
Mingers J. (1996). An Evaluation of Theories of Information with Regard to the Semantic and Pragmatic Aspects of Information Systems. Systemic Practice, 9(3), 187-209. Mithas, Sunil, Ramasubbu, Narayan, Sambamurthy V. (2011). How Information Management Capability Influences Firm Performance. MIS Quarterly, 35(1), 237-256.
OECD Organisation for Economic Co-Operation and Development (2001). Measuring Productivity. OECD Publications, Paris.
Oppenheim C., Stenson J., Wilson R.M.S. (2003). Studies on information as an asset I: definitions. Journal of Information Science, 29 (3), 159–166.
Politecnico di Milano (2017). Industria 4.0: la grande occasione per l'Italia. Observatory Industry 4.0, Milano.
PwC (2016). PwC’s 2016 Global Industry 4.0 Survey: Building the Digital Enterprise. PwC, London.
Spruit M., Pietzka K. (2015). MD3M: The master data management maturity model. Computers in Human Behavior, 51, 1068-1076.
Strassmann P.A. (2004). Defining and Measuring
Information Productivity. The Information Economics Press, New Canaan.
Taylor R.S. (1986). Value-added Processes in Information Systems. Praeger, Westport.
Vance D. M. (1997). Information, knowledge and wisdom: The epistemic hierarchy and computer-based information systems. Proceedings of the 1997 America’s Conference on Information Systems.
Von Krogh G. (1998). Care in Knowledge Creation. California Management Review, 40(3), 133-153.
