Data Man­age­ment: Why a Stra­tegic Mind­set Is Essential

More and more com­pan­ies are com­ing to real­ise that “Data Man­age­ment” is play­ing an evolving role in their day-to day busi­ness. But, at present, it is hard to pin­point a clear, sin­gu­lar defin­i­tion of this term. When look­ing up the defin­i­tion of Data Man­age­ment one is likely to stumble across dif­fer­ent pro­pos­als and ideas regard­ing how to approach Data Man­age­ment at your com­pany. In this blog we will sum­mar­ise our key-find­ings and exper­i­ences to provide you with some guidance.

Today, com­pan­ies are gen­er­at­ing more data than ever before. In 2020 up to 2.5 quin­til­lion bytes of new data are being cre­ated every single day[1]. To keep up with this huge amount of gen­er­ated data it needs to be prop­erly man­aged and stored. In 2006 Clive Humby, an UK Math­em­atician and archi­tect of Tesco’s Club­card is cred­ited with coin­ing the phrase: “Data is the new oil. It is valu­able, but if unre­fined it can­not really be used. It must be changed into gas, plastic, chem­ic­als, etc. to cre­ate a valu­able entity that drives prof­it­able activ­ity; data must be broken down, ana­lyzed for it to have value.”

This state­ment should encour­age com­pan­ies to con­sider their data as an asset. There­fore, proper man­age­ment is a pre-requis­ite for both ensur­ing trust, as well as for incor­por­a­tion into busi­ness pro­cesses. To go even fur­ther, accord­ing to a Gart­ner sur­vey, nearly 80% of exec­ut­ives agree com­pan­ies will lose com­pet­it­ive advant­age if they do not effect­ively util­ize data, and 49% say data can be used to decrease expenses and cre­ate new aven­ues for innov­a­tion.[2] Hence, lack­ing adequate Data Man­age­ment could not only lead to dis­or­gan­ised data swamps, poor data qual­ity, and incom­plete data sets, but could also limit organ­isa­tions in the scope of their busi­ness ana­lyt­ics and long-term plan­ning. Data has also grown in import­ance as busi­nesses are sub­jec­ted to an increas­ing num­ber of reg­u­lat­ory com­pli­ance require­ments, includ­ing data pri­vacy and pro­tec­tion laws such as GDPR.

Tak­ing this into account, it should become clear why Data Man­age­ment applies not only to the IT depart­ment but rep­res­ents a struc­tural issue for an entire com­pany. We will do our best to give a high-level over­view of which areas to take into con­sid­er­a­tion when (re-)structuring your company’s Data Man­age­ment strategy.

“Top-Down” Data Management

Firstly, organ­isa­tions should work on align­ing their Data Strategy with their busi­ness strategy. A busi­ness strategy refers to the actions and decisions that a com­pany takes to reach its busi­ness goals and be com­pet­it­ive in its industry. As data hand­ling becomes more import­ant, it is cru­cial that data strategy closely sup­ports busi­ness strategy, as opposed to being built on “latest and greatest” tech­no­logy aspir­a­tions. Data which is pur­pose­fully col­lec­ted and ana­lysed with respect to the pains your busi­ness might have serves the over­all scope of your busi­ness. As an example, if your busi­ness object­ive is to double sales, your busi­ness strategy may be ori­ented towards build­ing more appeal­ing products. To find out how to optim­ize the appeal of your products, your data strategy should focus on col­lect­ing and ana­lys­ing data which describes how your cus­tom­ers use and bene­fit from your products. Do not for­get to con­sider external data as well.

To eval­u­ate and extend your company’s Data Man­age­ment helps to be aware of which data is used in day-to-day oper­a­tions, as well as poten­tially required data for future pro­jects. This is one reason why input from the busi­ness depart­ments is cru­cial. Not only should the IT depart­ment be involved in set­ting up a Data Man­age­ment struc­ture, but also man­age­ment and selec­ted busi­ness lead­ers. Jointly they should cooper­ate on com­pany-wide data over­sight struc­tures – you might want to con­sider cre­at­ing a steer­ing com­mit­tee and a Chief Data Officer (CDO) pos­i­tion for this activ­ity. The busi­ness provides the input needed to rank and pri­or­it­ize data-related activ­it­ies: a method we have seen to work well, both to get star­ted as well as an iter­at­ively repeated activ­ity, is Use Case clus­ter­ing. Here, “Use Cases” are the data-related activ­it­ies of con­crete busi­ness ini­ti­at­ives. By rank­ing these you pri­or­it­ize them accord­ing to busi­ness relevance/value which enables your organ­iz­a­tion to be data-driven. This com­piled list should provide guid­ance regard­ing over­all data strategy as well as Data Man­age­ment structure.

Essen­tially, the setup of “top-down” Data Man­age­ment is not only rel­ev­ant for IT, but also impacts the entire enter­prise through col­lab­or­at­ive future planning.

Data Gov­ernance

Data Gov­ernance is another cent­ral com­pon­ent of Data Man­age­ment. More than half of organ­iz­a­tions, how­ever, lack a formal Data Gov­ernance frame­work.[3] It com­prises the pro­cesses and respons­ib­il­it­ies rel­ev­ant to the qual­ity and secur­ity of the data used in an organ­isa­tion. These con­sid­er­a­tions are espe­cially rel­ev­ant these days due to GDPR regulations.

Data Gov­ernance pro­cesses and respons­ib­il­it­ies entail a num­ber of dis­tinct roles. These roles, such as data stew­ard or data owner, are clearly defined to effect­ively ensure Data Gov­ernance. When set­ting up a frame­work, hav­ing a clear defin­i­tion of Data Gov­ernance is only the first step. Equally import­ant is imple­ment­ing the frame­work into daily busi­ness. The real chal­lenge here is for the pro­cesses to be used and lived, and roles to be owned. Gart­ner analyst Nick Heu­decker estim­ates that 85% of data pro­jects fail[4] due to dif­fi­culty in cul­ture and chal­lenges in change man­age­ment, as well as man­age­ment block­ing pro­gress. These stat­ist­ics show that top-down guid­ance and sup­port is key. If man­age­ment sets a good example, oth­ers in their organ­isa­tion are more likely to follow.

Data Lever­aging

An addi­tion­ally import­ant task in work­ing towards true Data Man­age­ment is mak­ing data eas­ily access­ible to users. You can com­pare this to logist­ics: the required mater­ial (data) should be avail­able in suf­fi­cient volume, neces­sary qual­ity, via the cor­rect channel/tool at the needed time. This is not only true for BI and report­ing. Data Sci­ent­ists often find them­selves spend­ing a huge amount of time (up to 60% of their time accord­ing to a Data Sci­ence report[5]) in alloc­at­ing the right data for ana­lysis, whereas they could prefer­ably go straight to dig­ging for insights.

The key here is effi­ciency and con­veni­ence: to achieve these, enabling self-ser­vice is the way to go. This entails build­ing data infra­struc­ture which enables, for example, Data Sci­ent­ists to retrieve the right data in the right format for their ana­lysis. There are plenty of ways to make this pos­sible – be it via data lakes, cata­logues, repos­it­or­ies, ware­houses and prob­ably more – this is indi­vidual to every com­pany and their situ­ation. You need to find out what works best for your organisation.

Together with proper Data Gov­ernance you can make effort­less access­ible data for author­ized users hap­pen. Plus, with access­ib­il­ity comes easier collaboration​ for your team.

Data Source Integration

Enabling self-ser­vice of data is already a big effi­ciency improve­ment. Nev­er­the­less, com­pan­ies usu­ally deal with a huge diversity of data types and sources, which can also be extremely dynamic, updat­ing often. Hence, flex­ib­il­ity is hugely advantageous.

It is easy to under­es­tim­ate the work required to cor­rectly man­age Data Source Integ­ra­tion. For a given pro­ject, the effort spent on integ­rat­ing data sources and formats can rep­res­ent 60% of the entire effort[6].

The key here is to reduce the com­plex­ity of data integ­ra­tion. Tools can help with increas­ing flex­ib­il­ity. If you are not spe­cial­ized in build­ing sys­tem integ­ra­tions, and you have other busi­ness focusses, we recom­mend using stand­ard solu­tion sys­tems such as Ab Ini­tio, Cloudera, Cloud nat­ive ser­vices, etc. In-house integ­ra­tion solu­tions are usu­ally higher main­ten­ance and changes in the data sys­tem land­scape requires too much effort.

Bene­fits of Proper Data Management

To sum­mar­ize, for one to gain value from their data it needs to be man­aged (or, along the lines of the oil ana­logy, refined) accord­ingly. A well-executed Data Man­age­ment strategy can boost your entire busi­ness, not only IT.

Data Man­age­ment can help com­pan­ies to gain a com­pet­it­ive edge over their busi­ness rivals, both by improv­ing oper­a­tional effect­ive­ness and enabling bet­ter decision-mak­ing. Organ­isa­tions which make extens­ive use of their data can also become more agile, mak­ing it pos­sible to spot mar­ket trends and quickly adjust towards tak­ing advant­age of them.

Ulti­mately, a solid approach to Data Man­age­ment res­ults in more effect­ive use of data, a bet­ter under­stand­ing of cus­tom­ers, as well as value cre­ation through provid­ing new (data-) products and offer­ings. This can open up new sources of rev­enue and game-chan­ging oppor­tun­it­ies in an increas­ingly com­pet­it­ive busi­ness world.

Links, Ref­er­ences

[1] https://www.dihuni.com/2020/04/10/every-day-big-data-statistics‑2–5‑quintillion-bytes-of-data-created-daily/#:~:text=Every%20Day%20Big%20Data%20Statistics%20%E2%80%93%202.5%20Quintillion%20Bytes%20of%20Data%20Created%20Daily,-This%20post%20was

[2] https://www.gartner.com/en/newsroom/press-releases/2019–11-7-gartner-says-data-and-cyber-related-risks-remain-top-worries-for-audit-executives

[3] https://www.gartner.com/en/newsroom/press-releases/2019–11-7-gartner-says-data-and-cyber-related-risks-remain-top-worries-for-audit-executives

[4] https://www.techrepublic.com/article/85-of-big-data-projects-fail-but-your-developers-can-help-yours-succeed/

[5] https://visit.figure-eight.com/data-science-report.html

[6]https://www.researchgate.net/publication/293299036_Integrating_data_sources_from_different_development_environments_An_E-LT_approach