Data sharing in industrial ecosystems

Data sharing in industrial ecosystems (final part)

by Prof. Dr. Boris Otto, Prof. Dr. Niko­laus Mohr, Mat­thi­as Rog­gen­dorf, PhD and Tobi­as Gug­gen­ber­ger

Suc­cess ingre­dients of suc­cess­ful data sharing models

Many indus­tri­al data eco­sys­tems are still in their nas­cent sta­ges and many play­ers are only taking their first steps in this area, often in the form of bila­te­ral data sharing. Howe­ver, the ear­ly suc­ces­ses that do exist pro­vi­de a basis for iden­ti­fy­ing steps to incre­a­se the chan­ces that an indus­tri­al eco­sys­tem will pro­sper. As more com­pa­nies par­ti­ci­pa­te in data eco­sys­tems, this list will surely evol­ve:

  1. Sys­te­ma­ti­cal­ly assess busi­ness oppor­tu­nities from sharing your own data and gai­ning access to new data: In a first step, it is important for orga­niz­a­ti­ons to under­stand an ecosystem’s poten­ti­al in terms the busi­ness it will gene­ra­te rather than the data the com­pa­ny will share. Such an approach requi­res loo­king broad­ly at both inter­nal app­li­ca­ti­ons and oppor­tu­nities that can be co-crea­ted with cus­to­mers, or even new busi­ness they can build out­side the core. In some cases, addi­tio­nal value may come from sel­ling data alo­ne, while in others it may invol­ve some give and take. Com­pa­nies often fail to arti­cu­la­te their expec­ta­ti­ons clear­ly and unde­re­sti­ma­te the time nee­ded to achie­ve bene­fits. Also, it is cru­cial to under­stand the main actors in the poten­ti­al busi­ness eco­sys­tem – for examp­le, who will need to play the orches­tra­tor and who are par­ti­ci­pants – and what mecha­nisms would incen­ti­vi­ze them to share their data (for examp­le, if value crea­ti­on is uneven­ly dis­tri­bu­t­ed).

  2. Defi­ne your own role and opti­ons: After under­stan­ding the busi­ness oppor­tu­ni­ty offe­red by the eco­sys­tem and its ope­ra­ting pro­to­col, com­pa­nies need to tho­rough­ly assess their stra­te­gic opti­ons. The­se might invol­ve beco­m­ing a data owner, a par­ti­ci­pant in an emer­ging eco­sys­tem, or a pro­vi­der of ser­vices. The choice of this role lar­ge­ly deter­mi­nes the requi­red invest­ments, the orga­niz­a­tio­nal set­up, and the resour­ces nee­ded to dri­ve imple­men­ta­ti­on from both the busi­ness and tech­no­lo­gy side.

  3. Crea­te a value- and risk-dri­ven approach to data: A sys­te­ma­tic scan of cor­po­ra­te and tran­sac­tio­n­al data can reve­al the value that can be crea­ted and the intel­lec­tu­al pro­per­ty that might be at sta­ke. Opti­ons should be con­si­de­red for saniti­zing or aggre­ga­ting the data so it can still be used with part­ners. Methods for mana­ging data value and risk should be put in place for data assets asses­sed as stra­te­gic.

  4. Estab­lish sus­tainab­le data gover­nan­ce: Once rele­vant data has been iden­ti­fied, sca­ling up the approach requi­res sui­ta­ble data gover­nan­ce. This invol­ves estab­li­shing busi­ness owners­hip of inter­nal data so that it can be shared extern­al­ly and inte­gra­ted quick­ly with data from new sources. Com­pa­nies may also want to set up a set of data manage­ment tools, such as a data cata­log (or inter­nal data mar­ket­place), as well as a basic tools for topics such as tracking data lineage, clas­si­fy­ing data risk, or docu­men­ting data access poli­ci­es. Data part­ners­hip mana­gers may be nee­ded to dri­ve date iden­ti­fi­ca­ti­on and com­mer­cia­liz­a­ti­on.

  5. Embark on a data plat­form stra­te­gy: Often, a company’s lega­cy IT sys­tems are high­ly com­plex, and its data is silo­ed in ope­ra­tio­nal app­li­ca­ti­ons. Infle­xi­ble sys­tems and data models deter­mi­ned by soft­ware ven­dors make the data pro­duc­tion side of the equa­ti­on dif­fi­cult to chan­ge. The data con­sump­ti­on side, in turn, is easier to evol­ve. The key to gai­ning trac­tion is a fle­xi­ble data con­sump­ti­on archi­tec­tu­re that enab­les easy blen­ding of data, real-time data pro­ces­sing, and the use of ana­ly­tics and AI. In many cases, the main work ent­ails making exis­ting data acces­si­ble to such an ana­ly­tics envi­ron­ment rather crea­ting the envi­ron­ment its­elf, as cloud-based ana­ly­tics solu­ti­ons are fast to set up and use.
  6. App­ly a test-and-learn approach befo­re sca­ling: Com­pa­nies should not be afraid to learn by doing. Set an aggres­si­ve goal of taking a first mini­mal via­ble pro­duct live with inter­nal or exter­nal cus­to­mers wit­hin a few weeks. Start with use cases of limi­ted com­ple­xi­ty that con­tri­bu­te to over­all stra­te­gy. Find a friend­ly part­ner or cus­to­mers who is exci­ted to learn with you. Com­mu­ni­ca­te your suc­ces­ses and grow your capa­bi­li­ties along the way.

The mar­ket for data sharing plat­forms that gua­ran­tee data sov­er­eig­n­ty is still nas­cent. Fur­ther­mo­re, com­mer­cial approa­ches are dif­fi­cult to estab­lish (except in niches) due to the need for a neu­tral sup­ply­ing enti­ty (a pre­re­qui­si­te that may inter­fe­re with com­mer­cial inte­rests).

Play­ers that enga­ge ear­ly can shape the mar­ket and build high­ly inno­va­ti­ve busi­ness models and pro­ducts or ser­vices that com­pe­ti­tors can­not easi­ly copy. In some cases, such solu­ti­ons beco­me add-ons to the exis­ting busi­ness; in others, the supe­rio­ri­ty of the new solu­ti­ons could ful­ly dis­rupt the indus­try.

A key requi­re­ment for the pro­li­fe­ra­ti­on of indus­tri­al eco­sys­tems is the exis­tence of stan­dards for data inter­ope­ra­bi­li­ty, por­ta­bi­li­ty and sov­er­eig­n­ty. Pro­mi­sing stan­dar­di­z­a­ti­on initia­ti­ves such as Inter­na­tio­nal Data Spaces (IDS) and Trus­ted Cloud – as well as the recent­ly star­ted GAIA‑X pro­ject – are steps into the right direc­tion.