(c) jay Metadata Tech_transfermedia

Info­tain­ment and in-stream shopping

Some­ti­mes you spot a pair of snea­k­ers in a sce­ne, that pret­ty red dress the actress is wea­ring, or a pie­ce of fur­ni­tu­re you’d like to buy later on. Or you won­der whe­re tho­se impres­si­ve cliffs by the coast are that one can always see in the back­ground. And what was the name of the actor again who’s play­ing the sup­port­ing role in this film?

How start-up jay is making meta­da­ta from film pro­duc­tion actionable

The pro­duc­tion of films, series and docu­men­ta­ries also gene­ra­tes a lot of infor­ma­ti­on about film loca­ti­ons, prot­ago­nists, cos­tu­mes and props that has hard­ly been exploi­ted and dis­play­ed until now. If you’­re loo­king for a red dress, you goog­le and shim­my through shop­ping plat­forms using the typi­cal search words. You can use the second screen of your mobi­le pho­ne to look up the actors on Wiki­pe­dia and IMDb, but you won’t find any addi­tio­nal infor­ma­ti­on about the shoo­ting loca­ti­ons. Howe­ver, all of this data exists, it’s gathe­red tog­e­ther for the pro­duc­tion of a film - but then it goes to was­te becau­se the­re isn’t any cen­tral, digi­tal tool to store it.

One tech solu­ti­on for all film data

The Pots­­dam-based com­pa­ny trans­fer­me­dia has now laun­ched jay, a meta­da­ta tech­no­lo­gy offe­ring film pro­duc­tion com­pa­nies and dis­tri­bu­tors an auto­ma­ted solu­ti­on for coll­ec­ting and making data available. Whe­ther it’s data about film loca­ti­ons, props, actors or cos­tu­mes: ever­y­thing is lin­ked and stored on a sce­­ne-by-sce­­ne basis thanks to meta­da­ta soft­ware. The audi­ence no lon­ger has to use a second screen to access the infor­ma­ti­on becau­se ever­y­thing can now appear in the film - and inte­gra­ted in-stream shop­ping means that they can even buy the dis­play­ed pro­ducts directly.

Cus­to­mers of strea­ming ser­vices know a part of the prin­ci­ple from Ama­zon X-Ray whe­re you can access infor­ma­ti­on about the actors who are curr­ent­ly acting in a par­ti­cu­lar sce­ne so that it appears on the edge of the screen. Howe­ver, this is the result of an ela­bo­ra­te pro­cess whe­re sce­nes are divi­ded manu­al­ly and the infor­ma­ti­on about the prot­ago­nists is sub­se­quent­ly added. In addi­ti­on, Ama­zon only offers infor­ma­ti­on about the cast or the sound­track. jay, though, makes it pos­si­ble for the first time to have all of a film’s meta­da­ta gathe­red during pro­duc­tion - and to do this as part of an auto­ma­ted work­flow. Dis­tri­bu­tors then recei­ve the film direct­ly with the lin­ked data.

New busi­ness models con­se­quent­ly open up for broad­cas­ters and video-on-demand plat­forms with such ser­vices as info­tain­ment and in-stream shop­ping. “One of the big­gest chal­lenges is to help initia­te the chan­ge and show what pos­si­bi­li­ties we have,” says Irmela Wro­ge­mann, who speaks with broad­cas­ters in her role as Chief Mar­ke­ting Offi­cer (CMO) at trans­fer­me­dia. “This is very much in the direc­tion of chan­ge manage­ment, i.e. how to think, plan and play out con­tent in the future. And it’s also about ope­ning up struc­tures. Edi­to­ri­al, pro­duc­tion, dis­tri­bu­ti­on and tech­no­lo­gy have few inter­faces in the world of broad­cas­ting. But we could bene­fit a lot if we were to work tog­e­ther in a more joi­n­ed up way. And ulti­m­ate­ly offer the audi­ence a bet­ter con­tent experience.”

trans­fer­me­dia has been rese­ar­ching in the field of meta­da­ta for a long time. The team hea­ded up by Peter Effen­berg draws on many years of expe­ri­ence in film pro­duc­tion, tech­ni­cal exper­ti­se and inten­si­ve rese­arch on metadata.

In-stream shop­ping: buy­ing direct­ly ins­tead of spen­ding a long time googling

We know the prin­ci­ple of shop­pable con­tent from social media. jay is now making in-stream shop­ping pos­si­ble for the film indus­try as well. This is of bene­fit to the strea­ming mar­ket becau­se ser­vices like Net­flix, Dis­ney and Co. are loo­king for new busi­ness models. Sub­scrip­ti­on num­bers are start­ing to fall and the prin­ci­ple of con­stant­ly pro­du­cing new con­tent is rea­ching its limits. Some pro­vi­ders are alre­a­dy thin­king about intro­du­cing adver­ti­sing as a way of gene­ra­ting new sources of inco­me. This is whe­re jay comes in with the pos­si­bi­li­ty of in-stream shop­ping. Pro­duct pla­ce­ment as regu­lar­ly seen in major block­bus­ters from James Bond to Spi­der Man works in a simi­lar way - but in-stream shop­ping is much more acces­si­ble, direct and mea­sura­ble. The view­ers don’t have to rese­arch the manu­fac­tu­rer or brand after wat­ching the film; ins­tead of that, they can have the right pro­duct offe­red to them in a direct way with just one click whilst they are strea­ming. The basis for this is digi­tal meta­da­ta coll­ec­ted by jay during pro­duc­tion. We will pro­ba­b­ly soon see new pro­fes­si­ons appearing on film sets to gather the data, and a data wrang­ler may then beco­me part of the stan­dard line-up in a film crew. Becau­se it’s worth brin­ging all of the data tog­e­ther from the screen­play or script con­ti­nui­ty as part of the ongo­ing pro­cess. Data reco­gni­ti­on by arti­fi­ci­al intel­li­gence isn’t curr­ent­ly an opti­on and won’t be the case in the near future eit­her. A red dress, for exam­p­le, is easy to mix up, and the algo­rithm here is still overst­ret­ched and pro­ne to error.

Under­stan­ding the view­ers’ needs

jay not only faci­li­ta­tes the cen­tral ope­ra­ti­on and sto­rage of data - it also ana­ly­ses and veri­fies it by using the Data­Qua­li­ty fea­ture to adapt the infor­ma­ti­on to the dis­tri­bu­tors’ stan­dards. It is then the respon­si­bi­li­ty of the broad­cas­ters and strea­ming ser­vices to inte­gra­te the data cor­rect­ly into the respec­ti­ve plat­form. jay sup­ports them in an advi­so­ry capa­ci­ty in the imple­men­ta­ti­on and is curr­ent­ly con­duc­ting user rese­arch to have a bet­ter under­stan­ding of the view­ers’ needs: what addi­tio­nal infor­ma­ti­on are they par­ti­cu­lar­ly inte­res­ted in? How do they want to inter­act with the data? And what is important to them in terms of design?

The chan­ge in thin­king by the broad­cas­ters is not the only issue; trans­fer­me­dia has also encoun­te­red reser­va­tions about a pos­si­ble sell­out of a film. A cri­ti­cism that is under­stan­da­ble, says Wro­ge­mann. In-stream shop­ping does not suit every for­mat. It is not sui­ta­ble for art­house films, but it could work, for exam­p­le, for roman­tic come­dies or series whe­re mer­chan­di­sing plays a role. Moreo­ver, info­tain­ment is now just as much a pos­si­bi­li­ty as shop­ping. jay has alre­a­dy rea­li­sed a first use case with the rbb docu-dra­­ma “Kennedy’s Love for Euro­pe”. Thanks to the new con­tent level, the view­ers can access fur­ther infor­ma­ti­on about the actors, his­to­ri­cal events and loca­ti­ons direct­ly in the play­er. rbb is thus the first Ger­man broad­cas­ter to under­ta­ke this kind of in-stream interaction.

Dis­tri­bu­tors such as broad­cas­ters and VoD plat­forms as well as pro­du­cers are jay’s key tar­get groups along with the view­ers as end users. While jay’s tech­no­lo­gy opens up new levels of con­tent for the audi­ence, the B2B tar­get group can gene­ra­te sources of inco­me through in-stream shop­ping, with film pro­duc­tions able to tap into a wider ran­ge of finan­cing sources fur­ther upstream. The first major use case for in-stream shop­ping is now being plan­ned by jay with a broad­cas­ter for next year.

About MTH Blog

The media technologies of the future are already being used today – not only in the entertainment sector, but also in a wide variety of industries. Christine Lentz meets up with tech enthusiasts, established companies and researchers for our monthly MediaTech Hub Potsdam blog to tell the stories behind the innovative business models.