Netflix – Redefining Creativity

Read­ing time: 4 mins

Net­flix was one of the first online stream­ing ser­vices and is by far the most pop­u­lar, but is that all it deserves cred­it for? By track­ing what visu­al and aur­al trig­gers you respond to, Net­flix have tak­en con­tent cre­ation and pro­mo­tion to the next lev­el. Mac­quar­ie Uni actu­ar­i­al stu­dent Abhishek Maran explores the role of data track­ing in Netflix’s cre­ative suc­cess.

Net­flix was one of the first cor­po­ra­tions to com­bine data with the enter­tain­ment indus­try.

Mov­ing into the online stream­ing sphere in 2007, Net­flix quick­ly realised the poten­tial for the col­lec­tion of large amounts of data by launch­ing its USD$1M com­pe­ti­tion, the ‘Net­flix Prize’.

The com­pe­ti­tion was aimed at cre­at­ing the best algo­rithm to improve the accu­ra­cy of pre­dic­tions about how much some­one was going to enjoy a movie based on their pri­or pref­er­ences.

Since then, Net­flix has evolved into a data min­ing behe­moth, bring­ing you con­tent that you can binge watch repeat­ed­ly with­out leav­ing the com­fort of your bed.

As Net­flix pop­u­lar­i­ty and user base has expand­ed, the service’s pre­dic­tive capa­bil­i­ties have improved sig­nif­i­cant­ly.

This has only been pos­si­ble through the col­lec­tion of a wider range of data points.

Data Set A

  • Times when you stop, pause, rewind or fast for­ward the con­tent.
  • Days and times when you watch cer­tain con­tent such as rom-coms on Sat­ur­day night at 7pm, and Fam­i­ly Guy on Tues­days at 10pm.
  • The spe­cif­ic dates you watch (e.g. what movies are pop­u­lar on Valen­tines’ day)
  • Your loca­tion when you watch such as your home or at work.
  • What device you use to watch con­tent. (e.g. TV for movies, Lap­top for binge watch­ing shows in bed)
  • At what points dur­ing the show you stop watch­ing and move on. In addi­tion, they also track whether you resume watch­ing lat­er.
  • What rat­ing you assign a piece of con­tent.
  • Your search his­to­ry.

Whilst the above seem stan­dard for com­pa­nies of Netflix’s ilk, a few oth­ers are also known to be tracked:

Data Set B – where Net­flix sets itself apart from com­peti­tors

  • How you browse and scroll through selec­tions. I.e. Do you pause and read descrip­tions, or just skim through until you see a title/cover you like?
  • The types of trail­ers, pro­mo­tion­al posters, words, colours and sounds you respond best to i.e. most like­ly to click on, and fol­low through.

 Data Set A can give you a tai­lored ser­vice, but Net­flix has gone above and beyond with Data Set B, which focus­es on cre­at­ing a com­plete­ly unique iter­a­tion of its ser­vice for each of its users.

The ‘per­fect’ TV show

In 2013, the first ‘Net­flix Orig­i­nal’ was released: House of Cards.

Pro­duced by David Finch­er and star­ring Kevin Spacey, the show was part of a metic­u­lous­ly orches­trat­ed plan Net­flix con­coct­ed as they aimed to release the ‘per­fect’ TV Show.

From their var­i­ous points of data, it became clear that their users pre­ferred movies star­ring Kevin Spacey and movies pro­duced by David Finch­er who is also involved in ‘The Fight Club’ and ‘The Social Net­work’. So as the say­ing goes ‘por que no los dos’?

Soon enough, the series was released in typ­i­cal binge watch­ing fash­ion such that all the episodes were avail­able at once. Cur­rent­ly the series has been renewed for a 5th sea­son and is rat­ed 84% on Rot­ten Toma­toes. How­ev­er, the series’ suc­cess isn’t entire­ly reliant on Spacey or Finch­er, but rather the pro­mo­tion­al strat­e­gy cre­at­ed from study­ing Data Set B.

From analysing the gath­ered data, dif­fer­ent ver­sions of the poster con­tain­ing dif­fer­ent themes/colours/actors were cre­at­ed to appeal to a wider range of view­ers.

Addi­tion­al­ly, the dis­tri­b­u­tion of the mul­ti­ple pre­view trail­er was based on people’s inter­ests e.g. those who adored Spacey, were shown a trail­er where he was the promi­nent fig­ure.

Tak­ing the indus­try to new heights

For­tu­nate­ly for Net­flix, ‘House of Cards’ was not a fluke as it has gone on to release oth­er suc­cess­ful TV shows such as “Marvel’s Dare­dev­il”, “Orange is the New Black”, “Stranger Things” and many more.

More­over, as Dave Hast­ings, Netflix’s direc­tor of prod­uct ana­lyt­ics, and data sci­ence said at a Whar­ton Cus­tomer Ana­lyt­ics Ini­tia­tive Con­fer­ence in 2015 “You do not make a $100 mil­lion invest­ment these days with­out an awful lot of ana­lyt­ics,”.  

Net­flix have fur­ther forged the road of big bud­get pro­duc­tions hav­ing released 126 ‘orig­i­nal series’ and films in 2016 alone (far more than any oth­er Amer­i­can Net­work), pump­ing $6 Bil­lion into the indus­try.

Clear­ly, the col­lec­tion and analy­sis of data is a key ingre­di­ent in the recipe to rev­o­lu­tionise the tele­vi­sion indus­try and push it to new heights of orig­i­nal­i­ty and cre­ativ­i­ty.

Like all enter­tain­ment com­pa­nies, Net­flix wasn’t 100% suc­cess­ful. It expe­ri­enced its fair share of poor rat­ings. But, unlike oth­er pro­duc­tion firms, Net­flix is play­ing on anoth­er lev­el by exper­i­ment­ing with data insights to pro­duce new con­tent. This begs the ques­tion; does data or cre­ativ­i­ty rule deci­sion mak­ing? As Net­flix con­tin­ues to pro­duce and release their Orig­i­nal films and shows, it is impor­tant to realise the effect cre­ativ­i­ty has on the suc­cess of a show. Going back to “House of Cards” it is evi­dent that a lot of inspi­ra­tion was derived from the orig­i­nal British ver­sion of the show. More impor­tant­ly, the data gath­ered from peo­ple watch­ing the British ver­sion was used to implement/retain (i.e. the audi­ence reten­tion times, and the paus­ing times of view­ers) the pop­u­lar fea­tures whilst using cre­ativ­i­ty to improve the oth­er unpop­u­lar fea­tures.

On the oth­er hand, the Net­flix reboot of the pop­u­lar series “Arrest­ed Devel­op­ment” was received extreme­ly poor­ly by exist­ing fans due to its lack of loy­al­ty to the orig­i­nal for­mat of the show. Whilst Net­flix may have had good inten­tions with reboot­ing the series, in this case it’s like­ly that the data cre­at­ed from binge watch­ing users was gross­ly mis­in­ter­pret­ed. How­ev­er, such strat­e­gy of reviv­ing dead TV shows does bring expo­sure from out­side Netflix’s user base, hence gen­er­at­ing mon­ey from them in the short run, but is this strat­e­gy the most prof­itable in the long run?

With more new shows and films to be com­mis­sioned, the expense of doing so will almost cer­tain­ly increase on a year­ly basis. Hence it is like­ly that Net­flix may shift their focus from cre­at­ing good pieces of con­tent inspired by data towards one entire­ly dic­tat­ed by data for the sake of pro­tect­ing their bank bal­ance. Whilst Big Data is inher­ent­ly use­ful towards cast­ing and set­ting cre­ative guide­lines, it must still be remem­bered that the flow and illog­i­cal nature of cre­ativ­i­ty must be present for a piece of con­tent to be deemed excel­lent and human.

CPD Actuaries Institute Members can claim two CPD points for every hour of reading articles on Actuaries Digital.

About the author

Abhishek Maran

Abhishek Maran, currently in his third year of Actuarial Studies at Macquarie University, is the President of the Actuarial Students' Society (ASSOC). His pursuits in entrepreneurship saw him start his own tutoring agency in July 2016 and is also an avid reader. He is interested in pursuing a career as an Investment Analyst and is looking forward to using his actuarial skills in this field.

Comment on the article (Be kind)


No Comments

Also this month