Big data is the buzz word of 2017 but the art in­dus­try has been slow to re­spond to it. We ask why, and its fu­ture in a mar­ket where play­ers have thrived on its opac­ity.

The Peak (Hong Kong) - - Contents - STORY CHRISTIE LEE

Big data is the buzz word of 2017, but the art in­dus­try has been slow to re­spond to it. We ask why, and its fu­ture in an in­dus­try where play­ers have thrived on its opac­ity

The art world is on a data sci­en­tist-buy­ing spree. While data has al­ways been pre­cious, auc­tion houses and deal­ers op­er­at­ing at the top-end of the mar­ket are lever­ag­ing quan­ti­ta­tive mar­ket analy­ses, al­go­rithms and high fre­quency price in­dices to gain a com­pet­i­tive edge in an in­creas­ingly crowded mar­ket place. The spate of ac­qui­si­tions over the past year at­tests to this. Fol­low­ing Sotheby’s ac­qui­si­tion of The Mei Moses Art In­dices (now Sotheby’s Mei Moses) last Oc­to­ber, on­line gi­ant art­net poached Fabian Bo­cart, co-founder of Tutela Cap­i­tal, to head art­net An­a­lyt­ics a month later. Ear­lier this year, Artsy ac­quired Ar­tad­vi­sor, the data sci­ence start-up founded by AI ex­pert Hugo Liu.

If the above is of any in­di­ca­tion, the big data arms race is on, but the ques­tion re­mains: in an opaque in­dus­try where peo­ple guard their pri­vacy like a po­lar bear guards its cubs, what could be big data’s big­gest po­ten­tial? And what are its lim­its?

To an­swer that, we first need to de­fine what big data is. Ac­cord­ing to Daniel Doubrovkine, chief tech­ni­cal of­fi­cer at Artsy, big data is "sta­tis­ti­cally sig­nif­i­cant in­for­ma­tion that can yield real in­sights with the help of al­go­rithms and ma­chines.” Thus, while a cen­tury-old auc­tion house might have been metic­u­lously not­ing down the names of their clients, their habits and de­tails of their pur­chases since day one, this in­for­ma­tion – how­ever rich – can­not au­to­mat­i­cally be clas­si­fied as big data.

To achieve ef­fec­tive data min­ing, a siz­able amount of data is re­quired. In this re­gard, it does ap­pear that the art in­dus­try is ripe for big data dis­rup­tion. In 2005, the TEFAF Art Mar­ket Re­port put the value of the art mar­ket at US$35.9 bil­lion; in 2015, that num­ber has risen to US$63.76, trans­lat­ing to a growth of 77.6 per cent.

If we pan in onto the on­line art mar­ket – where data is ar­gubly, more eas­ily track­able – the num­bers are even more en­cour­ag­ing. On­line sales are at 3.75 bil­lion in 2016, up 15 per cent from the pre­vi­ous year, ac­cord­ing to the His­cox On­line Art Trade Re­port 2017. On­line sales now take up a 8.4 per cent share of the over­all art mar­ket, com­pared with 7.5 per cent in 2015.

Art has be­come an al­ter­na­tive as­set class of its own but there are still those who be­lieve that it should ‘rise above’ mon­e­tary con­cerns, ac­cord­ing to Ro­man Kraussl, a fi­nance pro­fes­sor at the Uni­ver­sity of Lux­em­bourg.

“Pure fi­nance con­sid­er­a­tions are still be­ing treated as evil, as non-arty, as de­stroy­ing the pure, the true value of art,” Kraussl, a re­search fel­low at the Emory Cen­ter for Al­ter­na­tive In­vest­ments, notes. “It has noth­ing to do with the fact that we do not have enough data avail­able. I am working with a data­base of over 10,000,000 auc­tion records, go­ing back to 1900.”

It’s also no se­cret that the art in­dus­try thrives on its very opac­ity. Even though pub­lic auc­tion records are well, pub­lic, the dealer seg­ment of the mar­ket re­mains pri­vate. Deal­ers are un­der no le­gal obli­ga­tion to dis­close the amount they prof­ited from a sale, and it has long been a mantra of the shrewdest in the trade to ‘buy low, sell high’. Added to this the re­al­ity that more and more col­lec­tors are look­ing to the dis­cre­tion that dealer sales of­fer. Ac­cord­ing to the 2017 TEFAF re­port, dealer sales had a 20 per cent growth, from 23.25 bil­lion in 2015 to 27.9 bil­lion in 2016. Mean­while, auc­tion sales are down, from 21.08 bil­lion in 2015 to 16.9 bil­lion in 2016.

“In such a mar­ket en­vi­ron­ment, some peo­ple do make big prof­its, akin to an in­sider IPO mar­ket. My re­search, my analy­ses, my in­dices, based on big data brings trans­parency to the mar­ket, and not ev­ery­one is happy about that,” notes Kraussl.

“Pri­vacy - and the ab­sence of reg­u­la­tion - is what makes the mech­a­nisms of the art mar­ket very dif­fer­ent from those for crude oil, stocks or prop­erty. There is no cen­tral reg­istry where sales are recorded,” adds David Hop­per, founder of Art Logic,

And it doesn’t look like there will be one soon. In a ne­go­ti­a­tion-based mar­ket­place, where one dealer might be able to sell a work for US$10,000, and a less shrewd one could also sell it for US$ 9,000, no gal­lerist would want to dis­close the fi­nal sell­ing price for fear that the buyer would de­mand a re­fund for the dif­fer­ence.

“Changes in tech­nol­ogy have made data anal­y­sis and pre­dic­tion more com­mon­place in the art mar­ket, yet de­mand for large amounts of data upon which pre­dic­tions can be based will out­strip sup­ply,” adds an am­biva­lent Hop­per.

As ex­pected, Artsy, vo­cal about its goal to ‘democra­tise’ art through tech­nol­ogy since its found­ing in 2009, re­mains the most op­ti­mistic about the dis­rup­tive po­ten­tial of big data. “For the first time we have data from both the pri­mary and the sec­ondary mar­kets that can power ma­chine learn­ing mod­els and some early AI,” Doubrovbine notes.

In an in­dus­try where there are huge data gaps, Doubrovkine also ar­gues that data breadth is more im­por­tant than data depth. “mar­kets typ­i­cally be­have con­sis­tently at scale and with broad enough data we can in­fer the over­all sys­tem.”


Dif­fer­ent com­pa­nies have dif­fer­ent ways of lever­ag­ing data, and at Artsy, that comes in the form of The Art Genome Pro­ject (TAGP). De­vised by six art his­to­ri­ans, one thing that TAGP does is to cal­cu­late the im­por­tance of an art­work within the artist’s body of work. Tak­ing into ac­count var­ied fac­tors, in­clud­ing the size of the work, the artist’s pop­u­lar­ity, and whether the work is in a mu­seum or pri­vate col­lec­tor’s

hand, a iconity score is gen­er­ated. It’s in­ter­est­ing (and re­fresh­ing) to see that even in the age of so­cial me­dia, the art world still abide by cer­tain tra­di­tions – while a work might score 2 points for be­ing in a mu­seum, it would only score 0.5 points if its artist has 10,000-plus fol­low­ers on so­cial me­dia.

Prior to join­ing art­net, Bo­cart of Tutela Cap­i­tal was per­haps most fa­mous for putting forth the Fair Value Mea­sure, or IFRS 13, an in­ter­na­tion­ally recog­nised val­u­a­tion for works of art that en­ables en­ti­ties like banks and hedge funds to lever­age art while con­form­ing to reg­u­la­tory de­mands for dis­clo­sure. In an in­ter­view with The Art News­pa­per, the for­mer trader stated his am­bi­tion to cor­rect the art mar­ket’s ma­jor dis­crepany: while the 2015 TEFAF re­port val­ues the art mar­ket at US$63.76 bil­lion, he es­ti­mates the po­ten­tial value to be some­where nearer $1 tril­lion. “What we want to do is to pro­vide the right in­for­ma­tion to stim­u­late that mar­ket, to help push mar­gins lower and to give banks, hedge funds and in­vestors ac­cess to a mar­ket [that is] in a sim­i­lar po­si­tion as the real es­tate mar­ket was in the 1980 and 90s."

While data col­lect­ing has been at the heart of Christie's busi­ness strat­egy since Day One, it is only in the last five to six years has the 250-year-old started do­ing large scale data-min­ing, ac­cord­ing to Re­becca Wei, Christie's Asia Pres­i­dent. She re­mains cau­tiously op­ti­mistic about the role of big data in the auc­tion mar­ket, “big data mat­ters when you have to de­cide whether to move on­line more ag­gres­sively, whether we need to close down part of the com­pany or where we should en­ter a spe­cific cat­e­gory, but when it comes to our day-to-day op­er­a­tions, small data is still much more im­por­tant.” Af­ter all, the auc­tion busi­ness is still con­sid­ered a niche mar­ket, where one client’s pur­chases could rep­re­sent 10 per cent of an auc­tion’s an­nual sales. Data of in­di­vid­ual clients are still more im­por­tant. “We aren’t in the FMCG in­dus­try. We are sell­ing a US$20 mil­lion di­a­mond, so we need to know which client has the funds.”

Wei’s cau­tious op­ti­mism also hints at where big data will best ef­fect change in the mar­ket. Com­pared to the top-end of the mar­ket, there is less pub­lic in­for­ma­tion about the mid­dle to lower mar­ket, but it is much larger, vol­ume-wise, and Doubrovbine points out that they’re see­ing lot of re­peat­able and con­sis­tent pat­terns for art that sells un­der US$10,000. Quant analy­ses also ap­pear par­tic­u­larly ben­e­fi­cial to the con­tem­po­rary mar­ket, where the risk is largest, ac­cord­ing to Kraussl.


Some have sug­gested that blockchain tech­nolo­gies might be a so­lu­tion to the in­dus­try’s opac­ity – it could

pro­vide se­cu­rity with­out com­pro­mis­ing on the pri­vacy of in­di­vid­ual clients. “Blockchain tech­nol­ogy has a lot to of­fer to any in­dus­try that deals with prove­nance, sup­ply chain, chain of own­er­ship, au­then­tic­i­ties,” says Je­han Chu, art ad­vi­sor turned in­vestor of blockchain com­pa­nies.

“If art is on the blockchain, it wouldn’t just be prove­nance we’re talk­ing about, we can also add user ex­pe­ri­ence, de­riv­a­tive prod­ucts such as in­surance,’ notes Chu. While a sophis­cated sys­tem is re­quired to ac­cu­rately marry ac­tual ob­jects, be it a paint­ing, a sculp­ture, an in­stal­la­tion, with the dig­i­tal iden­tity on the blockchain. Chu says the big­gest chal­lenge is adop­tion. “The tech­nol­ogy it­self is very sound, but how do you get ev­ery­one from stu­dent artist who are do­ing their own art show, to ma­jor sec­ondary mar­ket play­ers to get onto it?” Eleesa Da­di­ani of the Uk-based Da­dini Fine Art might be a trail blazer in this re­gard. On June 29, she an­nounced that buy­ers will be able to use any of the six lead­ing cry­tocur­ren­cies on the mar­ket, in­clu­idng Bit­coin and Ethereum, to buy art from Cork Street gallery.


The rise of AI is of­ten linked to the death of the hu­man ad­vi­sor. Those we talked all em­pha­sised that a big data-driven world would bet­ter, not elim­i­nate the role of the ad­vi­sor. “We can an­tic­i­pate…a shift in the role of ad­vi­sors. Col­lect­ing and con­sum­ing data will be­come a larger part of their rou­tine, although their strengths will re­main their knowl­edge of the work and the artist, con­nec­tions, and re­la­tion­ships,” notes Hop­per.

And in­deed, while hard num­bers might ap­pear more ob­jec­tive next to their live-in the flesh hu­man ad­vi­sors, big data can eas­ily be ma­nip­u­lated. “We all know that it is pos­si­ble to gen­er­ate any wished­for out­come,” quips Kraussl. “TEFAF and Art Basel should set up a web­site, in­di­cat­ing the gal­leries, and peo­ple they ap­proached. Oth­er­wise these stud­ies could be mis­lead­ing.”

An over­re­liance on big data and al­go­rithms might also lead to a ‘nar­row­ing’ of the mar­ket. “[Pur­chases can be in­flu­enced by] no­table press men­tions, crit­i­cal ex­po­sure and a cer­tain vol­ume of sec­ondary sales. This dis­tor­tion will con­sol­i­date es­tab­lished artists at the ex­pense of emerg­ing artists and favour gal­leries that rep­re­sent es­tab­lished artists, not least be­cause of the grow­ing cadre of buy­ers who see art as an in­vest­ment,” Hooper ex­plains. Hu­man ad­vi­sors might serve to cor­rect this, as Kraussl says, “the good ad­vi­sors will sur­vive, those who know what the art is re­ally about, in­stead of those only know how to spot a bar­gain­ing deal.”


For all its power, big data comes with its own set of lim­its. In an in­dus­try where col­lec­tors are of­ten driven by very per­sonal de­sires, it is im­pos­si­ble to pre­dict what the next Jean-michel Basquiat, Claude Monet or Jack Pollack will sell for. Few would have pre­dicted that a 1982 paint­ing by Basquiat would soar way above its US$60 mil­lion es­ti­mate to reach a record-break­ing US$110.5 mil­lion at a Sotheby’s sale, nor that Zao Wouki’s 29.09.64 would punch three times its US$6 mil­lion es­ti­mate to score Us$19.7mil­lion at Christie’s.

“It’s emo­tional, it’s im­pul­sive, and you never know how two peo­ple will be bid­ding against each other. More than 10 per cent of it [art buy­ing be­hav­ior] is ir­ra­tional, be­cause love is. In the case of Basquiat and Zao Wou-ki, a sin­gle jump in price al­tered the whole curve,” notes Wei.


– Ro­man Kraussl, fi­nance pro­fes­sor

Zaowou-ki’s 29.09.64went un­der­the ham­mer­for Us$19.7mil­lion atchristie’s Hongkong, morethan three­time­sits Us$6mil­lion es­ti­mate.

Newspapers in English

Newspapers from China

© PressReader. All rights reserved.