Since the art world’s widespread physical shutdown triggered a flood of online viewing rooms, much of the industry’s attention has been directed toward typical sales reports and heartening progress on price transparency. At this point, we can safely say two things: that business is still being done in the depths of the crisis; and that the average observer has never before had more visibility into the availability of artworks and their costs. These are inherently good things.

The trade’s step into e-commerce also means that another valuable commodity is being exchanged at an unprecedented rate, even when no artworks are being sold: user data. In contrast to decades worth of data-gathering in the auction sector (let alone the larger sphere of online retail), private art dealers have lagged far behind on this front. 

But the great data gold rush of 2020 might just lead galleries and fairs into a long-overdue era of quantitative analysis—that is, assuming the transition is handled conscientiously and strategically.

To slash a clear path through this thorny thicket of technological and legal questions, our trusty art business editor has assembled a handy Q&A in plain English (with a dash of good humor) so that even total beginners can get a sense of how to navigate this terrain. 

 

OK egghead, what information can galleries and fairs collect about me right now? 

That depends partly on the law in your country of residence, and partly on you. 

A growing body of state, national, and international laws restricts companies’ abilities to profile users based on their browsing activity without their consent. The most prominent of these is the European Union’s General Data Protection Regulation (GDPR). While it is technically an EU law for EU-based companies, any online platform in any country can run afoul of GDPR if that platform violates the rights of an EU citizen—which is why every website in existence had to be retrofit for compliance before the regulation went into force in May 2018. 

According to Kenneth Mullen, a partner at the international law firm Withersworldwide, the core purpose of GDPR is to ensure that online retailers “only collect the amount of data needed to fulfill their lawful purposes” via transparent collection practices. Moreover, companies are legally allowed to preserve information about a user only as long as it’s needed. GDPR also prevents retailers from sharing user data with others unless there are “legal safeguards in place to ensure it’s only used for purposes that the customer reasonably expects.”

But that still leaves more wiggle room than you might think.

Translation, please?

Here’s the good news: Thanks to GDPR, the California Consumer Privacy Act, and similar laws elsewhere, every website (including every online viewing room) offers a privacy policy that outlines each type of data being gathered from you, how it’s being gathered, which third parties it may be shared with, and to what ends—as well as (under certain conditions) the process by which you can procure a copy of all your personal data, restrict its being shared with others, and/or demand its wholesale deletion.

Here’s the less-good news: As a representative example, David Zwirner’s privacy policy swells to over 2,000 words. For reference, that’s about the max I can submit to my very patient editors without them looking at me like I just dumped a bubbling cauldron of salsa on their favorite rug. [Editor’s note: this one is just over 2,300 words, but it gets a pass.]

Which means basically none of us ever reads these things. As a result, we tend to hand websites permission to harvest additional info as eagerly as we would hand our celebrity crushes our phone numbers. 

 

Wait, I’m voluntarily giving out extra data to art dealers? How?

You know that pop-up window asking if you consent to the use of “cookies” whenever you visit a website for the first time? Cookies are files downloaded to your device that hold a tiny amount of info about you and/or the website you’re visiting. The second you blindly click “OK” or “Approve,” you’re agreeing to install “non-essential” cookies that deliver data about your browsing activity to the website owner. 

 

Oh… well, what kind of data do these cookies deliver?

They do things like track users’ movements from page to page on their websites, log which items a user clicks on, and record the amount of time spent on each page. All of this data offers potentially valuable insights into consumer behavior. But the key is that none of it is actually linked to consumers’ identities. 

“The purpose of the data is to make correlations,” explains Penny Gillespie, vice president of research at the multinational business-intelligence firm Gartner. Analyzing data “can lead to personalized experiences without knowing the customer, which ultimately leads to greater conversion and even cross-sell and upsell opportunities.” You just have to achieve sufficient scale with the anonymized data first.

Uh, do I have to go to business school, or can you clarify that with an example? 

Please allow me to save you two years and about $100,000.

Imagine your website sells women’s clothing, and you’ve had 200,000 potential shoppers search your site with the phrase “black dress.” That almost infinitely variable term may not mean much alone. But based on which search results users actually click on, it becomes clear that what most people intend to find is a black cocktail dress. Based on that data, your engineers can tweak the search algorithm to prioritize black cocktail dresses for any new search for “black dress” going forward. 

This is where it gets interesting, as Gillespie shows by continuing the example to its endpoint: 

“Over time, I also learn that the intent of the customer is not just to buy a black dress (or even cocktail dress), but rather to buy an ensemble for a wedding or black-tie event. Then, when the customer searches for ‘black dress,’ I show a black cocktail dress in the context of an event (wedding, gala, etc.) and have an opportunity to sell the customer an entire outfit (dress, shoes, bag, jewelry, undies, makeup, etc.).”

 

So galleries are doing this kind of anonymized trend analysis now? 

They’re starting to, often with the help of art fairs or other partners. 

According to a spokesperson, Art Basel will provide exhibitors in the June iteration of its online viewing rooms with “basic analytics on the number of page views per room, as well as per artwork, while exploring further possibilities.” Dealers in the inaugural Frieze online viewing room earlier this month received aggregated information on which works in their digital booths received the most views, and for how long. 

Similarly, a David Zwirner spokesperson confirms that the gallery has been relaying to the dealers making use of its Platform initiative some overall visitor metrics, including total page views, average time spent on site, and the number of users hailing from each city or country represented. Sotheby’s will also share anonymized data, such as the number of visits and page views, to members of the auction house’s recently launched Gallery Network.

But again, the central point here is that all of this information is nameless and faceless. It only allows dealers to make inferences about which artists and works generally make the strongest first impression (by provoking a click), best sustain viewers’ interest (by extending time spent on each page), and ultimately incite some kind of action (such as an email inquiry or, on Sotheby’s Gallery Network, placing a work under $150,000 in your virtual shopping cart). 

 

Are dealers doing anything more targeted and personalized?

Officially, the answer appears to be “not really,” but the ingredients are in place to change that if and when policies and priorities allow.

Zwirner’s Platform and the Sotheby’s Gallery Network don’t require users to submit any identifying info at all to view the works on offer. But other initiatives do, and that would be the starting point. 

Art Basel’s online viewing rooms, for example, are only accessible after users log into their Art Basel accounts, which means putting their first and last names and a valid email address on file. Frieze requires a name, email address, and telephone number (though the number is only used to verify registration); and unlike Platform, Zwirner’s own online exhibitions can only be unlocked by providing an email address, which is relatively standard practice among galleries of all sizes.

This identifying information could empower dealers to begin building individual profiles of users’ tastes based on their browsing activity. But everyone I spoke to for this story emphasized that, in adherence to GDPR and other privacy laws, they are strictly gathering the aggregated (meaning, anonymized) data discussed earlier. 

Do galleries ever get data linked to a specific customer? 

Pretty much only when users send direct email inquiries about specific works. In Art Basel’s online viewing rooms, users can provide additional data points about themselves (most notably, their VIP status with the fair), but only if they wish. Otherwise, the dealer they reach out to only receives their name and email.

It’s a similar story when visitors to the Frieze online viewing room sign a gallery’s virtual guest book. At that point, their name and registered email are supplied to the dealer in question (who must then only use them in accordance with online privacy regulations). And of course, individual galleries get the same information from direct inquiries on their own in-house platforms.

“Everyone wants to know who the customer is because the experience can be even more personalized,” Gillespie says. But for now, the art market doesn’t appear to be taking the next step by building profiles of specific users, which is the Excalibur of smart online retail.

 

What would user-specific personalization look like for the online art market? 

We can get some clues by leveling up the “black dress” example from earlier. Instead of simply intuiting that an anonymous user is searching for wedding attire and responding with semi-tailored options, a specific customer’s browsing (and buying) history on the same site could be mined to serve up wedding fits from their favorite brands, styles, price points, and more—all in real time, so that the site is further personalized with each new click.

The next logical upgrade is to implement personalization engines that would enable this type of customization for private art dealers. Rather than give every collector the same grid of available works in the same order, or ask them to filter lists of artists themselves, algorithms could automatically sequence online viewing rooms to prioritize specific artists (or types of artists), media, or price ranges according to what each particular user has preferred in the past. 

That feature might not seem especially useful in, say, a single small gallery’s viewing room featuring 10 or so works by one artist. But its utility increases dramatically for a mega-gallery showcasing dozens of works by multiple artists at once—and exponentially for an online art fair offering thousands of works by hundreds of artists in dozens of galleries. 

 

Technologically speaking, how far away are we from that future? 

General retailers are basically there now. Gillespie has already reviewed case studies for various personalization engines, and says the results are “typically impressive” and generate meaningful return on investment “within months.” 

I just can’t tell you whether any dealers, art fairs, or auction houses are seriously exploring the possibilities. Either fittingly or ironically for a story about data permissions, no one I spoke to for this story was particularly keen on getting into specifics about the future. The best I can do is relay that they’re all using both quantitative analysis and good old-fashioned qualitative feedback from customers and exhibitor-partners to enhance their future offerings. 

Is data collection and analysis an arena where the biggest, richest players have a huge advantage over the little guys? 

Yes and no. According to Gillespie, a stark divide between haves and have-nots used to be the default in the wider retail realm, but now “more sophisticated technology is being brought to the masses.” 

Case in point: many retailers long ago opted out of using their own proprietary software to analyze the aggregated data we’ve already discussed. Art Basel uses Google Analytics to dissect the traffic in its online viewing rooms, and given Google’s prominence in the field, I’ll bet you a pallet of exorbitantly priced black-market Clorox wipes that our Swiss friends aren’t the only ones in the private market doing so. Why build something (likely) worse from scratch if you can plug and play a better toolkit from a trusted expert for a reasonable fee?

The same basic calculus appears in personalization engines: individual companies can either try to develop their own tools and unpack the data themselves, or they can pay for third-party technology, expertise, and analysis. Ditto for the types of robust cybersecurity measures needed to safeguard sensitive user data, especially for online viewing rooms linked to an actual payment and fulfillment apparatus. In all cases, just be advised that the results and prices of these services will vary.

 

That’s a lot to digest. What’s the big takeaway? 

Whether we’re talking about the anonymized trend analysis available today, or the real-time personalized experience achievable in the future, online dealers need to collect large amounts of data before it becomes especially useful. Right now, that’s much more achievable for a Frieze or a David Zwirner than it is for a single-location basement gallery hoping to break through online. 

Which means that thriving on the frontier of user data still comes down to fundamental questions about who your audience is, how engaged they are with your digital platform, and how long any dealer or fair can stick around to learn from them. Those aren’t new lessons; online viewing rooms and their associated data discoveries just make them newly relevant.