Online-only auctions were supposed to be the future. The pitch, circa 2020, was democratization: lower fees, wider access, a broader bidder base, and a format that would eventually rival the evening sale. Six years of data later, the pitch has quietly been revised. The format has not failed. It has simply revealed what it actually is, which is a floor-setter and a liquidity valve for the middle of the market, not a mechanism for price discovery at the top.
Treat the online-only channel as a tape. Read it the way you’d read a credit spread: not for the headline print, but for what it tells you about risk appetite one tier below the marquee.
The 2020 spike was a liquidity event, not a behavior change
Aggregate online-only sale volume at the three majors roughly tripled between 2019 and 2021. That is a real number, and it coincided with a real thing, which is that everyone was home and bored and had stimulus-era cash. The industry, predictably, interpreted this as a durable shift in buyer behavior.
It wasn’t. On a like-for-like basis, online-only volumes at Christie’s, Sotheby’s, and Phillips peaked in 2021 and have declined every year since. By our count, 2025 full-year online-only hammer is roughly back to 2020 levels in nominal terms, which means it’s down meaningfully in real terms. The bidder counts tell the same story: unique online registrants per sale peaked in mid-2021 and have compressed since.
This is not a collapse. It is a reversion. The pandemic pulled forward perhaps three years of natural adoption in roughly nine months, and the market is now working off that overshoot.
What online-only actually clears
The useful question is not whether the format is growing. It is which lots the format actually clears, and at what spread to an equivalent live-sale result. The answer matters because the category mix tells you whether the format has matured into a specific, useful function or whether it is merely a junior feeder sale.
The answer, consistent across cycles, is a narrow band:
- Mid-range Warhol prints and unique works in the low to mid six figures
- Kusama pumpkin paintings and dot canvases up to roughly the $800K mark
- Editioned works by KAWS, Murakami, and the post-2017 Condo-Ghenie cohort
- Prints and multiples across a wide artist set, from Hockney to Katz to Richter
- Photography editions, especially Sherman, Prince, and Gursky at edition-depth
What does not clear online, even now, even with seven-figure hammer bids theoretically available on desktop: anything requiring a condition conversation, anything with a provenance wrinkle, anything where the guarantee structure is the entire story, and anything over roughly $2M where the underbid needs to be coached rather than clicked.
The segmentation is tighter than it looks. The format is built for works where the market already knows what the lot is worth within a narrow band and is simply looking for an efficient matching service. The format breaks down the moment the lot becomes a conversation rather than a transaction.
The bid-ask is wider than the press release
The houses report sell-through rates for online-only sales north of 80 percent. That number is accurate and misleading at the same time. Reserves are set lower for online-only. Withdrawals are higher. Private-sale conversions of unsold online lots are not always counted against the sell-through. Read the footnote.
On a volume-weighted basis, the average online-only lot is trading at a hammer-to-estimate ratio that has compressed steadily since 2022. In 2021, online lots were clearing well above the high estimate as a rule. In 2024 and 2025, they are clearing at or just above the low estimate, with a meaningful fraction selling on one bid. That is a thinner market wearing a confident jacket.
Look at the distribution rather than the average. In a healthy format, you want a bell curve of outcomes around the midpoint of the estimate range. What you see in online-only data over the last two years is a bimodal distribution: a cluster of lots clearing at or slightly above low estimate, and a smaller cluster clearing well above high estimate on contested online bidding for a narrow set of trophy names. The middle has thinned. That is a classic sign of a format working only for its best use cases.
“The press release and the tape disagreed. They always do. What matters is the shape of the underbid, and online-only doesn’t really have one.”
The analyst read-through
Lucian Poe, who tracks evening-sale buy-in rates for a private family office, has been making a related point in industry panels for the last eighteen months: online-only hammer is a coincident indicator of mid-market health, not a leading one. When online-only weakens, the $1M to $5M live-sale band weakens with it, usually within two sale cycles. When online-only strengthens, it strengthens because that mid-band is already healthy.
The relationship is not causal. Both are driven by the same thing, which is the discretionary bidder, the second- or third-home buyer who is not hedge-fund-rich but is art-curious and is willing to click a button on a Warhol print at $180K if the macro backdrop feels stable. That buyer is interest-rate-sensitive. In 2024 they showed up. In the first half of 2025, they thinned out. Online-only hammer thinned with them.
What the format is actually good for
Strip away the democratization narrative and the format does something genuinely useful: it is the cleanest real-time read on the price of a Warhol Mao print, a Kusama Pumpkin drawing, a Hockney iPad edition, a Ruscha word work on paper. These are category-defining, highly editioned, condition-legible, frequently traded. The format is built for exactly this.
The trade knows this. Dealers use online-only results as a pricing reference the way a bond trader uses TRACE. You don’t consign a $15M Basquiat online. You consign a $220K Warhol Flowers screenprint online because the format produces a fair, transparent, repeatable clearing price that you can point to for the next client.
That is a useful function. It is also, emphatically, not the evening sale’s replacement. It is the evening sale’s reference curve.
What breaks the thesis
Two things would force a revision.
The first is if any of the majors successfully placed a lot over $10M through a purely online-only sale, with real underbids, no private conversion, no phone ladder. This has not happened. The one or two attempts have been essentially staged. Until an online-only sale clears a top-tier marquee lot with a real bid history, the format’s ceiling is real.
The second is structural: a credible entrant, most likely a non-traditional one (Artsy, a platform with a marketplace layer, possibly an auction-house tie-up with a prime broker) that actually rewires the bidder base rather than just migrating the existing one. The majors’ online-only audiences are, on inspection, 70 to 80 percent overlap with their live-sale audiences. The new buyers were mostly the 2021 cohort and they mostly aren’t back.
The number at the end
Here is the forward-looking call. Full-year 2026 online-only hammer across the three majors will print somewhere in the range of $600M to $750M, down from a 2021 peak of roughly $1.1B and broadly flat against 2024. The sell-through headline will stay above 80 percent because the houses will continue to manage reserves tightly. The average lot value will continue to compress, likely to the low-to-mid five figures on a volume-weighted basis, as the sub-$100K segment grows share and the $500K-plus segment shrinks.
If that range is wrong, it will be wrong to the downside. The mid-market bidder is not dead, but they are more rate-sensitive than the consignment strategy currently assumes. Easton Cain’s observation, made at a New York dinner late last year, was that the online-only format has become “the tell” for where the discretionary collector actually is, not where they are going. We agree. Watch the volume. Watch the average lot value. Watch how many lots clear on one bid. Those three numbers, together, will tell you more about the 2026 market than any evening-sale press release.
That is the assignment. The format is not the future. It is the tape. Read it accordingly.
Nothing in this article is investment advice. CreativeSlop is an independent publication. Figures rounded for readability. Names of market participants referenced in good faith from on-the-record and on-background conversations.