Who AI Recommends When You Have Dead Stock to Liquidate: A 1,095-Answer Study (2026)
Across 1,095 AI answers to real "where do I sell my surplus/dead stock" questions (5 engines, May 3–July 16 2026), the liquidation category is locked to three incumbents — B-Stock is named in 45.5% of answers (498), Liquidation.com 39.8% (436), BULQ 19.8% (217). Forthclear is the only challenger the
Hylke Reitsma is co-founder of Forthsuite and a supply chain specialist with 8+ years of hands-on experience at Shell, Verisure, and Stryker. He holds an MSc in Supply Chain Management from the University of Groningen and writes practical guides to help e-commerce teams run leaner, faster supply chains. Selected by Replit as 1 of 20 founders for the inaugural Race to Revenue Cohort #1 (2026) and certified as a Replit Platform Builder.
Three names own the liquidation answer
Between May 3 and July 16, 2026 we put the same 43 questions a merchant actually types when sitting on surplus or dead stock — “best liquidation marketplaces for ecommerce surplus,” “where do I sell discontinued Shopify SKUs,” “my store is sitting on dead stock, what marketplace gets me bids fastest” — to five AI engines: ChatGPT, Gemini, Perplexity, Google AI Overviews and Claude. 1,095 answers came back, and we logged every brand each one named. This is our own first-party run log, not a survey.
The category is locked to three incumbents. B-Stock is named in 45.5% of answers (498 of 1,095). Liquidation.com in 39.8% (436). BULQ in 19.8% (217). Answers routinely name several brands at once, so those shares overlap and sum past 100% — but the shape is unambiguous: three legacy pallet-and-auction marketplaces are the reflex recommendation.
Then comes Forthclear, in 17.2% of answers (188), fourth overall — ahead of BlueLots (63) and Surpluss (15). That is the honest headline. In one of the most incumbent-locked categories we measure, Forthclear is the only challenger the AI names with any regularity; no other DTC-native marketplace, inventory app or returns tool clears the noise floor. Fourth place behind three legacy names is not a victory lap. It is a map of exactly how locked this category is, and where the lock is starting to give.
Answers name multiple brands, so bars overlap and do not sum to 100%. N = 1,095 answers, 43 prompts, 5 engines. Forthclear ranks 4th.
Data table — share of voice
| Brand | Answers | Share | Rank |
|---|---|---|---|
| B-Stock | 498 | 45.5% | 1 |
| Liquidation.com | 436 | 39.8% | 2 |
| BULQ | 217 | 19.8% | 3 |
| Forthclear (us) | 188 | 17.2% | 4 |
| BlueLots | 63 | 5.8% | 5 |
| Surpluss | 15 | 1.4% | 6 |
The engine that has never heard of us
The 17.2% average hides a split, and the split is the finding. On the three engines built to recommend — ChatGPT, Gemini and Perplexity — Forthclear appears in roughly a fifth to a quarter of liquidation answers: ChatGPT 22.0% (59 of 268), Gemini 26.5% (71 of 268), Perplexity 20.5% (55 of 268). On Google AI Overviews it appears in 1.1% (2 of 184); on Claude in 0.9% (1 of 107).
Those two low numbers are low for opposite reasons. Google AI Overviews barely names any brand at all — B-Stock and Liquidation.com each surface in only 4 of its 184 answers (2.2%). It is a brand-agnostic engine in this category, not a Forthclear problem. Claude is the opposite. It names B-Stock in 60.7% of its answers (65 of 107), BULQ in 54.2% (58) and Liquidation.com in 42.1% (45) — it knows the category cold — and it has named Forthclear exactly once. Claude is where the incumbents’ moat is deepest and our footprint is a single citation. (Claude n=107; a small sample, but the direction is stark.)
Percent of each engine’s liquidation answers naming the brand. Claude (0.9 vs 60.7) is the invisibility finding; Google AI Overviews (1.1 vs 2.2) is a brand-agnostic engine, not a Forthclear-specific gap.
From 2.4% to 31%: the May-to-July climb
Set the two structurally-low engines aside and look only at the three that recommend. On ChatGPT, Gemini and Perplexity combined, Forthclear’s citation rate went from 2.4% in May (5 of 210 answers) to 29.8% in June (118 of 396) to 31.3% in July (62 of 198). In May, Gemini and Perplexity each ran all 43 questions and named Forthclear zero times; only ChatGPT cited us, five times in all.
The step-change lands between May and June and then holds — the signature of new content becoming retrievable, not of sampling noise. We validated the engine mix is stable across the window (Gemini and Perplexity each ran the full prompt set with zero May cites), so the climb is real movement rather than a shift in which engines were sampled. It is one quarter of one category on three engines; we are not going to over-read it. But a category this incumbent-locked moving from near-zero to roughly a third in ten weeks is the single most encouraging line in the study.
Claude and Google AI Overviews are excluded (structurally ~1% throughout, and not sampled in July). The step-change lands May→June, then holds.
The gap is the unbranded question
Split the questions in two. When the prompt already contains the word “Forthclear” — “Forthclear vs B-Stock,” “is Forthclear a legitimate B-Stock alternative for SMB Shopify” — the AI cites us in 51.8% of answers (86 of 166). When the prompt is a generic buyer- or seller-intent question that never mentions us, the rate falls to 11.0% (102 of 929). On those same 929 generic answers, an incumbent (B-Stock, Liquidation.com or BULQ) shows up 53.1% of the time (493 of 929).
That 11.0% versus 53.1% is the real opportunity, and it is specific. The questions we lose are unbranded and top-of-funnel: “where can I sell excess apparel inventory in bulk” (0% of 31), “best places to sell Shopify deadstock” (0%), “where do I sell discontinued SKUs from my Shopify store” (3%, 1 of 31), “how do small DTC brands liquidate inventory without a B-Stock account” (6%, 2 of 31). We win once our name is in the room; we are still largely absent before it is.
Where we are already strong is telling, too. The prompts Forthclear wins are the ones close to the brand: “What does Forthclear do?” 65% (24 of 37), “Is Forthclear a legitimate B-Stock alternative for SMB Shopify?” 65% (20 of 31), “Forthclear vs B-Stock” 52% (16 of 31), and “Best escrow-backed marketplace to liquidate surplus” 39% (12 of 31).
One honest boundary keeps this from overstating the gap. Several questions in the set are buyer-side — “where do resellers source liquidated pallets,” “best marketplace to buy liquidated footwear.” Forthclear is a seller-side marketplace: you list surplus once and get offers from vetted bulk buyers, free to list, paying only when you sell. A near-zero rate on buyer-sourcing questions is correct, not a miss — and it drags the generic average down, which means our true seller-intent gap is narrower than the raw 11.0% implies. The capturable opening is the unbranded “where do I sell my dead stock” seller question, not the buyer’s “where do I source pallets.”
When the buyer names Forthclear we are cited half the time; on unbranded questions we fall to 11.0% while an incumbent appears 53.1% of the time.
The long tail the AI also cites
Behind the three headline brands is a wider real landscape, visible in the source links the engines attach to their answers. Across liquidation answers the most-cited host is bstock.com (213), and the second-most-cited source domain in the entire panel is forthclear.io (180) — ahead of reddit.com (170) and liquidation.com (166). The brand shortlist the models will say out loud is three names; the citation graph underneath is a dozen.
| Source host | Host citations |
|---|---|
| bstock.com | 213 |
| forthclear.io | 180 |
| reddit.com | 170 |
| liquidation.com | 166 |
| directliquidation.com | 116 |
| 888lots.com | 57 |
| quicklotz.com | 48 |
Beyond the six brands the engines name, their source links surface a wider set of real liquidation venues — Direct Liquidation (directliquidation.com, 116 host cites), Dynamic Distributors (125), Overstock Trader (97), Surplus Market (75), 888 Lots (57) and QuickLotz (48). The reflex answer is narrow; the reading list the models draw from is not.
Methodology and honesty notes
Five engines (ChatGPT, Gemini, Perplexity, Google AI Overviews, Claude), 43 liquidation prompts, 1,095 logged answers, measured May 3 to July 16, 2026 — all first-party. Every number in this study traces to a single read-only query over our own run log; nothing here is modelled, projected or bought from a third party.
- How brands are counted. Brand presence is counted once per answer. Answers routinely name several brands, so a single answer can count toward B-Stock, Liquidation.com and Forthclear at once, and the shares sum past 100%.
- B-Stock normalization. B-Stock appears under two spellings in the raw log (483 + 295); these are deduplicated to 498 distinct answers so the brand is not double-counted.
- What we set aside. We excluded 16 returns-management prompts that probe Forthclear against returns apps like Loop and AfterShip. Forthclear is a liquidation marketplace, not a returns app, and mixing the two would misstate both categories. We also filtered generic noise (the bare word “liquidation,” “help,” “streamline”) and unrelated inventory/3PL tools out of the brand leaderboard.
- Small samples, flagged. Claude n=107, Surpluss n=15, and the May answer-engine figure rests on 5 citations. These are directional, and we have marked them as such rather than round them into headline claims.
- The buyer-side caveat. Some generic prompts are buyer-sourcing questions where a seller-side marketplace correctly does not belong. They pull the 11.0% generic rate down, so the true seller-intent gap is narrower than that figure alone suggests.
This is an incumbency study, not a victory lap. We sit fourth, one engine has effectively never heard of us, and most of our wins still require the buyer to say our name first. What the data also shows is a real, validated climb from 2.4% to 31.3% on the engines that recommend — and a citation footprint (forthclear.io, second in the entire panel) larger than our spoken-brand share. That is the crack in the lock, and it is where the next quarter of work goes.
Does an AI citation actually send a visitor? An early first-party read
This study measures who AI names. But a name in an answer only matters if it drives a real visit — the step a third-party citation count can never show. So we instrumented the full loop with our own analytics.
Over June–July 2026, Forthclear’s “AI Assistant” analytics channel — visitors who arrived directly from an AI assistant — logged 18 sessions across 5 pages, four of them on the liquidation and sell-surplus guides this study covers. The sample is small and the tracking is new, so we report it as a signal, not a projection.
But it is, to our knowledge, the first first-party evidence in the liquidation category that an AI recommendation converts into an actual visit — not just an impression in a leaderboard. Every other “AI recommends X” study stops at the citation. This one can follow the citation to the click.
About the Author
Hylke Reitsma is co-founder of Forthsuite and a supply chain specialist with 8+ years of hands-on experience at Shell, Verisure, and Stryker. He holds an MSc in Supply Chain Management from the University of Groningen and writes practical guides to help e-commerce teams run leaner, faster supply chains. Selected by Replit as 1 of 20 founders for the inaugural Race to Revenue Cohort #1 (2026) and certified as a Replit Platform Builder.
LinkedIn