← back to rankingETSY · Etsy, Inc.
Specialty Retail · mkt cap $6.6B · calls: Q1 FY2026 vs Q4 FY2025
39.0 conviction · conf-adj 39
conf 4/10 conflict
enthusiasm:24.0 · trend:8 · quantifies:0 · impact:0 · under_radar:0 · credibility:5 · business_impact:8 · disruption:-6 · commitment:0 · confirmation:0
Enthusiasm latest 8 / prev 7 (rising)
Etsy's AI story is substantive matching/discovery (ML/LLMs, AI buyer profiles, personalized app and owned channels), monetization (ML-driven Etsy Ads), seller productivity (listing/shop AI, two shipped agents), and agentic distribution (OpenAI/Microsoft/Google; tiny but fast-growing traffic with high intent and on-site flow-through). Enthusiasm is rising as Q1 adds shipped product (profiles, agents, conversational tests) beyond Q4's partnership narrative. Management quantifies channel scale (15x agentic traffic, sub-1% share) and adjacent engagement/GMS metrics but does not attribute dollars of revenue or margin solely to AI; credibility is moderate—real deployment and hybrid-model discipline, but frequency has not inflected and AI outcomes remain early-stage.
GROUNDED NEXT-FY IMPACT vs CONSENSUS
Grounded on actual base — revenue $2.9B · net income $0.2B · net margin 5.7% · diluted EPS 1.39
These are next-fiscal-year annual uplift estimates, not next-quarter numbers.
Aggregate next-FY est. rev uplift: % · next-FY EPS uplift: % · vs analysts: unclear · priced in: high (model's call-read: medium; verdict above is the hard-data one used for ranking) · confidence: 4/10
| Claim | Figure | Arithmetic | Next-FY Rev % | Next-FY EPS % |
|---|
Agentic traffic 15x YoY (Q4), still <1% of traffic
engagement · soft | ~15x YoY; <1% of total traffic, 'a small part of GMS' | Bounded but not cleanly sizable as net-new AI revenue. If agentic traffic share converted at the blended take rate and were fully incremental, post-15x it is <1% of $2,883.5M (<$28.8M gross); a midpoint 0.75% share ×14/15 incremental = 0.70% of traffic → ~$20.2M (Y's arithmetic). BUT management calls it channel-shifted and 'a small part of GMS,' so most is not net-new and traffic share overstates revenue share; no traffic→revenue conversion or GMS$ disclosed. Net-new AI-attributable revenue is a fraction of the ceiling → soft. | | |
Agentic traffic share <1% of total traffic
engagement · soft | <1% / fraction of a percent | Descriptive level/ceiling, not a delta: caps agentic-attributable topline at <1% of revenue (<$28.8M gross). Overlaps the 15x claim (not additive). No incremental conversion $ disclosed. | | |
App GMS +11.2% YoY (Q1'26) vs +6.6% prior qtr; app ~47% of GMS
engagement · soft | +11.2% vs +6.6%; app=47% of GMS | Arithmetic is HARD on disclosed rates (Y: 0.47×(1/1.066−1/1.112)=0.386% of GMS → ~$11.1M) BUT this is GMS not revenue and the 4.6pp acceleration is multi-causal (mix, pricing, promotions, ML); no AI-attributable share disclosed, so the AI portion cannot be isolated → soft. | | |
App share of GMS ~47%, +240bps YoY
engagement · soft | 47%, +240bps | Structural mix shift. Y sizes 2.40%×40% LTV=0.96% of rev (~$27.7M) but flags 'not 100% AI-causal'; LTV gap is cross-sectional, not an AI-driven delta, so AI attribution isn't isolatable → soft. | | |
App users 40% higher LTV than non-app
engagement · soft | +40% LTV | Cross-sectional benchmark (app vs non-app cohort), not an AI-driven delta; standalone $ would require invented incremental volume. Unsizable. | | |
Homepage clicks/visit +14% YoY (Q4'25)
engagement · soft | +14% | Upper-funnel engagement; no disclosed homepage share of visits/GMS or click→conversion→revenue elasticity. Cannot map to $ without inventing funnel share. | | |
Push & email clicks +25%, volumes disciplined (Q4'25)
engagement · soft | >+25% clicks | Owned-channel engagement lift from personalization; no owned-channel % of revenue or conversion delta disclosed. Volumes flat → no opex saving to size either. | | |
Owned-channel personalization coverage <1/4 -> >3/4
engagement · soft | from <25% to >75% personalized (~50pp) | Capability-coverage metric, not a financial output; drives the +25% click lift above (would double-count). No incremental $. | | |
Take rate 25.7%; marketplace take-rate expansion led by Etsy Ads ML relevance/pacing
other · soft | take rate 25.7% (level), AI-led expansion (bps undisclosed) | Revenue-relevant and on Etsy's own ad product, but 25.7% is the take-rate LEVEL; no prior-period take rate or AI-attributable bps of expansion disclosed, so incremental revenue can't be isolated. Take-rate gains already flow to reported revenue/consensus. | | |
Search/discovery early tests improve add-to-cart & conversion
engagement · soft | qualitative, no % | Explicitly 'early tests,' no % disclosed. Vague/unanchored. | | |
ChatGPT-originated orders skew higher value
revenue · soft | qualitative, no % | Directional only; no AOV delta or volume disclosed. Subset of the <1% agentic channel. | | |
Internal build velocity 'weeks not months' from AI tooling
productivity · soft | qualitative | No $ opex saving, FTE count, or capitalized-dev figure disclosed; management frames AI/ML spend as staying within existing margin outlook → no net bottom-line uplift to book. Unsizable. | | |
Assumptions: Next-FY (FY2026) sizing. All claims are adopter-side (AI improving Etsy's own marketplace, app, ads, marketing, tooling); no supplier-side AI revenue. Where arithmetic is possible it is HARD on disclosed engagement/mix rates (Y showed: agentic 0.75%×14/15→~$20.2M; app mix 2.4%×40% LTV→~$27.7M; app accel 0.47×(1/1.066−1/1.112)→~$11.1M, at incremental net margin 5.6522% = FY2025 162,982,000/2,883,501,000, tax 21% N/A as no $ saves) — BUT in every case the AI-attributable share is not isolated and the metric is GMS/engagement/traffic rather than disclosed incremental revenue, so no figure is booked as a clean AI uplift. The agentic channel is the only genuinely bounded item: <1% of traffic even post-15x and largely channel-shifted ('a small part of GMS'), capping AI topline well under 1% of the $2,883.5M base (<$28.8M gross, far less net-new). EPS basis: do NOT compute EPS% off depressed GAAP net income $162.98M (GAAP EPS $1.39); consensus runs on a much higher adjusted/forward base (FY2026 NI ~$415.6M, EPS $3.57), so a GAAP-denominated EPS% would be inflated and is moot with no anchored figure.
Top line: No cleanly AI-attributable incremental revenue is disclosed. The only bounded item — agentic/ChatGPT traffic — is explicitly <1% of total traffic even after 15x YoY growth and is largely channel-shifted, capping near-term AI topline well under 1% of the $2,883.5M base (<$28.8M gross, far less net-new). The real engagement story (app GMS accelerating to +11.2% YoY, app ~47% of GMS +240bps, homepage clicks/visit +14%, push/email clicks +25% on personalization rising <25%→>75%, Etsy Ads take rate 25.7% with ML-led expansion) is material to GMS/revenue but multi-causal, with no AI-attributable share disclosed, so it cannot be sized as a clean AI uplift. Y's HARD arithmetic sums to ~1.66% of revenue (~$47.9M) only by treating agentic traffic share as fully-incremental revenue and app mix/LTV as AI-causal — both attribution assumptions the disclosures do not support — so the consensus does not adopt that as the AI tailwind. Against a consensus FY2026 revenue of ~$2,778.5M (≈−3.7% YoY), any AI tailwind is immaterial and not separately identifiable.
Bottom line: No quantified cost saving or FTE reduction. 'Weeks not months' build velocity is qualitative, and management explicitly commits to keeping ML/agentic compute spend WITHIN the existing margin outlook — i.e. reinvested, not dropped to EPS. EPS uplift is therefore null, not zero-by-tiny-denominator: there is simply no anchored figure, and the depressed GAAP base ($162.98M NI / $1.39 EPS vs consensus $3.57 adjusted) would distort any % anyway.
[impact n/m (all claims soft/unanchored)] Consensus prices essentially NO revenue growth — 2026 revenue $2,778.5M is -3.7% vs 2025 $2,884.9M, recovering only to ~$2,893.0M by 2027 — while EPS jumps to $3.57 (2026) on margin/buyback, not AI revenue. The AI engagement initiatives (app GMS +11.2%, take-rate expansion) argue for upside to that flat-to-declining revenue line, and agentic shopping is essentially unpriced optionality. But every AI claim is an engagement/coverage/qualitative metric with zero isolated dollar attribution, so the gap-to-consensus cannot be quantified — the upside case is real directionally but unmeasurable from disclosure. Engagement gains already partly embedded in reported GMS; agentic (<1%) too small to move the number yet.
MODEL CONSENSUS (impact)
conflict
Sided with X's attribution discipline: real metrics, but no disclosed AI-attributable incremental $; kept Y's bounding arithmetic in basis as ceilings, not booked uplift.
Conflicts reconciled
- est_rev_uplift_pct: X=null vs Y=1.66 -> used null because Y's 1.66% treats agentic traffic share as fully-incremental revenue (mgmt: channel-shifted, 'small part of GMS') and app mix/LTV as AI-causal, but no AI-attributable share is disclosed
- est_eps_uplift_pct: X=null vs Y=0.63 -> used null; same unisolated attribution and no disclosed AI-incremental $; also avoids GAAP-denominated distortion
- agentic 15x rev/eps: X=null/soft vs Y=0.7/0.7 -> kept soft with a stated <1% (<$28.8M gross) ceiling because traffic→revenue conversion is undisclosed and the channel is largely not net-new
- app mix/accel: X=soft vs Y=0.96/0.39 HARD -> kept soft; arithmetic is sound but GMS-not-revenue and AI attribution not isolated
| Field | Opus 4.8 | GPT-5.5 |
|---|
| Rev uplift % | – | – |
| EPS uplift % | – | – |
| Priced in | medium | – |
| vs analysts | unclear | – |
| Confidence | 3 | – |
| Top line | No quantified AI-attributable revenue. The one bounded item — agentic/ChatGPT traffic — is explicitly <1% of total traffic even after 15x YoY growth and is largely channel-shifted, capping near-term AI topline well under 1% of the $2,883.5M base (<$28.8M gross, far less net-new). The real engagement story (app GMS accelerating to +11.2% YoY, app now ~47% of GMS +240bps, homepage clicks/visit +14%, push/email clicks +25% on personalization rising from <25% to >75% coverage, Etsy Ads take rate 25.7% with ML-led expansion) is plausibly the largest lever but is reported in GMS/engagement units with no isolated AI attribution, so it is not separately sizable and already flows into reported results. | – |
| Bottom line | No quantified cost saving or FTE reduction. 'Weeks not months' build velocity is qualitative, and management explicitly commits to keeping ML/agentic compute spend WITHIN the existing margin outlook — i.e. reinvested, not dropped to EPS. EPS uplift is therefore null, not zero-by-tiny-denominator: there is simply no anchored figure, and the depressed GAAP base ($162.98M NI / $1.39 EPS vs consensus $3.57 adjusted) would distort any % anyway. | – |
| Reasoning | Consensus prices essentially NO revenue growth — 2026 revenue $2,778.5M is -3.7% vs 2025 $2,884.9M, recovering only to ~$2,893.0M by 2027 — while EPS jumps to $3.57 (2026) on margin/buyback, not AI revenue. The AI engagement initiatives (app GMS +11.2%, take-rate expansion) argue for upside to that flat-to-declining revenue line, and agentic shopping is essentially unpriced optionality. But every AI claim is an engagement/coverage/qualitative metric with zero isolated dollar attribution, so the gap-to-consensus cannot be quantified — the upside case is real directionally but unmeasurable from disclosure. Engagement gains already partly embedded in reported GMS; agentic (<1%) too small to move the number yet. | – |
Rows highlighted where the two models disagreed.
QUANTIFICATIONS
Agentic traffic growth: about 15x year-over-year in Q4 (Q4 FY2025 vs prior year, topline)
“While still a very small part of our overall traffic in GMS, agentic traffic to Etsy in Q4 was about 15x what it was last year, underscoring just how rapidly this channel is emerging.”
Agentic traffic share: less than 1% of total traffic (also described as a fraction of a percent in Q1) (Q4 FY2025 / Q1 FY2026, topline)
“it's still very, very small. It's a fraction of a percent of our total traffic.”
App GMS growth (personalization/ML-centered app strategy): up 11.2% year-over-year in Q1 2026 vs up 6.6% last quarter (Q1 FY2026, topline)
“Mobile app GMS was up 11.2% year-over-year in the first quarter of 2026 versus up 6.6% last quarter.”
App share of GMS: about 47% of total GMS, expanding 240 basis points year-over-year (Q1 FY2026, topline)
“App GMS is continuing to significantly outpace non-app growth and now makes up about 47% of total GMS, expanding 240 basis points year-over-year.”
App user lifetime value premium: 40% higher LTV than non-app users (Q1 FY2026 (structural benchmark cited in Q&A), both)
“our app users have 40% higher LTV than non-app users.”
Homepage engagement (personalized app home): homepage clicks per visit increasing 14% year-over-year (Q4 FY2025, topline)
“homepage clicks per visit increasing 14% year-over-year”
Owned marketing engagement: push and e-mail clicks up more than 25% (Q4 FY2025, topline)
“push and e-mail clicks up more than 25% while message volumes stayed disciplined.”
Owned-channel personalization coverage: from less than 1/4 to over 3/4 of push notifications and e-mails personalized (by Q4 FY2025, topline)
“our push notifications and e-mails have gone from less than 1/4 of them being personalized to now over 3/4 of them being personalized.”
Etsy Ads / ML relevance (take rate): take rate 25.7%; Etsy marketplace take rate expansion led by Etsy Ads with machine learning-driven relevance and pacing (Q1 FY2026, both)
“Etsy marketplace take rate expansion was led by Etsy Ads where we continue to benefit from machine learning-driven improvements to relevance and seller budget pacing.”
Search/discovery product impact on conversion (qualitative only): improvements in add to cart rates and conversion (no % disclosed) (Q1 FY2026 early tests, topline)
“With early tests showing improvements in add to cart rates and conversion.”
ChatGPT-originated order value: skew higher value vs some mature acquisition channels (no % disclosed) (Q4 FY2025, topline)
“Orders originating from ChatGPT also tend to skew higher value compared to some of our more mature acquisition channels.”
Internal build velocity from agents/AI tooling: weeks, not months (Q1 FY2026, bottomline)
“they're allowing us to move faster, building and iterating in weeks, not months”
PAST (realized)
- Q4 FY2025: Agentic traffic to Etsy was about 15x year-over-year in Q4 while still a very small part of overall traffic and GMS.
- Q4 FY2025: Push and email personalization coverage increased from less than 1/4 to over 3/4 of messages, driven by machine learning models.
- Q4 FY2025: App homepage clicks per visit increased 14% year-over-year; app GMS growth accelerated to 6.6% in Q4.
- Q4 FY2025: Seller-side AI already automating routine tasks; AI listing assistance already in market.
- Q1 FY2026: Introduced AI-generated buyer profiles and built two agents in Q1 (buyer gift finder; seller insights/operations).
- Q1 FY2026: Developed an integrated Etsy app for ChatGPT (not yet launched as of the call).
- Q1 FY2026: Early ML/AI search and home-feed tests already showing improvements in add-to-cart and conversion (no magnitudes disclosed).
CURRENT (now)
- Q1 FY2026: Shifting search from moment-conversion optimization to relevance/personalization powered by machine learning and AI.
- Q1 FY2026: Testing conversational AI on Etsy; live buyer gift conversational agent; seller shop-management AI tools expanding.
- Q1 FY2026: Active integrations/partnerships with OpenAI, Microsoft, and Google; monitoring agentic traffic at a fraction of a percent of total traffic.
- Q1 FY2026: Machine learning improving Etsy Ads relevance and seller budget pacing, contributing to take-rate expansion.
- Q1 FY2026: Hybrid deployment of open-source, commercial, and internal models; internal AI used to ship agents in weeks.
- Q4 FY2025: Expanded agentic partnerships (Microsoft Copilot, Google, Stripe agentic payments); ML-personalized app home and owned marketing driving engagement now.
FORWARD (guidance)
- Q1 FY2026: Plans to build on AI-powered seller tools for listing and shop management; integrate seller identity/stories into core shopping experience this year.
- Q1 FY2026: Expect agentic/genAI shopping to become a meaningful discovery channel over time; continue proactive investment with ROI discipline on compute/tokens.
- Q1 FY2026: ChatGPT integrated app to launch; continued on-site conversational/agent experiences paired with Etsy data.
- Q4 FY2025: Road map includes more product experiences on Etsy beyond partnerships; conversational interfaces to capture richer intent on-site.
- Q4 FY2025: Management expects continued investment in ML/agentic within existing margin outlook while frequency/retention inflection takes time.
TRACK RECORD — PROMISE vs DELIVERY
42/100 track record over-promises 6 calls reviewed
Across six calls Etsy almost never sets AI/ML goals with both a number and a deadline; the one explicit personalization target (near-100% tailored email/push by end 2025) was not affirmed as achieved and disappeared from the narrative despite adjacent engagement gains.
Raise personalized email/push share from ~40% to nearly all buyer communications by year-end 2025 — promised Q2 FY2025
quietly-dropped Later calls cited stronger owned-channel performance (e.g., ~25% higher push/email clicks in Q4 FY2025) and continued ML-driven relevance but never reported reaching near-universal personalization and stopped citing the metric.
PRICED-IN (REFINED)
HIGH (already in)Est. revisions rising · Fwd P/E 53.0 · EV/Sales 2.1x
AI claim maps to Services Revenue, Marketplace Revenue
Price targets step up (last month $71.50 > last quarter $71.27 > last year $67.64) and holds have edged down while buy counts are stable, indicating modest upward revision momentum. Consensus already embeds a sharp EPS step-up ($1.32 in 2025 to $3.57 in 2026) on roughly flat revenue, so efficiency/upside is in the numbers—not a blank slate. At ~53x next-FY P/E and ~2.1x EV/Sales, the stock trades rich versus a mature specialty retailer, so AI-driven margin/services gains (chiefly Services Revenue, with some Marketplace GMV lift) look largely reflected.
COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS
7Q4 FY20248Q1 FY20258Q2 FY20257Q3 FY20257Q4 FY20259Q1 FY2026
AI enthusiasm across 6 calls — trend ↗ rising
From search ML and LLM craftsmanship tests to algatorial discovery and agentic checkout, then Kruti shipped buyer/seller agents and AI profiles with conversion lifts.
RECENT AI CATALYSTS & NEWS
BUSINESS IMPACT - QUALITATIVE MATERIALITY
7/10 qualitative impact material medium-term · mixed evidence
Where AI matters: personalized discovery, app engagement & Etsy Ads
Real shipped ML/LLM in search, buyer profiles, owned-channel personalization, seller agents, and ads pacing with hard engagement signals (app GMS +11.2%, push/email clicks +25%, take-rate expansion) but no isolated AI revenue/EPS and agentic traffic remains <1% after 15x growth.
Caveats: No dollar attribution—engagement gains are multi-causal and may already be in reported GMS; Agentic channel is tiny but structurally disintermediating if platforms own checkout/discovery; Credibility gap: explicit near-universal email/push personalization target quietly dropped; Compute/token costs and frequency inflection unproven—ROI discipline keeps spend inside margin outlook, not EPS drop-through
AI DISRUPTION / CANNIBALIZATION RISK two-sided · 5/10
Agentic discovery (ChatGPT/Google) can intermediate the shopper relationship and shift traffic off owned surfaces, while GenAI lowers listing friction and can dilute 'special' inventory/trust; durable offset is that gift/taste-led matching to heterogeneous seller SKUs is harder to commoditize than generic content or billable-hours services.
OPTIONS / MARKET STRUCTURE
option liquidity: fair
proxy inputs — dollar-ADV $212M · beta 1.898 · px $69.73
source: proxy (no options chain on FMP)
FMP /stable/ exposes no options-chain endpoint on this key, so ATM IV, bid-ask spread and open interest are unavailable. Liquidity below is a PROXY from dollar-ADV, beta and price level (a stand-in for option depth), not measured option-market data.
CONFIRMATION — INSIDERS · 13F · LANGUAGE
Undercutting — insiders selling, institutions flat, management language 5/10 measured.
INSIDERS selling 26 open-market sell(s) vs 0 buy(s) — net distribution
INSTITUTIONS (13F) flat as of 2026-03-31: 64 new / 101 closed positions; 231 increased / 176 reduced; institutional ownership -3.63pp; -48 net 13F holders
MGMT LANGUAGE 5/10 measured Real Q1 shipped AI/ML work and present-tense impact, but future agentic and seller tools stay early, testing, and plans.
commit “In Q1, we introduced AI-generated buyer profiles that help us go beyond the shoppers past activity.”
commit “In Q1, we built two agents, one focused on helping buyers find the perfect gift and another that brings together insights for sellers”
commit “we continue to benefit from machine learning-driven improvements to relevance and seller budget pacing”
VERBATIM AI QUOTES
“It's where personalization, machine learning and direct relationships come together most effectively, and we're seeing that show up in the numbers.”
— Kruti Patel Goyal, Q1 FY2026
“We're shifting toward a more personalized, relevance driven approach, powered by machine learning and AI that better understands what a buyer is looking for right now and their taste over time and the vast inventory offered by sellers on Etsy.”
— Kruti Patel Goyal, Q1 FY2026
“We're already starting to see positive signals from models that bring these elements together. With early tests showing improvements in add to cart rates and conversion.”
— Kruti Patel Goyal, Q1 FY2026
“In Q1, we introduced AI-generated buyer profiles that help us go beyond the shoppers past activity. With the goal of expanding the categories they explore and inspiring new purchases.”
— Kruti Patel Goyal, Q1 FY2026
“we're expanding ownership of our loyalty initiatives across product, engineering, marketing and operations because all of those moments together determine whether a customer chooses to return. For sellers, we're looking to reduce friction and enable growth with plans to build on our AI-powered tools to simplify listing and shop management so they can spend more time creating and connecting with buyers.”
— Kruti Patel Goyal, Q1 FY2026
“We are also intentional early movers in Agentec deeply focused on developing integrated experiences. We're encouraged by early engagement and traffic signals from Etsy's integrations with OpenAI, Microsoft and Google. And we recently developed an integrated Etsy app for Chat GPT.”
— Kruti Patel Goyal, Q1 FY2026
“At the same time, we're testing conversational AI experiences directly on Etsy because we see agents as a powerful way to simplify discovery and decision-making for both buyers and sellers, particularly when paired with our own data and insights.”
— Kruti Patel Goyal, Q1 FY2026
“In Q1, we built two agents, one focused on helping buyers find the perfect gift and another that brings together insights for sellers to make better decisions, access the right resources and reduce operational friction.”
— Kruti Patel Goyal, Q1 FY2026
“These are early examples of how ML and AI can make the marketplace meaningfully better for our customers. Just as importantly, they're allowing us to move faster, building and iterating in weeks, not months, which helps us learn more quickly and drive growth.”
— Kruti Patel Goyal, Q1 FY2026
“Etsy marketplace take rate expansion was led by Etsy Ads where we continue to benefit from machine learning-driven improvements to relevance and seller budget pacing.”
— Charles Baker (Lanny Baker), Q1 FY2026
“product improvements have also benefited AOV, including changes to our search and discovery algorithms that better surface higher quality, more relevant and differentiated inventory.”
— Charles Baker (Lanny Baker), Q1 FY2026
“we've introduced these new model that reflects much deeper buyer understanding. Basically -- develops a profile of a buyer of MAP to your interest in has over time and then map those interest inventory so that we can present to you when you first open the app, inventory that clearly connects to your taste but introduces you to new shopping missions and to discover new categories.”
— Kruti Patel Goyal, Q1 FY2026
“we've made a lot of good progress there in terms of richer inventory understanding, leveraging LLMs. The second is fire understanding or understanding our buyers' interest and taste -- and understanding those over time, that's part of what you were just describing as that richer context that persists.”
— Kruti Patel Goyal, Q1 FY2026
“one of the exciting developments that we shared in the shareholder letter in the prerecorded plus was about the conversational agent that we've just introduced for buyers to help them find gifts.”
— Kruti Patel Goyal, Q1 FY2026
“the investments we're making in AI and the features that we roll and the costs associated with the compute behind that is all going to be framed with reference to its impact on GMS growth and long-term user growth and frequency growth.”
— Charles Baker (Lanny Baker), Q1 FY2026
“we continue to see strong traffic growth and high-intent traffic. So that's great. but it's still very, very small. It's a fraction of a percent of our total traffic.”
— Kruti Patel Goyal, Q1 FY2026
“we really actively leverage a hybrid of options, open source models, commercial models and our internal models, really balancing capabilities with costs -- and we expect to continue to do that.”
— Kruti Patel Goyal, Q1 FY2026
“We have seen this already with the agents that we were able to launch in the last quarter really just built in weeks -- as we continue to see more of that opportunity to leverage tools to accelerate our work, we will deploy them actively.”
— Kruti Patel Goyal, Q1 FY2026
“I mentioned this in terms of some of the AI tools that we're launching, shopping shop management assistance, we've already talked about AI listing assistance.”
— Kruti Patel Goyal, Q1 FY2026
“Just over 20 years later, we're at an inflection point, one where the power of AI technology has the potential to make commerce on Etsy more human than ever, enhancing our differentiation and strengthening our unique customer value.”
— Kruti Patel Goyal, Q4 FY2025
“On the seller side, AI is already helping to automate routine tasks. So sellers can spend more time on what only humans can do, creating, designing and connecting with customers around the globe.”
— Kruti Patel Goyal, Q4 FY2025
“On the buyer side, it's making discovery easier and more relevant, helping Etsy show up in more of the moments where inspiration begins.”
— Kruti Patel Goyal, Q4 FY2025
“At the same time, AI-powered and agentic shopping presents meaningful opportunities for the unique items on Etsy to shine.”
— Kruti Patel Goyal, Q4 FY2025
“Since our last call, we've expanded our agentic shopping partnerships, adding integrations with Microsoft Copilot and Google as well as an agentic payments agreement with Stripe.”
— Kruti Patel Goyal, Q4 FY2025
“While still a very small part of our overall traffic in GMS, agentic traffic to Etsy in Q4 was about 15x what it was last year, underscoring just how rapidly this channel is emerging.”
— Kruti Patel Goyal, Q4 FY2025
“Using ChatGPT as an example, we're seeing evidence that in addition to bringing new buyers to Etsy, a meaningful share of buyers engaging through ChatGPT have a prior relationship with us, including lapsed buyers, indicating that a genetic shopping could be a great unlock for retention and better lifetime value. Orders originating from ChatGPT also tend to skew higher value compared to some of our more mature acquisition channels.”
— Kruti Patel Goyal, Q4 FY2025
“it's really critical to understand that we're really leveraging the capabilities, new capabilities of AI and the advancement in LLMs to really do things that were very, very much harder to do in the past, really deeply understand our inventory, which is incredibly unique and broad-based, much more deeply understand our buyers, their interest and their taste and match that with a stronger understanding more quickly of their intent to deliver a much more personalized content, really at every touch point off Etsy and on Etsy.”
— Kruti Patel Goyal, Q4 FY2025
“we've increased the coverage of those personalized recommendations, so that our push notifications and e-mails have gone from less than 1/4 of them being personalized to now over 3/4 of them being personalized. And all of that is driven by the advancements and investments that we've made in our machine learning-driven models.”
— Kruti Patel Goyal, Q4 FY2025
“What we're talking about in addition to the commerce is agentic discovery. And then the third piece that we're not talking about here is agentic capabilities. And those capabilities are not proprietary or owned by those platforms. There are capabilities that any company can use.”
— Kruti Patel Goyal, Q4 FY2025
“We're also using that to make our selling experience much better, much more efficient so that our sellers can spend their time doing what only they can do and really enhancing and amplifying the differentiation of Etsy and using it behind the scenes to make how we operate more effective, to make our support and trust and safety better, to make how we all operate day-to-day much more effective and efficient.”
— Kruti Patel Goyal, Q4 FY2025
“our first step is leveraging AI capabilities to pull that deeper understanding of our buyers' interests and of our inventory and to get signals of intent. But then the next step is to bring that same kind of interface to be able to capture even richer understanding of intent.”
— Kruti Patel Goyal, Q4 FY2025
ANALYST QUESTIONS ON AI
Q (Q1 FY2026, Nicholas Jones (BNP Paribas)): On internal AI: if AI unlocks richer persistent buyer context for customization, where are you in that evolution, and could expenses rise (e.g., tokens) to drive conversion?
A: Kruti: Exciting opportunity across three contexts—inventory (LLMs progress), buyer taste over time (early but good progress on taste profiles), and in-session intent; gift conversational agent gathers richer context in one interaction. Different maturity by area. Lanny: Embrace and experiment; manage costs with rigorous ROI tied to GMS, user growth, frequency, and long-term profitability.
Q (Q1 FY2026, Bryan Smilek (JPMorgan)): OpenAI may de-emphasize instant checkout vs discovery/SDK; you built a ChatGPT app—initial conversion impact? Last quarter agentic traffic was up multiples but still drives more on-platform conversions.
A: Kruti: ChatGPT app is not launched yet. Shift reinforces hypothesis that agentic shopping is a discovery channel; strong traffic growth and high intent but still a fraction of a percent of traffic. Stay early movers aligned with showing up where shoppers discover.
Q (Q1 FY2026, Jason Helfstein (Oppenheimer)): Where is Etsy on internal AI deployment scale, build vs buy, and should we expect a more aggressive push to automate internal workflows?
A: Kruti: Proactive holistically; hybrid of open source, commercial, and internal models balancing capability and cost. Internal AI is a force multiplier accelerating build cycles (Q1 agents in weeks); expect disciplined, proactive continued lean-in rather than one big step-change.
Q (Q4 FY2025, Michael Morton (MoffettNathanson)): What are you seeing in traffic/behavior from AI platforms, and could smarter off-platform answers compress the funnel and hurt on-site ads (CPC)?
A: Kruti: Very early; agentic traffic ~15x YoY but tiny; high intent, higher AOV, new + existing buyers, strong flow-through to Etsy for high-consideration items. Early mover partnerships help shape experiences and learn behavior. Lanny: Etsy Ads still under-monetizes seller demand; agentic often drives discovery then on-site journey; ecosystem may add ads over time—defensible on-site ads position. Kruti: Also emphasize agentic capabilities applied on Etsy, not only off-platform commerce.
Q (Q4 FY2025, Kenneth Gawrelski (Wells Fargo)): Is this a race between on-site/on-app discovery vs agents for unstructured Etsy listings—how do you ensure Etsy search beats external agents?
A: Kruti: Etsy should offer richer experience when shoppers know Etsy fits; agents excel at conversational intent capture—Etsy is using AI for inventory, interests, and intent signals, then bringing similar conversational interfaces on-site where Etsy has more data.
Q (Q4 FY2025, Deepak Mathivanan (Cantor)): What are you learning from agent traffic for native conversational experiences on Etsy, and how do you think about own/open-source models vs Gemini/GPT?
A: Kruti: Off-site learnings are discovery-channel quality (intent, AOV, engagement, flow-through); deeper learnings on conversational intent come from on-site experiments. Lanny: Agentic users show repeat usage—validates building agentic on Etsy for retention.
Q (Q4 FY2025, Shweta Khajuria (Wolfe Research)): Longer term, could AI agents disrupt take rates or let sellers transact off Etsy; how durable can frequency get with product improvements?
A: Kruti (agentic part): Very early; indicators support incremental discovery channel; impossible to predict every business-model pressure point; confidence in adapting as the world evolves because Etsy has done so before.