Overview
Mage Market (legally Mage App Inc., originally Snapcardster / Snapp.ai GmbH) was a Y Combinator W19 company that built an authenticated bid/ask marketplace for Magic: The Gathering cards. Founded in Kiel, Germany in 2016 by Peer Richelsen and Malte Delfs, the company operated for nearly four years — evolving from a mobile card-scanning app into a centralized warehouse authentication platform modeled explicitly on GOAT's sneaker resale business. At its peak, it served over 50,000 customers across Europe and the United States.[1]
Mage failed because its core value proposition — authenticated physical goods with centralized warehousing — required capital-intensive infrastructure that its approximately $880K in total funding could never adequately support.[2] When COVID-19 triggered a 60% revenue decline in March 2020, the company had no profitability buffer and no path to raise additional capital, making shutdown the only viable option.[3]
Eight months of post-shutdown acquisition negotiations with two potential acquirers, one investor, and one potential new CEO all failed.[4] The company formally dissolved in December 2020. Richelsen went on to co-found Cal.com, a venture-backed open-source scheduling platform — a business with none of Mage's structural constraints.
Founding Story
The origin of Mage Market is unusually specific: a single week in March 2016 at opencampus.sh, a coworking space in Kiel, Germany.[5] During a "Prototyping Week" event, Peer Richelsen — a Magic: The Gathering player himself — developed the initial concept for an app that would make buying and selling MTG cards less painful. The idea was not born from market research or a gap identified in a pitch competition. It came from a hobbyist's frustration with an existing, fragmented secondary market.
Richelsen co-founded the company with Malte Delfs, who served as CTO.[6] Public records do not detail how the two met or their professional backgrounds prior to Snapcardster, but the division of labor — Richelsen as product and business lead, Delfs as technical architect — held throughout the company's life. The founding team remained small: at the time of YC acceptance, the company listed three co-founders in the YC directory, though the team grew to nine employees at its peak.[7]
The company's early identity was entirely community-driven. It launched under the name Snapcardster, ran a Kickstarter campaign in February 2017 with a modest €6,000 goal, and raised €6,171 from 121 backers — enough to validate demand, not enough to build infrastructure.[8] The legal entity was Snapp.ai GmbH, a German company. In October 2017, a Swiss seed investor provided the first institutional capital, increasing share capital to €25,000.[9]
The pivot from hobbyist app to institutional marketplace accelerated when the team applied to Y Combinator. Acceptance into the W19 cohort — making them the 7th German team ever admitted — prompted a full rebrand from Snapcardster to Mage and a relocation to Silicon Valley for the program.[10] The YC experience reshaped the product vision. Richelsen later described the program's impact: "Y Combinator had multiple events per week and offered private 1-to-1 meetings with the partners. These partners have decades of experience building and scaling startups and had an incredible impact on the product and idea."[11]
The YC period crystallized the company's strategic analogy: do for Magic: The Gathering what GOAT had done for sneaker resale.[12] This framing was intellectually coherent — both markets involved high-value collectibles with authenticity risk and fragmented peer-to-peer supply — but it imported GOAT's capital requirements alongside its business logic. GOAT had raised tens of millions of dollars before achieving the warehouse scale that made its authentication model viable. Mage would attempt the same model with a fraction of that capital.
The company's name arc — Snapcardster → Snapp.ai GmbH → Mage App Inc. — traces a genuine maturation: from niche community app to institutional marketplace with global ambitions. Each rename reflected a real strategic shift. The final name, Mage, was clean, brandable, and market-agnostic enough to eventually expand beyond MTG. That expansion never came.
Timeline
- March 2016 — Peer Richelsen develops the initial concept for a Magic: The Gathering card app during Prototyping Week at opencampus.sh in Kiel, Germany.[13]
- February 2017 — Snapcardster launches a Kickstarter campaign with a €6,000 goal; raises €6,171 from 121 backers.[14]
- October 2017 — Snapcardster receives its first institutional seed investment from a Swiss investor; share capital increased to €25,000.[15]
- February 2018 — The Snapcardster app reaches 10,000 users for the first time.[16]
- January 2019 — Company joins Y Combinator Winter 2019 cohort as the 7th German team accepted; rebrands from Snapcardster to Mage.[17]
- June 2019 — Mage announces plans for v2.0 marketplace with bid/ask features, centralized warehouse authentication, vault storage, and North American/European warehouse expansion; plans to close a second funding round by end of year.[18]
- 2019 — Mage raises seed funding from Y Combinator and angel investors from Twitch; relaunches marketplace as v2.0 with ask/bid features.[19]
- January 2020 — North American warehouse expansion paused due to visa issues; Mage announces controversial Klarna financing partnership.[20]
- March 29, 2020 — COVID-19 causes 60% revenue decline; Peer Richelsen announces controlled shutdown of all operations.[21]
- March–April 2020 — Future purchases and sales deactivated; platform remains accessible until at least April 30, 2020. Mage begins acquisition discussions with several competitors.[22]
- December 3, 2020 — After 8 months of failed negotiations with two potential acquirers, one investor, and one potential new CEO, Richelsen announces formal dissolution of the company.[23]
What They Built
Mage Market's product evolved through two distinct phases, each representing a more ambitious — and more capital-intensive — interpretation of the same core problem: making the Magic: The Gathering secondary market safer and more liquid.
Version 1: The Scanning App
The original Snapcardster product was a mobile application built around a proprietary computer vision system. Players could point their smartphone camera at a Magic card, and the app would identify it within seconds, retrieve current market pricing, and enable listing for sale.[24] This addressed a genuine friction point: MTG cards number in the tens of thousands across dozens of sets, and manually identifying and pricing a collection was time-consuming and error-prone. The scanning feature reduced the listing process from minutes to seconds per card.
The v1 product was a peer-to-peer marketplace with Mage acting as a transaction facilitator. It was a software-first business with relatively low operational overhead — the kind of product that could scale without proportional cost increases.
Version 2: The Authenticated Exchange
The v2.0 relaunch in Summer 2019 was a fundamental architectural shift. Mage moved from peer-to-peer facilitation to a centralized authentication model.[25] The mechanics worked as follows:
- A seller lists a card with an "ask" price on the platform.
- A buyer places a "bid" at their target price.
- When bid and ask meet, the transaction executes automatically — like a stock exchange.[26]
- The seller ships the card not to the buyer, but to a Mage warehouse.
- Mage's authentication experts inspect the card for counterfeits and condition misrepresentation.[27]
- Only after passing authentication does the card ship to the buyer — or enter the vault.
The "vault" feature was the most sophisticated element of the v2.0 product. Buyers could purchase cards and elect to store them at Mage's warehouse rather than take physical delivery.[28] Vaulted cards could be relisted for sale directly from the warehouse, eliminating the shipping step entirely for subsequent transactions. For high-value cards that traded frequently — a Black Lotus or a set of Power Nine — this created a capital-efficient instrument: the card could change hands multiple times without ever leaving Mage's custody.
The GOAT analogy was explicit in Mage's positioning.[29] GOAT (and its eventual merger partner StockX) had demonstrated that authenticated resale could command a meaningful price premium over unverified peer-to-peer transactions. The insight was sound: MTG players had long suffered from counterfeit cards, particularly in high-value formats, and a trusted authentication layer had genuine value.
In early 2020, Mage added a Klarna buy-now-pay-later integration, allowing customers to finance card purchases.[30] The move was described as "controversial" in contemporary coverage, though the specific nature of the controversy — whether it was community backlash against financing collectibles, regulatory friction, or operational complications — is not documented in available sources.
What distinguished Mage from alternatives like TCGPlayer was the authentication guarantee and the exchange-style pricing mechanism. TCGPlayer operated as a traditional marketplace where buyers selected from individual seller listings at fixed prices. Mage's bid/ask model created a single clearing price, reducing the cognitive load of price discovery and eliminating the authenticity risk that plagued peer-to-peer transactions.
Market Position
Target Customers
Mage's primary customers were serious Magic: The Gathering players and collectors — specifically those engaged with the "paper" (physical card) format rather than the digital game. The average customer spent $120 per month on Magic cards,[31] a figure that places them firmly in the "competitive player" or "speculator" category rather than the casual hobbyist. At that spend level, authenticity risk is material: a single counterfeit card in a competitive deck can result in disqualification, and high-value singles (cards worth $50–$500+) are the primary targets for forgery.
The secondary customer segment was card speculators — players who bought cards as financial instruments, anticipating price appreciation from tournament results or set rotations. For this group, the vault feature had particular appeal: it enabled rapid trading without the friction and cost of repeated shipping.
Market Size
The Magic: The Gathering secondary market is large by hobby standards but small by venture standards. Wizards of the Coast (a Hasbro subsidiary) does not publish secondary market data, but industry estimates placed the global MTG singles market in the hundreds of millions of dollars annually during Mage's operating period. TCGPlayer, the dominant US marketplace, was acquired by eBay in 2022 for a reported $295 million — a valuation that implies meaningful but not enormous scale.
Richelsen himself identified this constraint in retrospect: "While Magic: The Gathering is a fun hobby, it's still fairly niche and I think it's equally hard to build a startup in a small market than in a big market, you still deal with the same shit every day anyway!"[32] The operational burden of running a warehouse, authenticating cards, managing logistics, and supporting customers is roughly constant regardless of whether the total addressable market is $500 million or $50 billion. Mage's TAM could not generate the revenue needed to justify the infrastructure required to serve it.
Competition
Mage operated in a competitive landscape with a clear structural hierarchy. The relevant axes were distribution reach (number of active buyers and sellers) and product depth (authentication, price discovery, user experience).
TCGPlayer was the dominant incumbent in North America, with a large established network of buyers and sellers, deep liquidity, and a trusted brand. It operated as a traditional fixed-price marketplace without centralized authentication — meaning Mage's authenticated exchange model was genuinely differentiated on product depth. However, TCGPlayer's distribution advantage was enormous: its liquidity network was self-reinforcing, and sellers listed there because buyers were there, and vice versa. Mage needed to overcome this cold-start problem while simultaneously building physical infrastructure.
Card Kingdom occupied the premium end of the market — a retailer and buylist operator known for high-quality grading and customer service. It competed with Mage on trust and authentication but operated as a principal (buying and selling its own inventory) rather than a marketplace.
eBay provided the long tail of peer-to-peer transactions with no authentication guarantee. It was the market Mage was explicitly trying to improve upon, not compete with directly.
The structural problem for Mage was that TCGPlayer's distribution advantage was not something product quality could overcome at seed-stage funding levels. TCGPlayer could, in principle, add authentication features to its existing platform — and indeed, eBay's acquisition of TCGPlayer in 2022 brought resources that could fund exactly that kind of product investment. Mage was competing on a dimension (authentication) where it had a temporary lead, but incumbents had the distribution and capital to absorb that feature once the market validated it.
The competitive landscape also shifted because of the underlying platform: the paper MTG market itself was dependent on in-person play infrastructure — Magic Fests, local game stores, Friday Night Magic events. When COVID-19 eliminated that infrastructure, demand for physical cards collapsed simultaneously across all competitors. But Mage, unlike TCGPlayer or Card Kingdom, had no balance sheet to weather the disruption.
Business Model
Mage operated as a transaction-fee marketplace. The company took a percentage of each completed transaction — the standard model for authenticated resale platforms. No public revenue figures were ever disclosed; the absence of revenue data in all available sources is itself a signal that the company never reached a scale at which it was comfortable publicizing financial metrics.
The unit economics can be partially reconstructed from available data. With $120/month average customer spend[33] and 50,000 customers,[34] gross merchandise value (GMV) at full engagement would be approximately $72 million annually — but this figure almost certainly overstates active GMV, as the 50,000 customer figure likely represents cumulative registrations rather than monthly active buyers. These are inferences, not reported figures.
On the cost side, the v2.0 model required warehouse leases, authentication labor, shipping logistics, and payment processing — fixed and semi-fixed costs that do not scale down easily during revenue shortfalls. With approximately $880K in total funding across three years,[35] and a team of up to nine employees plus warehouse operations, the company was almost certainly burning through capital faster than it was generating revenue. Estimated annual burn at up to 9 employees plus warehouse operations would likely exceed $700K–$900K, implying roughly one year of runway from the YC raise alone — again, an inference from headcount and operational model, not a reported figure.
The planned second funding round — intended to close by end of 2019 to fund US and EU warehouse expansion[36] — appears never to have closed. No announcement, press coverage, or Crunchbase entry documents a post-YC institutional round. If accurate, this means Mage entered 2020 running on depleted YC-era capital, with a warehouse model it could not fully fund and a North American expansion already stalled by visa issues.
Traction
Mage achieved meaningful user growth for a niche collectibles marketplace. The Snapcardster app reached 10,000 users by February 2018,[37] approximately two years after founding. By the time of shutdown, the platform had served over 50,000 customers across Europe and the United States.[38]
The $120/month average customer spend figure[39] is the most telling traction metric available. It indicates that Mage had successfully attracted the high-value segment of the MTG market — competitive players and speculators who transact frequently and in volume — rather than casual collectors. This is the right customer segment for an authenticated exchange model: these users care most about authenticity guarantees and price efficiency.
The demand signal was real. The problem was not that customers didn't want the product. The problem was that the 50,000-customer base, even at $120/month average spend, generated a GMV that could not support the operational infrastructure required to serve them at the authentication quality the model promised. No GMV or revenue figures were ever publicly disclosed, which prevents a precise assessment of whether the unit economics were ever viable at any achievable scale within the MTG market.
Post-Mortem
Primary Cause: A Capital-Intensive Model Built on a Seed-Stage Balance Sheet
The most structurally damning fact about Mage Market is the gap between its business model and its funding. The v2.0 authenticated warehouse model — the product that Mage bet its future on after YC — required physical warehouses, authentication labor, logistics operations, and working capital to hold inventory in transit. GOAT, the explicit inspiration for this model, had raised approximately $60 million before achieving meaningful warehouse scale. Mage attempted the same architecture with approximately $880K in total funding across three years.[40]
The planned second funding round, announced in June 2019 as the mechanism to fund US and EU warehouse expansion,[41] appears never to have closed. No post-YC institutional round appears in any public record. This is not a minor operational detail — it is the central financial fact of the company's final year. Without that round, Mage was attempting to operate a capital-intensive physical-goods business on a depleting seed-stage balance sheet.
The team tried to address this by launching the Klarna financing integration in early 2020, presumably to increase average order value and transaction volume — generating more fee revenue without raising more capital.[42] The move was described as "controversial," suggesting it generated community friction rather than the revenue boost the team needed. The attempt failed to solve the underlying capital problem.
Secondary Cause: The North American Expansion Stalled Before COVID Arrived
The v2.0 strategy required North American warehouse operations to serve the largest MTG market. By January 2020 — two months before COVID — that expansion was already "paused" due to visa issues.[43] The nature of the visa problem is not documented in available sources, but the consequence is clear: the growth engine of the post-YC strategy was non-operational at the moment the company most needed revenue growth.
This is significant because it means Mage was already in a weakened position when COVID hit. The company had not reached profitability after nearly four years of operation, its planned fundraise had not closed, and its primary expansion initiative was stalled. Richelsen acknowledged this directly: "we've not found a way to build a profitable business fast enough in order to survive such global crisis and revenue fallout."[44] The word "fast enough" is doing significant work in that sentence — it implies the team believed profitability was achievable, but that the timeline had run out.
Tertiary Cause: Structural Dependence on In-Person Play Infrastructure
The paper MTG market is not a standalone e-commerce category. It is downstream of in-person play. Cards gain value when they perform well in competitive tournaments (Magic Fests, Grand Prix events, local game store Friday Night Magic). Players buy cards to play them, and they play them in person. When COVID-19 eliminated in-person gatherings, it did not merely reduce demand — it eliminated the primary mechanism by which card values were discovered and validated.
Richelsen identified this explicitly in his shutdown announcement: the cancellation of Magic Fests, local game store shutdowns, and the end of private game nights due to quarantine were cited as reasons the company "do not believe the market will recover in the next 12 months, potentially never."[45] This was a reasonable assessment in March 2020, though the paper MTG market did eventually recover as in-person play resumed. The problem was that Mage had no runway to wait for that recovery.
This structural dependency was not unique to Mage — TCGPlayer and Card Kingdom faced the same demand shock. But those incumbents had the balance sheets and revenue scale to survive a 60% revenue decline. Mage did not.
Quaternary Cause: Niche Market TAM as a Structural Ceiling
Richelsen's own post-mortem reflection identifies a lesson that goes beyond COVID: "While Magic: The Gathering is a fun hobby, it's still fairly niche and I think it's equally hard to build a startup in a small market than in a big market, you still deal with the same shit every day anyway!"[46]
This is a precise diagnosis. The operational complexity of running an authenticated physical-goods marketplace — warehouse management, authentication labor, logistics, customer support, fraud prevention — is roughly constant regardless of market size. A business serving a $500 million TAM faces the same organizational challenges as one serving a $50 billion TAM, but with a fraction of the revenue ceiling to justify the investment. Mage's market was large enough to attract real customers but too small to generate the revenue needed to fund the infrastructure its model required.
The company could not expand into adjacent collectibles categories (sports cards, Pokémon) without additional capital — and that capital was not forthcoming. The niche market was not a temporary constraint to be grown through; it was a ceiling that the business model could not break through at the funding level available.
The Shutdown Decision and Its Aftermath
Richelsen announced the controlled shutdown on March 29, 2020, approximately one month after the COVID revenue decline became apparent.[47] The decision was deliberate and ethically considered: rather than pivot and cause layoffs, the founders used remaining capital to pay extended severances. "My cofounder Malte and I believe that in these times, the right thing to do is to announce a controlled shutdown of operations and use the little money left to pay all employees extended severances."[48]
This decision foreclosed any recovery attempt. Once the capital was committed to severances, there was no runway to execute a pivot or wait for acquisition terms to improve. The subsequent eight months of negotiations — with two potential acquirers, one investor, and one potential new CEO — all failed.[49] The identity of the potential acquirers is unknown, but the failure of all four negotiation tracks suggests the business had no standalone value that could be transferred: the technology (computer vision, exchange infrastructure) was not proprietary enough to command an acqui-hire premium, and the customer base was too small and too niche to justify a strategic acquisition.
The company formally dissolved in December 2020.[50] Mage Market was the first major Magic: The Gathering marketplace to close due to COVID-19.[51]
Key Lessons
-
Applying a capital-intensive model to a niche market creates an unfundable gap between required infrastructure and achievable revenue. Mage explicitly modeled itself on GOAT's authenticated sneaker resale business — a sound strategic analogy — but GOAT raised tens of millions of dollars before achieving warehouse scale. Mage attempted the same physical infrastructure with approximately $880K in total funding across three years. The MTG secondary market was large enough to attract 50,000 customers spending $120/month, but not large enough to generate the revenue needed to justify the warehouse operations the authentication model required. The lesson is not "don't build physical-goods marketplaces" — it is that the GOAT model requires GOAT-level capital, and a niche TAM cannot support that fundraise.
-
A planned fundraise that doesn't close is an existential event, not a scheduling problem. Mage announced in June 2019 that it planned to close a second funding round by end of year to fund US and EU warehouse expansion. No such round appears in any public record. The company entered 2020 with a warehouse model it could not fully fund, a North American expansion already stalled by visa issues, and a balance sheet depleted by three years of operations. When COVID hit in March 2020, there was no buffer. Startups building capital-intensive models should treat a failed fundraise as a forcing function to immediately reduce burn to a sustainable level — not as a temporary delay.
-
Niche market size is a structural constraint on the operational model, not just a growth challenge. Richelsen's own post-mortem reflection is the clearest articulation of this lesson: building a startup in a small market carries the same operational burden as building in a large one, but with a fraction of the revenue ceiling. Mage's warehouse, authentication labor, logistics, and customer support costs were roughly constant whether the TAM was $500 million or $50 billion. The company could not expand into adjacent collectibles categories without additional capital, and the MTG market alone could not generate the revenue to fund that expansion organically. Richelsen's subsequent company, Cal.com, is software-only, globally scalable, and not dependent on physical infrastructure — a direct structural inversion of Mage's constraints.
-
Physical-goods marketplaces serving hobbyist communities carry hidden demand dependencies on in-person infrastructure. The paper MTG market is downstream of in-person play: cards gain value through tournament performance, and players buy cards to play them in person. Mage's demand was structurally coupled to Magic Fests, local game stores, and Friday Night Magic events. When COVID eliminated those gatherings, it did not merely reduce demand — it eliminated the price discovery mechanism that made the secondary market function. Any marketplace serving a community whose primary activity is in-person should model the scenario in which that activity is disrupted, and stress-test its runway against a 60% revenue decline.
-
The ethical shutdown decision, while admirable, permanently foreclosed recovery options. Richelsen's choice to use remaining capital for extended employee severances rather than a pivot attempt was a principled decision that protected his team during a crisis. It also meant there was no runway left to execute on the eight months of acquisition negotiations that followed. Founders facing shutdown should sequence decisions carefully: exploring acquisition and pivot options before committing capital to wind-down costs preserves optionality. Mage's controlled shutdown was handled with integrity — GDPR-compliant data destruction, Stripe-mediated fund protection, transparent communication[52] — but the sequencing meant that by the time acquirers were engaged, there was nothing left to negotiate with.
Sources
- Undefeated Underdogs Podcast, Episode 57 — Mage Market
- Peer Richelsen — Shutting Down Mage Market (Medium)
- Hipsters of the Coast — Mage Market to Shut Down (March 30, 2020)
- Y Combinator — Mage Company Profile
- Snapcardster Company Blog
- Hipsters of the Coast — Mage to Launch a Stock Market for Magic Cards (June 20, 2019)
- Crunchbase — Mage Market
- Crunchbase — Peer Richelsen
- Mercury — Founder Spotlight: Peer Richelsen of Cal.com (July 6, 2023)
- opencampus.sh / Medium — Mit seinem Startup von Kiel ins Silicon Valley: Snapcardster zeigt wie es geht
Corrections
The following correction was provided directly by Mage's founder:
- Team size: At its peak, Mage had 9 employees — not 3 as originally stated in this report. The YC directory only lists co-founders, not all employees.