Upgrade to Pro to get implementation-ready specs for every company, the full report library, and 5 on-demand report requests per month.
If you only have a few minutes to spare, here’s what investors, operators, and founders should know about Goosebump (S17).
Goosebump was a Paris-based Summer 2017 company that helped people find electronic and live-music events through Facebook Messenger. Its observed consumer surface was a Messenger deep link; its event catalog came from Facebook's Graph API.[1][2]
That symmetry made the product efficient and brittle. Facebook supplied both the only documented interface and the inventory behind the recommendations. When it removed strategic event endpoints after Cambridge Analytica, Goosebump did not merely lose a cheap acquisition channel. It lost the catalog the bot needed to answer users.[2]
Co-founder Axel Delafosse dated the chatbot's death to around April 2018. The Goosebump and La Bringue brands still appeared in live-event promotion through September 2018, with an ambiguous partner-list mention in September 2019.[3][4][5] YC now lists the company inactive, but no corporate dissolution date, acquisition, or asset sale was found.[1]
YC names Raphael Doulonne as founder and CEO, Axel Delafosse as founder and CTO, and Henry Fontanier as founder and engineer, while the company card separately reports a two-person team.[1] The observed sources do not resolve that discrepancy, explain when each founder joined, establish how they met, or document the employers and schools that preceded the company.
Delafosse described Goosebump as a project built by electronic-music lovers to solve a problem they personally had. The team bootstrapped it while they were students and short on money, saw early traction, and positioned the product as more than a conventional event guide.[2] Its longer-term ambition was a marketplace around people's social lives.
The company also had a content and community precursor. A June 2016 SoundCloud playlist under Goosebump was branded La Bringue and paired an editorial interview with a mix from the Paris collective Classic As Fuck.[6] A January 2018 French music-startup directory linked Goosebump to hello@labringue.fr, named Doulonne, described a focus on techno and house in Paris and Lyon, and called the product a Messenger chatbot.[7]
The founders' own scene knowledge helped Goosebump secure event visibility. Resident Advisor called Goosebump and La Bringue an event partner in January 2018 and promoted the bot as a way to answer where the afterparty was.[8] Shotgun event pages in January, March, and April linked Goosebump recommendations, discovery campaigns, giveaways, or contest pages.[9][10][11]
The evidence index contains founder-authored reporting but not two preserved verbatim founder quotations about the company's origin, so the canonical two-quote requirement cannot be met without invention. The exact founder phrase used later in the post-mortem is retained because it is explicitly observed.
Goosebump answered a social planning question through Messenger: what electronic or live-music event should someone attend? YC describes discovery in a single message and points directly to m.me/higoosebump.[1] A bot directory and the French music-startup directory likewise identify Messenger, while no separate consumer web or mobile application was found.[7][14]
Facebook supplied the underlying event data through its Graph API. The observed sources do not document the exact endpoints, ranking model, personalization inputs, conversational script, or event-data schema.[2] The visible product proposition was simpler: replace browsing listings with a conversational recommendation.
Goosebump's discovery system also fed an editorial and promotional layer. Event listings called the company a recommender, partner, contest operator, or promoter. La Bringue connected the brand to playlists, interviews, and scene curation. This suggests, without proving a formal rebrand, that Goosebump combined software discovery with community media and live programming.
The product's most consequential design choice was platform concentration. Messenger reduced the cost of creating an interface, identity layer, and notification channel. The Graph API reduced the cost of assembling a catalog. Both advantages came from one provider. Goosebump did not own a direct user surface or a durable event feed when policy changed.
Goosebump targeted electronic-music fans in Paris and Lyon who wanted a fast answer rather than another event calendar.[7] Its event partnerships centered on techno and house. Promoters were the likely supply-side relationship, but no source establishes whether those relationships were paid placements, affiliate deals, data partnerships, or informal media swaps.
No verified user count, query volume, retention, conversion, revenue, or addressable-market figure was found. Event pages show figures such as 1,700 interested or sold-out ticket tiers, but those numbers belong to the events and organizers, not Goosebump.[9][11]
Resident Advisor now combines electronic-music discovery, personalized recommendations, editorial, and ticketing, and says it connects promoters in 50 countries with millions of music lovers each month.[15] DICE offers personalized discovery, ticketing, promoter analytics, and checkout integrations with Spotify and YouTube; it says 41% of tickets sold through DICE are prompted by Discovery.[16] Shotgun imports music preferences from Spotify or SoundCloud and connects with Songkick and Bandsintown.[17]
Those incumbents own more of the transaction. They combine inventory, audience signals, promoter tools, recommendations, and ticket conversion. A generic event chatbot would now be a thin interface over stronger systems. A credible rebuild needs a narrow social-planning job and direct or licensed inventory.
Goosebump's monetization was not established. The company participated in recommendations, contests, giveaways, partnerships, and event promotion, but no source shows affiliate revenue, ticket commissions, subscriptions, placement fees, or promoter contracts.
A French startup directory claims the company raised €500,000 from Kima Ventures and private investors on June 1, 2017.[12] This is uncorroborated secondary data. Delafosse independently confirmed that Goosebump had investors who were helping the team seek a continuation path after product failure, but he did not name them or disclose terms.[2]
No revenue, gross margin, acquisition cost, burn, runway, or unit economics were found. The economics cannot be reconstructed from event interest counts or ticket prices because Goosebump's share, if any, is unknown.
The strongest traction evidence is market presence, not usage. Event pages from Resident Advisor and Shotgun repeatedly named Goosebump as a recommender, event partner, promoter, or giveaway channel from January through September 2018.[8][10][4]
The July 2018 Café Barge event shows that brand activity survived the chatbot's failure. Independent coverage credited the Goosebump team with programming seven hours from house to techno, alongside a new outdoor sound system, activities, and a ticket giveaway.[13] This is evidence of community and promotional capability, not proof the Messenger product remained alive.
On June 29, 2018, Delafosse wrote that Goosebump had “died 2 months ago,” dating the chatbot's end to around April.[2] He directly connected the timing to Facebook's removal of event endpoints after Cambridge Analytica.
The structural cause was deeper than platform-dependent acquisition. Facebook supplied the only observed consumer interface through Messenger and the event inventory through Graph API. When the endpoints disappeared, Goosebump could still have a bot shell, but not the catalog required to recommend events. The change removed a production input.
The team had no documented fallback feed from promoters, ticketing systems, licensed aggregators, or its own crawler. Rebuilding the catalog would have required new contracts and normalization while the consumer product was already failing. Multiple interfaces would not have solved missing inventory; the architecture needed independent supply before policy changed.
Delafosse's same account says co-founder misalignment and conflict put Goosebump on a path toward death.[2] That is direct evidence of a second failure factor. The source does not quantify its contribution, describe specific decisions it blocked, or establish whether conflict alone would have ended the company.
The evidence does not rank conflict above or below platform loss. The platform event explains a specific product break in April 2018. Conflict explains internal fragility around the same period. Both can be material without a fabricated allocation of blame. No source documents mediation or a governance reset; Delafosse said only that investors were helping the team seek a way to continue its mission after the product failed.[2]
The cleanest timeline separates product, promotion, and corporation. The Messenger product died around April by founder account. Goosebump promoted a live Café Barge event in July and remained on a partner list in September.[3][4] A September 2019 third-party listing included Goosebump and Pool among partners, but does not prove an active operation.[5]
This counter-narrative matters. Goosebump was not erased the day the API closed; its scene relationships had residual value. Yet event curation and promotion were not the same product as a scalable Messenger discovery bot. No evidence shows that the live-event activity became a durable business.
Resident Advisor, DICE, and Shotgun demonstrate that music discovery remains valuable.[15][16][17] They also show why Goosebump's product layer was exposed. Modern incumbents combine direct promoter relationships, ticket inventory, fan taste, conversion, and analytics. Messaging can be a useful surface, but it cannot be the system of record.
No primary account from Doulonne or Fontanier, investor retrospective, or independent user discussion was found. No corporate dissolution date, acquisition, or asset sale is documented. YC's inactive status is the only confirmed company-level end state.[1]
Ready to rebuild Goosebump?
Implementation-ready specs, every report, and 5 on-demand requests each month.