Airbyte — Airbyte Seed Deck

Deck context

About this deck

The 15-slide deck Airbyte used to raise its $5.2M seed round, led by Accel with Y Combinator and 8VC, announced 2 March 2021. Co-founder John Lafleur published the deck on Airbyte's own blog, and the file is hosted on the founder-owned SlideShare account. Note: the founders redacted parts of the deck before releasing it — the ask and prior raise appear as $XM and $YM, and customer names, angel investor names, roadmap features, headcount, KPI values and cash balances are replaced with placeholders. Study the narrative structure rather than the numbers. This seed presentation from 2021 contains 15 slides and records $5,200,000 in funding. Its sequence moves through cover, market, problem, solution, and later supporting sections. The outline below pairs every slide with its extracted text so researchers can study how the argument develops, not only how the pages look.

Key facts

Presentation structure

Slide-by-slide outline

1. 01

   Airbyte: simple and extensible open-source EL(T)Cover

   AIRBYTE. Simple & extensible Open-source EL(T). Diagram: extract from sources, load to destinations through our UI or API, transform raw schemas with compatibility with DBT or more.

2. 02

   Data is huge and growing, with commercial open source everywhere in the stackMarket

   Data is huge and growing, with COSS omnipresent in data infrastructures. Stack map across data sources (OLTP DBs via CDC, applications/ERPs, event collectors, logs, third-party APIs, file and object storage), ingestion and transformation (connectors: Fivetran, Airbyte - we are here; data modeling: dbt; workflow managers: Airflow; event streaming: Kafka; large-scale processing: Spark), storage (data warehouse, data lake) and data usage (data science platforms, ad hoc query engines, real-time analytics, dashboards and analytics).

3. 03

   Data integration is not a solved problemProblem

   But data integration is not a solved problem. Less than 2% of connectors are pre-built, so teams build and maintain other ones internally. No adaptability to custom use cases. Counter-productive pricing model based on volume. Third-party liability with data security and privacy.

4. 04

   Our goal: become the open-source standard for data integrationSolution

   Our goal: become the OSS standard for data integration. Diagram: connectors (DB, logs, file, URL, bucket, API) flow through optional normalization into a destination, turning raw data into analysis-ready schemas, and adapt to your stack in your infrastructure.

5. 05

   Open source makes data integration a commoditySolution

   Open-source enables us to finally make data integration a commodity. Cover the long tail of connectors. Extensible to address your exact needs. Feature-based pricing. Data security as a first-class citizen.

6. 06

   One product for analysts and for data engineersProduct

   A Fivetran-like UI for analysts and data scientists, shown as a sources list with connection frequency and last sync. An API for programmatic cases and data engineers, shown as a developer working against the Airbyte API.

7. 07

   The team that built LiveRampTeam

   The team that built LiveRamp. Michel Tricot, head of integration at LiveRamp and rideOS. John Lafleur, serial entrepreneur, 3 startups and 2 exits. Shrif Nada, technical lead at LiveRamp. Jared Rhizor, technical lead at LiveRamp. Charles Giardina, technical lead at LiveRamp and rideOS. Chris Duong, head of data at Comet. Logos: LiveRamp (NYSE: RAMP) and rideOS ($35M, Sequoia).

8. 08

   What we built until nowTraction

   What we built until now. Timeline from the pre-pivot data unblocking CDP: YC W20; seed of $1.8M; runway until June 2022; pivot to Airbyte on 07/23; MVP soft launch on 09/24; 0.8.0 release on 12/16. Progress by month: Aug 2020 three source connectors and three destination connectors with replication to one destination; Oct-Nov 2020 thirty-nine source connectors and four destination connectors with replication to multiple destinations; on 12/16, 380 GitHub stars, X companies syncing data, Y active contributors.

9. 09

   Roadmap through Q1 2022Milestones

   Roadmap. Quarterly grid from Q4 2020 to Q1 2022 covering product (main features and connectors), team (headcount), KPIs and cash in bank excluding revenues, ending at the runway date. The published version redacts the specific features, headcount, KPI values and cash figures as placeholders.

10. 10

    The ask and what it buysAsk

    Here is what we need today. We already raised $YM. Asking $XM to: win 3 months in KPIs to become the open-source leader for data integration; grow core engineering to help scale connectors faster; start building a community team to better engage it. The round size and prior raise are redacted as $XM and $YM in the version the founders published.

11. 11

    Appendix dividerCover

    Appendix. Section divider introducing the supporting slides the founders sent after meetings.

12. 12

    GitHub star growthTraction

    Our GitHub stars. Star history chart for airbytehq/airbyte climbing from near zero in September to roughly 380 stars by mid-December 2020. Source: https://star-history.t9t.io/#airbytehq/airbyte

13. 13

    Deployment growth by weekTraction

    Our deployments. Bar chart of weekly deployments rising steadily from 19 October to 15 December 2020, alongside a list of deploying companies whose names the founders redacted to "Company name" before publishing.

14. 14

    Airbyte versus Singer and MeltanoCompetition

    How Airbyte is unique in OSS. Comparison against Singer and Meltano on: number of pre-built connectors (goal 50+ by end of year after just 5 months, versus 96 mostly out of date in 3 years, versus 19 after 1+ year); for non-engineers (Fivetran-like UI with optional normalized data, versus standalone binaries, versus limited UI without normalization); for engineers (programmatic interactions with an API, versus a lot of manual work, versus a CLI interface for engineers); building new integrations (optimized for, can be in any language, versus as hard as building it yourself, versus same as Singer); integration within the existing data stack (supports K8s, Airflow and DBT, versus N/A, versus supports Airflow and DBT); project ownership and commitment (fully committed and VC-backed, versus no more commitment from Talend, versus a lab from Gitlab with one maintainer).

15. 15

    Our investors and angelsTeam

    Our investors. 8VC and Y Combinator, plus three angel investors whose names the founders redacted to "Investor Name, Founder of X, CEO of Y" before publishing, and others from Samba TV, Tango, Airbnb and Netflix.

Research notes

Questions about this deck