More Code with Jason episodes

334 - Merge Queues and Building Mergify with Julien Danjou thumbnail

334 - Merge Queues and Building Mergify with Julien Danjou

Published 10 Nov 2026

Duration: 00:55:02

The podcast explores the challenges of merging code changes in CI/CD, introducing the merge queue concept and discussing business strategies and future workflows.

Episode Description

In this episode I talk with Julien Danjou, CEO of Mergify, about merge queues: why PRs that pass CI separately can still break main together, and how...

Overview

The conversation explores merge queues as a solution to a specific CI/CD problem: when multiple pull requests pass tests independently but fail when merged together. Julian Donjo, CEO of Mergify, explains that the concept emerged from his work on OpenStack, where he and co-founder Medhi noticed almost nobody was solving this at scale. The product started around 2019 as a side project, initially serving open-source communities and small teams before growing through word of mouth.

The discussion also covers the business challenges of selling developer tools. Donjo describes how the first year was a "blue ocean" where no one knew what a merge queue was, making every conversation start with explaining the problem itself. He notes that GitHub's entry into the space in 2023 shifted the market, and the subsequent economic downturn forced a pivot from selling developer-experience improvements to building quantified ROI cases. The conversation also covers the parallel challenges of building Saturn CI, a CI platform, including finding initial customers and positioning against established products like CircleCI and GitHub Actions.

What If

  • What if you replaced your self-serve checkout with a single scheduling link and a short problem statement on your website?

    • Move: Add a Calendly-style booking link to your product page and de-emphasize the payment form, following Donjo's finding that it's "way more efficient to put a Calendly link actually on the website to talk to people."
    • Why Now? As a solo operator with an early-stage product, prospects need to verify you can solve their specific problem; a call lets you qualify and build trust in one interaction.
    • Expected Upside: This is a proposal, not a guarantee, but Donjo reports "like 90% of the sales we do is like talking to people," suggesting conversation-first could raise your close rate on inbound interest.
  • What if you launched into a narrow vertical where you are the only option, then expanded outward over time?

    • Move: Pick one specific pain point (e.g., "CI should parse and surface the exact test failure automatically") and market only to that slice, using the campfire logic: "You have to start with some tiny little twigs and then put some bigger twigs on top of that."
    • Why Now? As a solo developer you cannot compete head-on with GitHub Actions or CircleCI on breadth; a narrow niche means "you have no competitors" and lets you build a small audience before broadening.
    • Expected Upside: This is a possibility, not a certainty, but the source suggests starting narrow makes it "easier when you start" and gives you a defensible wedge to expand from.
  • What if you built a one-page ROI worksheet for each prospect before the sales call?

    • Move: Before each discovery call, draft a short document estimating the prospect's current CI failure cost (hours of debugging, compute waste) and projected savings, because for larger buyers "it's always boils down to what is my return on investment" and "you have to put numbers and dollars in front of that."
    • Why Now? If you are targeting teams beyond small startups, the buying environment shifted so that "there is no budget anymore" for vague experience improvements; a quantified case is the differentiator.
    • Expected Upside: This is a proposal; Donjo notes that when you can "build a case like that, it's pretty easy to sell," and the larger the customer, "the more you can build a ROI case, the easier it's going to be to sell your product."

Takeaway

  • Put a scheduling link (e.g., Calendly) on your product page and lead with a conversation rather than a payment form; Donjo found this "way more efficient" and reports "like 90% of the sales we do is like talking to people".
  • Launch into a single narrow niche where you face no direct competitors, then expand outward incrementally, following the pattern of starting with "tiny little twigs" before adding "bigger twigs".
  • Before pitching to teams larger than a few engineers, prepare a quantified ROI worksheet showing hours saved, compute costs avoided, and deployment-time reductions, because larger buyers ask "how much money I'm going to save" and you must "put numbers and dollars in front of that".
  • Work your personal network first - "friends on friends and acquaintance" - by scheduling demo calls with people you know, then expand to strangers and online marketing once those relationships are exhausted.
  • Revisit and reframe your value proposition when market conditions shift: if buyers stop funding "fun" developer-experience tools, pivot to measurable savings in CI time and engineering interruptions, as Mergify did after 2022.

Recent Episodes of Code with Jason

5 Sept 2026 327 - Will Carey, CTO of Brand New Box

"Explores how software agencies are shifting from coding to defining client needs, emphasizing AI's role in development while stressing the importance of conceptual design and real-world problem-solving."

More Code with Jason episodes