For founders selling devtools, OSS infrastructure, or developer SaaS · Updated June 2026

The buyer is a contributor.

Stars on competitor repos, issues asking for the feature you have, PRs by engineers evaluating both. GitHub is a public buying-committee map for anyone selling to developers.

7-minute read · 1 anatomy table · 1 reply template · 1 worked example

Part 1 · The diagnosis

Engineers evaluate in public.

For most B2B categories, the buyer evaluates privately. They sit in meetings, fill out vendor questionnaires, ask their network for opinions, and arrive at a decision the vendor can only infer. For devtools, the buyer's evaluation runs almost entirely in the open. They star repos they are considering. They file issues that describe their actual use case. They submit PRs to projects they want to learn the internals of. Every action is public, attached to their real identity, and frequently traceable to their employer.

That makes GitHub the densest signal layer in B2B. A developer who stars your competitor's repo this week is, with very high probability, in active consideration of that competitor and at least one alternative. A developer who files a feature-request issue on a competitor's repo is telling you, in technical detail, exactly the gap your product needs to address to win them. A developer who submits a PR is past consideration and into hands-on evaluation.

The play scales because the buyer and the user are the same person in devtools. Unlike most B2B where you have to navigate from the user up to the buyer, here you can write directly to the engineer who decides. Their team will defer to their technical opinion, and the procurement path is shorter than it looks because most devtools sit below the 50K-annual threshold where procurement gets formal.

The catch is that the writing has to be technical. An engineer who stars a competitor repo and gets generic sales outreach archives it inside a second. The same engineer who gets a substantive technical comparison email, ideally from another engineer, opens at a different rate entirely.

The rest of this page is the anatomy of which GitHub events convert and which do not, the engineer-to-engineer reply pattern, the composite case study of an open-source observability tool that built their entire qualified pipeline off competitor-repo stargazers, and how we would help run the play with you.

The four numbers that make this play work
Reply rate on engineer-to-engineer outreach
12 to 22 percent when the outreach is technical and concrete. Below 1 percent when it reads as sales. Source: outreach data from devtools campaigns, 2024-2026.
Conversion to evaluation
1 in 4 replies convert to a sandbox or trial. The signal is qualified by construction. Source: deal data from GitHub-sourced engagements.
Time from signal to evaluation
Roughly 7 to 14 days. Developer evaluation cycles are fast. Source: typical evaluation duration across devtools categories.
Cost of the GitHub feed
Free. GitHub REST and GraphQL APIs are public. Source: GitHub API documentation as of May 2026.
Part 2 · The anatomy of the signal

Different events, different intent.

GitHub events are not equal in intent. Stars are casual, issues are moderate, PRs are high-intent. The table below is the calibration matrix we use to decide which events to act on for which categories.

GitHub eventIntent strengthRight outreach shapeConversion
New star on competitor repoLow to moderateTechnical compare email1 in 10
Feature-request issue on competitorHighWe built exactly this, sandbox link1 in 3
Won't-fix comment by competitor maintainerVery highCompete on the gap directly1 in 2
PR submission to competitor repoVery high (hands-on eval)Engineer-to-engineer DM1 in 3
Star + employment at named coModerate, scaledAccount-level pursuit1 in 5
Star on adjacent tool you complementLowIntegration-first content1 in 12

The strongest case on the table is the won't-fix comment. When a competitor's maintainer closes a feature request with a "we won't build this," the engineer who filed it is immediately in the market for an alternative. The conversion rate on that cohort, when the outreach lands within 72 hours, is the highest of any play in this library.

The trap is unfiltered stargazer mining. A single new star on a competitor repo, by an engineer not employable at a target company, with no further activity, is a low-intent signal. Run the star-only motion against named accounts where the additional context can pre-qualify, not against the universe.

Part 3 · The reply pattern

Write like an engineer. Link to code.

The line between converting and getting blocked on this play is whether the outreach reads as engineer-to-engineer or as sales. Engineers smell sales copy from across the room and discard it. The pattern below is what works.

The engineer-to-engineer outreach pattern
For a star on competitor repo Channel: email, GitHub email if listed; LinkedIn DM as backup Subject: re: {competitor_repo} + {your_repo} Hey {first_name}, Saw you starred {competitor_repo} a couple of days ago. If you have not seen {your_repo} yet, the architectural difference that usually matters in your use case is {specific_technical_difference}. Concrete example: {link_to_specific_benchmark_or_demo_repo}. Not a pitch, just wanted to flag in case the compare saves you time. Happy to answer specific questions about {specific_concern_likely} if useful. {first_name_signoff} For a feature-request issue on competitor repo Channel: comment on the issue itself, then email/DM Public comment: "FWIW we built exactly this in {your_repo}, here is the implementation: {link}. Not trying to hijack the thread, just thought it might save you time while {competitor} decides on this." Email/DM follow-up (after the comment): Hey {first_name}, Noticed your feature request on {competitor_repo} for {specific_feature}. We built it in {your_repo} a couple of months ago, sandbox is at {sandbox_url} if useful. Would love your read on the implementation. {first_name_signoff} For a won't-fix comment Channel: DM via GitHub or Twitter, fast Hey {first_name}, saw {competitor} closed your request on {feature}. We have it: {link_to_specific_feature_page}. 20 minutes to compare? {first_name_signoff}

The reason the won't-fix DM is so short is that engineers value brevity. Two sentences with a useful link consistently outperforms a five-paragraph pitch to the same engineer. The medium rewards precision.

The public comment on the feature-request issue is a careful move. Done well it is a contribution to the community. Done poorly it reads as hijacking the thread and damages your reputation with both the issue filer and the competitor's community. Lead with humility, mention your repo without overselling, and step out if the OP signals annoyance.

Part 4 · A worked example

The OSS observability tool that ran all their pipeline off stargazers.

Composite drawn from open-source devtools founders running GitHub-signal outbound as the primary motion. Specifics anonymized; the arc is consistent with what the play produces in a developer-buyer category.

The team was a Series A open-source observability platform competing with Datadog and OpenTelemetry. Their open-source repo had 4K stars and a small but active contributor community. They had been running standard outbound to VPs of Engineering at target accounts and producing 1.8 percent reply rates.

They switched the motion entirely. Daily pipeline watching for new stars on the Datadog OSS repo and the OpenTelemetry repo, filtered to engineers employed at companies above 50 employees per LinkedIn enrichment. Engineer-to-engineer outreach within 48 hours, technical compare email, linking to a benchmark repo that ran the same workload on their tool versus the competitor.

Reply rate on the stargazer cohort came in at 14 percent, with 1 in 4 replies converting to a hands-on sandbox evaluation. Of evaluations, 30 percent converted to a paid pilot, then 60 percent of pilots converted to recurring revenue. Average ACV came in at 1.4x their cold-list cohort because the engineer was already technically pre-sold.

By month nine the GitHub-signal motion was producing roughly half their new logos at a tenth of the sender volume. They added monitoring for won't-fix comments on the competitor repos in month 10, which became the single highest-converting micro-cohort across the whole motion. They shut off cold-list outbound in month 11 and have not used a cold list since.

14%
Reply rate
1.4x
ACV uplift
~50%
Of new logos
Part 5 · How we would run this with you

The data is free. The writing is not.

GitHub's APIs are public and the integration work is straightforward. What separates teams that make this play work from teams that abandon it is the engineering-tone discipline of the outreach and the per-prospect technical depth. We pair with an engineer on your team so the outreach reads right.

Four pieces, repeated weekly, indefinitely:

01

GitHub feed across your category

Daily pull of stars, issues, PRs, and comments on competitor repos plus complementary OSS projects. Filtered to engineers at named target companies plus the open universe above a size threshold.

Output: daily event feed · enriched developers · ready
02

Per-event technical context

For each surfaced event, the surrounding technical context: the issue thread, prior commits, the engineer's GitHub activity pattern. Brief that lets the writer go straight to the point.

Output: per-event brief · technical depth · 5-min reply
03

Engineer-co-written outreach

We pair with an engineer on your team to write the technical outreach. Founder reviews the first 20, then we maintain the voice and your engineer reviews the new templates as new event types emerge.

Output: technical sequences · voice-matched · approved templates
04

Benchmark and demo asset library

The technical assets that earn the reply: benchmark repos, demo videos, side-by-side architectural breakdowns. We help build and maintain the library so each new outreach has a real artifact to point at.

Output: benchmark repos · demo library · architectural notes

The sizing call is short. You tell us your competitor repos and your engineering team's bandwidth, we tell you the typical event volume against your category and the realistic per-event conversion math, and you decide whether the engineer-pairing math works.

The 20-min sizing call

Tell us your competitor repos. We will tell you the event volume.

We will pull a sample week of stars, issues, and PRs on your competitor and complementary repos, send you the developer list with enrichment, and walk through the technical-tone outreach that has been working. If the volume justifies the engineer pairing, we can talk about running it.

Book the sizing call

Free for devtools founders. The sample event list is yours either way.