The Rival Issue-Thread Read: Pitch the Account, Time It, Brief Only, or Pass
Score any rival repo issue on six dimensions and reach Pitch, Time It, Brief Only, or Pass, with a dated re-check and a contact route that never touches the thread.
Key takeaways
- A closed GitHub issue marked 'completed' proves nothing on its own, because closing without a reason defaults to completed; only completed plus a linked pull request shows the rival shipped a fix.
- 'Closed as not planned' can mean under-voted rather than refused: VS Code's bot closes requests that miss 20 upvotes in 60 days, so read the closing comment before treating it as a considered refusal.
- Community-filed issues averaged 39 days to close against 5.9 days for team-filed ones in Kikas et al., so the 24-hour outreach rush that vendor guides push buys little and skips the checks that decide the verdict.
- In Refolk's index, 2 of 25 sampled US Pulumi profiles (8%) were Pulumi's own employees, so exclude MEMBER and COLLABORATOR filers before you try to resolve any account.
- In Refolk's German Terraform sample, 44% held founder or CXO titles and 12% were consultants, self-employed, or fractional, so many filers are either the buyer themselves or tied to no single account.
- Never post in the rival's thread and never email a commit address; reach the account through a work address found outside GitHub, with an Article 14 notice and an Article 21 opt-out.
This read is for founders, account executives, and SDR leads who have found an issue or discussion thread in a competitor's or incumbent's public repo where someone describes a pain their product solves. It scores that one thread on six dimensions and gives one of four verdicts: Pitch the account, Time It, Brief Only, or Pass. Each verdict comes with a dated re-check and a contact route that never goes through the thread itself or a scraped commit email.
What does the rival issue-thread read decide?
The read turns one rival issue thread into one of four verdicts: pitch the filer's account now, time the approach to a dated maintainer decision, put the pain in your battlecard only, or pass. A battlecard is the internal sheet your reps use to position against a named competitor.
Most advice on how to prospect from competitor GitHub issues treats any thread that mentions a rival as a lead. One ranking prospecting guide says that an issue asking how to migrate means "that's a purchase-ready prospect." That skips the two facts that actually decide the outcome:
- Whether the window is open. The maintainers may have closed the issue as not planned, put it on a milestone, or merged a fix that ends the pain before your email lands.
- Whether there is an account at all. The filer may work for the rival, consult for a client nobody names, or have an employer string that went out of date when their company was acquired.
This read judges a single thread: where it is in its lifecycle, who filed it, who else reacted, and how its window closes. It does not score an account's overall engagement across repos, or the stack files an account commits. Those are separate reads.
| Verdict | What it means | Re-check |
|---|---|---|
| Pitch | Open window, confirmed account, exact fit | Thread state on send day |
| Time It | Confirmed account, decision pending | Demand-bar deadline or linked PR |
| Brief Only | Real pain, no account or window shut | Employer source at 90 days |
| Pass | No fit, cosmetic pain, or noise | Only if reopened |
Window state against account resolution
A fix that has already shipped to a confirmed account usually drops to Brief Only. Time It applies only while the fix is scheduled or still being tested for demand.
How to read the maintainer state of a rival issue
Maintainer state tells you whether the window the thread opened is still open. A declined issue leaves it open, a shipped fix closes it, and a scheduled or demand-tested issue puts it on a clock. Read it before anything else, because it can end the read in five minutes.
Every field you need is in the issue payload and can be searched, as the GitHub REST issues reference documents:
- state: open or closed.
- state_reason: completed, not_planned, duplicate, reopened, or null.
- labels and milestone.
- linked pull requests, searchable with
linked:pr. - author_association: the filer's relationship to the repo, for example COLLABORATOR or MEMBER.
| State | Search qualifier | Meaning for the verdict |
|---|---|---|
| Fixed | reason:completed + linked:pr | Window closing |
| Declined | reason:"not planned" | Window open |
| Duplicate | state_reason duplicate | Read the canonical issue |
| Demand test | Backlog Candidates milestone | Time It, 60-day clock |
The search qualifiers come from GitHub's documentation. The verdict column is this framework's interpretation, not something GitHub says.
What each state proves, and how it lies
- Fixed proves the rival shipped a change, but only when a linked pull request goes with it. On its own, "completed" is weak evidence: closing an issue without picking a reason defaults to completed, even when the real reason was "not a bug" or "duplicate".
- Declined proves the rival has no current plan to build it, which leaves the pain to you. It lies when a bot closed the issue for missing a vote target rather than a maintainer judging it on merit.
- Duplicate proves the pain lives somewhere else. Closing as duplicate through the API requires a
duplicate_issue_id, so the canonical issue is always named. Follow it and start the read again there. - Demand test proves nothing yet. Milestone names are specific to each repo. In VS Code, "Backlog Candidates" means the request is still being tested for demand: "The community has 60 days to upvote the issue. If it receives 20 upvotes we will move it to our backlog." Requests that miss the bar are closed, as #322225 was. #295755 cleared the bar in 4 days, so a demand test can end within days or run its full 60.
These searches sort a repo's threads quickly. Replace OWNER/REPO with the rival repo:
repo:OWNER/REPO is:issue is:closed reason:completed linked:prfinds threads that were really fixed.repo:OWNER/REPO is:issue is:closed reason:"not planned"finds declined threads.repo:OWNER/REPO is:issue is:open sort:reactions-desclists open pain, most-reacted first.
Scoring dimension 1: maintainer response state (window)
- 3: Declined as not planned by a maintainer, or open for a long time with no milestone and no linked PR.
- 2: Open and not yet triaged, or reopened.
- 1: Scheduled on a milestone, or inside a published demand test.
- 0: Completed with a linked PR and a shipped release.
A closed issue tells you the maintainers stopped talking. Only a linked pull request tells you they shipped.
How to score pain severity, recency, and momentum
Pain severity tells you whether someone would pay to make the problem go away. Recency and momentum tell you whether the pain is still live and shared by more than one person. Score each 0 to 3.
Dimension 2: pain severity and workaround
- 3: Blocking, with no workaround in the thread.
- 2: Painful, and the only workaround posted is ugly or partial.
- 1: Painful, but a clean workaround is posted and confirmed to work.
- 0: Cosmetic.
What it proves: someone hit a cost they cannot route around. How it lies: angry wording on a cosmetic issue, or a workaround posted several comments below the original complaint that the filer then thanked someone for. Read the whole thread, not just the opening post.
A GitHub issue asking how to migrate off the rival usually scores 3 here, because the filer has stated a wish to leave. That makes it a strong prospect, not a confirmed one. It still has to pass the window and employer checks.
Dimension 3: recency and momentum
Record two dates and one count: when the issue was opened, the date of the last non-bot comment, and the number of reactions. Then compare them with any demand bar the repo publishes. No general threshold has been published for when a pain counts as recurring rather than a one-off. The VS Code bar of 20 upvotes in 60 days is the only maintainer bar documented in my sources. Use the rival's own bar where one exists, and say plainly when you are using a proxy instead.
Sort by reactions with sort:reactions-desc in search, or filter by count with the gh CLI's --reactions flag. Users have reported that the dropdown for sorting by reactions disappeared from the web UI while the search qualifiers kept working. Rely on the qualifier, not the dropdown.
Where the rival publishes no bar, the closest public evidence is how long issues take to close:
| Study | Sample | Median | Mean |
|---|---|---|---|
| Scientific workflow systems | 21,116 issues | 18.09 d | 190.65 d |
| ML issues | Controlled sample | 25 d | 101 d |
| Non-ML issues | Controlled sample | 11 d | 61 d |
| Kikas et al. | GitHub projects | 2 to 4 d | Community 39 d vs team 5.9 d |
The last row matters most for this read. Issues filed by outsiders stay open far longer than issues filed by the team: 39 days on average against 5.9. Vendor guides say issues stay high-intent for 7 to 14 days and send high-intent signals to outreach within 24 hours. The benchmarks point the other way. Pain filed by an outsider tends to stay unresolved well beyond that 7 to 14 day window, so there is time to do the checks.
Score momentum against the reference window. That is the rival's published demand window, or failing that, the 39-day community average as a proxy.
- 3: Last non-bot comment inside the window, and reactions at or past the repo's published bar.
- 2: Inside the window, with reactions short of the bar.
- 1: Outside the window, but reactions have built up from people other than the filer.
- 0: Silent past the window, with no reactions beyond the filer's own.
How to tie the filer to a buying account
A thread becomes an account signal only when the filer's employer is confirmed by two sources, with the second dated within 90 days, and the filer is neither rival staff nor a consultant filing for someone else. Same-org corroboration then tells you whether the pain belongs to one person or to a team.
Dimension 4: filer employer resolution
First, check author_association. A MEMBER or COLLABORATOR filer belongs to the rival. One ranking guide advises adding competitor company domains to a suppression list so their employees never appear as leads. Do that too, but author_association is the more direct check because it is printed on the issue itself.
Next, take the company field from the filer's profile and confirm it against a second source. Doing this by hand means cross-checking public professional records one filer at a time. A plain-English search in Refolk for the role, the skill, and the company returns matching profiles from public professional records and the public GitHub graph, which gives you that second source in one pass.
Employer strings in technical pools are messier than they look. In Refolk's index of professional profiles, small samples from three infrastructure pools show the problem:
| Pool | Profiles | Founder/CXO in sample | Ambiguous employer in sample |
|---|---|---|---|
| Terraform, US | 88,530 | 7/25 (28%) | 1 acquisition string |
| Pulumi, US | 950 | 7/25 (28%) | 2/25 rival's own staff |
| Terraform, Germany | 14,350 | 11/25 (44%) | 3/25 consultant/self/fractional, 1 "Not disclosed" |
Each sample is 25 profiles, so treat the percentages as indicative only.
What it proves: a named, current employer you can contact at account level. How it lies: an employer string like "Plural acquired by SAI360" (a real one from the US Terraform sample) usually means nobody has updated the field since the acquisition. A consultant may be filing on behalf of a client they cannot name.
- 3: Confirmed by two sources, the second dated within 90 days, at a named company that is not a consultancy.
- 2: One credible source only.
- 1: Consultant, self-employed, or fractional title, or an out-of-date "acquired by" string.
- 0: Unresolved, or MEMBER/COLLABORATOR (rival staff).
How many issue filers can be tied to an employer, and how often that employer turns out wrong, is not publicly established. Refolk's index cannot measure issue filers directly. Track your own resolution rate across your first batch of threads, and use that as your local base rate.
Dimension 5: same-org corroboration
Check whether anyone else from the filer's company has filed, commented on, or reacted to issues in this repo or neighbouring repos. One person's complaint is a data point. Three people from the same company is a team problem.
The size of the pool sets the base rate. In Refolk's index, 88,530 US profiles list Terraform and 950 list Pulumi, so Terraform outnumbers Pulumi 93.2 to 1. The US Terraform pool is also 6.2 times the German one. A single Pulumi thread from an account you can resolve is rarer than a Terraform one, and worth more. To check whether a resolved company has other users of the rival tool, ask Refolk for engineers at that company who list it. The count you get back shows whether the filer is alone.
- 3: More than one other person from the company filing or commenting.
- 2: One other person filing or commenting.
- 1: One other person reacting only.
- 0: Nobody else found.
How to score pitch fit and reach a verdict
Pitch fit asks whether your product solves this exact pain without waiting for the rival's fix. A "no" ends the read at Pass, whatever the other five scores say. After that, apply the verdict rules in order.
Dimension 6: pitch fit
- Yes: your product removes this pain today, with no change needed upstream.
- Partial: it removes part of the pain, or an adjacent one.
- No: it does not address this pain.
How it lies: answering "yes" because the thread is in your category when the specific failure described is one your product shares. Read the actual error or workflow, not the title.
Verdict rules, applied in order
- Pass if pitch fit is no, or pain severity is 0.
- Brief Only if the filer is MEMBER or COLLABORATOR. The pain still counts for the battlecard, but there is no account to approach.
- Brief Only if employer resolution is 2 or lower, if the window is 0, or if pitch fit is partial.
- Pitch if the window is 2 or higher, employer resolution is 3, pain is 2 or higher, momentum is 1 or higher, pitch fit is yes, and corroboration is 1 or higher or the filer holds a founder or CXO title at the account.
- Time It if employer resolution is 3, pain is 2 or higher, pitch fit is yes, and either the window is 1 or the Pitch rule failed only on momentum or corroboration.
- Brief Only in every other case.
Re-check rules by verdict
- Pitch: re-read the thread state on the day you send, because a linked PR can appear overnight. Refresh the employer source before it is more than 90 days old.
- Time It: re-check on the repo's demand-bar deadline (for VS Code, 60 days after the request entered Backlog Candidates) or at once if a linked PR appears. With no published bar, use the open date plus the median for the closest benchmark, labelled as a proxy.
- Brief Only: log the pain and thread link in the battlecard. Re-run the employer check when a second source turns up, or at 90 days.
- Pass: no date. Re-open only if the state_reason becomes reopened.
RIVAL THREAD READ Thread URL: Rival / repo: Read date: Reader (SDR / AE): 1. Window (maintainer state) score 0-3: __ state: open | closed state_reason: completed | not_planned | duplicate | reopened | null linked PR: yes | no milestone: ______ labels: ______ closed by bot vote rule? yes | no duplicate? canonical issue: ______ (re-run the read there) 2. Pain severity + workaround score 0-3: __ blocking | painful | cosmetic workaround in thread: yes | no 3. Recency + momentum score 0-3: __ opened: ____-__-__ last non-bot comment: ____-__-__ reactions: __ repo demand bar: ______ (or "none, proxy used") 4. Filer employer score 0-3: __ author_association: ______ (MEMBER / COLLABORATOR = Brief Only) source 1 (profile company field): ______ source 2 (dated within 90 days): ______ date: ____-__-__ founder/CXO title: yes | no consultant/fractional: yes | no 5. Same-org corroboration count: __ score 0-3: __ 6. Pitch fit yes | partial | no VERDICT: Pitch | Time It | Brief Only | Pass Re-check date: ____-__-__ trigger: ______ Proxy used for re-check: yes | no Contact route: work address found outside GitHub | none
One sheet per thread; paste it into your CRM note or tracker and fill every line before choosing a verdict.
Step-by-step: one rival thread to a verdict
Eight steps take a thread from a link to a verdict and a contact route. Going by the per-step estimates below, that is roughly 40 to 50 minutes of SDR and AE time.
The rival issue-thread read
- Read the lifecycle stateSDR, 5 minutes. Record state, state_reason, labels, milestone, and linked PR; class the thread as open/untriaged, scheduled, declined, fixed, or duplicate, and for a duplicate restart on the canonical issue.
- Score pain severity and workaroundAE, 5 to 10 minutes. Decide blocking, painful, or cosmetic, check the whole thread for a workaround, and score 0 to 3.
- Record recency and momentumSDR, 5 minutes. Note the last non-bot comment date and reaction count, and compare them with any published demand bar such as 20 upvotes in 60 days.
- Resolve the filer's employerSDR, 10 to 15 minutes. Exclude MEMBER and COLLABORATOR, then confirm the company field against a second source dated within 90 days, or mark it unresolved.
- Count same-org corroborationSDR, 10 minutes. Count other filers, commenters, or reactors from the same company in this repo and neighbouring repos.
- Judge pitch fitAE or founder, 5 minutes. Answer yes, partial, or no to whether your product solves this exact pain without the rival's fix.
- Set the verdict and re-check dateAE, 2 minutes. Apply the verdict rules in order and write a dated re-check trigger on the sheet.
- Choose the contact routeAE. For a Pitch, approach at account level through a work address found outside GitHub with an Article 14 notice and Article 21 opt-out; never post in the thread or use a commit email.
Sources disagree about speed. Vendor guides put a 24-hour outreach target ahead of steps 1 and 4. My view: issues filed by outsiders averaged 39 days to close in Kikas et al., so skipping the lifecycle and employer checks to save one day trades a small timing gain for a high chance of emailing rival staff, or pitching a pain that a merged PR has already fixed.
How the rival thread read goes wrong
Most bad reads come from trusting a default: a close reason that fell back to completed, a company field nobody updated, or an email address published for some other purpose. These seven failure modes cover nearly every false positive in this read.
- "Completed" read as fixed. The close reason defaults to completed, so a thread can look fixed when nothing shipped. Check: look for
linked:prand a release note. Without both, score the window 2, not 0. - "Not planned" read as rejected on merit. VS Code's bot closes anything that misses 20 votes in 60 days, so not_planned can simply mean under-voted. Check: read the closing comment. A bot closure leaves the window open but signals weak demand, so check corroboration before you pitch.
- Filer is rival staff. MEMBER and COLLABORATOR filers work for the rival, and their threads read like customer pain because they are often relaying it. Check: author_association first, then a company domain suppression list.
- Stale or acquired employer. A string like "acquired by" usually means the company field is out of date. Check: require a second source dated within 90 days before scoring employer resolution 3.
- Consultant filer. A consultant may be filing on behalf of a client they do not name. Check: count the pain toward the battlecard, but do not resolve it to the consultancy as an account.
- Hidden duplicate. A lone thread is often a duplicate of a canonical issue with ten times the reactions. Check: search the repo for the key error string, sorted by reactions, before scoring momentum.
- Contact through a commit email. This is the backlash pattern documented in public since 2015. The false reasoning is "they published it, so it's fine." Check: use only work addresses found outside GitHub.
A thread can also lie through its silence. A thread with no comments and no reactions may still describe a blocking pain that dozens of people read and never acknowledged. The read scores that as low momentum, not low pain. Brief Only is the honest verdict until someone else from the account shows up.
How to contact the account without touching the thread
Approach the account, not the thread. Use a work address found outside GitHub, rely on legitimate interest with an Article 14 notice and an Article 21 opt-out, and never post in the issue or email a commit address.
From a Pitch verdict to first contact
- Re-check the threadSame-day read confirms the state and no linked PR
- Pick the personDecision owner at the account, often the filer in founder-heavy pools
- Find a work addressSourced outside GitHub, never a commit email
- Write it by handName the public pain and your data source
- Honour objectionsAny opt-out ends all outreach to that person
The rules behind this:
- GitHub's Acceptable Use Policies define scraping as "extracting information from our Service via an automated process" and say that "Scraping does not refer to the collection of information through our API." Use of the data must still comply with the GitHub Privacy Statement. Earlier wording also barred using GitHub information for "selling GitHub users' personal information, such as to recruiters, headhunters, and job boards." I have not verified the current text of that section in full, so read it yourself before you build a pipeline on API data.
- GDPR: Recital 47 says processing for direct marketing may be regarded as carried out for a legitimate interest. You still owe the person an Article 14 notice saying where their data came from, and you must stop when they object under Article 21.
- ePrivacy: the rules for email are set nationally in each EU member state. Check the recipient's country before sending.
The backlash is documented and public. A developer described companies that "notice users contributing to repos in their field of business, send marketing emails to those users without receiving their consent", and the same thread traces similar incidents back years. Templates draw the sharpest objections. One commenter found it acceptable to be contacted, "but not by sending automatic/template emails." I found no primary source that measures how vendor comments in threads or automated issue filing are received. Their reception is unverified, and this framework rules both out regardless.
Subject: <the pain, in five words> at <Company> Hi <first name>, I'm <your name> at <your company>. A public issue in the <rival project> repo describes <the pain, in one plain sentence>, and it looks like the <team name> team at <Company> runs into it. <Your product> handles this today without waiting on an upstream fix: <one concrete sentence on how>. If it is a live problem for your team, would a short call be useful? If not, no reply needed. <your name> Why you are getting this: I saw the public GitHub issue and found your work address in public professional records. I am contacting you on the basis of legitimate interest. Reply "stop" and I will remove you from all outreach.
Rewrite every line in your own words for each account; mass-sent templated copy is the pattern developers complain about publicly.
Keeping thread reads current
A verdict is only as good as its last re-check. Put every Time It and Brief Only date in your calendar or CRM when you set it, and re-run the lifecycle step before acting on any thread older than its re-check date.
Three things move a verdict without telling you:
- A linked PR appearing. It turns an open window into a closing one. Saved searches using
linked:pron your watched repos catch this. - A reopened state. A Pass thread whose state_reason becomes reopened goes back to step 1.
- A duplicate closure. When your thread is closed as a duplicate, the live signal moves to the canonical issue, and your sheet should move with it.
Refresh your base rates as well. The pool sizes and employer-ambiguity rates above come from small samples in Refolk's index. Your own count of resolvable filers over your first batch of threads will tell you more about your market than any published figure, because none has been published.
Before you act on a verdict
- The thread's state and state_reason were re-read today, not copied from an earlier read.
- A "completed" closure was only scored as fixed when a linked PR and a release note exist.
- A "not planned" closure was checked for a bot vote-rule comment.
- Duplicates were followed to the canonical issue and scored there.
- The filer's author_association is neither MEMBER nor COLLABORATOR.
- The employer is confirmed by two sources, the second dated within 90 days.
- Consultant, fractional, and "acquired by" employers were not resolved to an account.
- Every Time It re-check date is labelled as a published repo bar or a proxy.
- The contact address was found outside GitHub and is not a commit email.
- The message names its data source and offers a one-word opt-out.
- Nothing was posted in the rival's thread.
Questions practitioners ask
Should I reach out to a developer who opened an issue on a competitor's repo?
Only once the thread clears the read: the window is still open, the filer's employer is confirmed by two sources, and your product fixes that exact pain. Even then, contact the account through a work address found outside GitHub, with a notice saying where you found their details and an easy way to opt out. Never reply in the thread and never use a commit email. Public complaints about that pattern go back to 2015.
Is a GitHub issue asking how to migrate off a competitor a purchase-ready lead?
No, not on its own. A migration question shows real pain and a stated wish to leave, so it usually scores high on severity. But the filer may work for the rival, consult for a client nobody names, or have a company field that is out of date. Rival maintainers may also already be fixing the cause. Run the lifecycle check and the employer check before you treat it as an account signal.
How fast do I need to act on a competitor issue before the intent goes cold?
Speed matters less than vendor tools claim. Some say issues stay high-intent for 7 to 14 days and that you should reach out within 24 hours. But in Kikas et al., community-filed issues averaged 39 days to close, against 5.9 days for issues the project team filed itself. The full read takes roughly 40 to 50 minutes. Do it first, then act on the verdict.
Can I comment in the issue thread to offer my product as an alternative?
No. This framework forbids it outright. I found no primary source that measures whether vendor comments in rival threads are welcome or resented, so the reception is not established. What is established is that developers object loudly in public to unsolicited, templated outreach built from their GitHub activity. A pitch in a rival's thread is permanent, public, and aimed at one person, so the possible damage far outweighs any gain.
Is it legal to email someone I found through a GitHub issue?
It can be, but it depends on the route you take. Under GDPR, Recital 47 says direct marketing may count as a legitimate interest. You still have to give an Article 14 notice saying where the data came from, and you have to stop if the person objects under Article 21. The rules for email itself are set by each EU member state, so check the recipient's country. GitHub's policies also govern how you collect the data.
What does 'closed as not planned' mean for my pitch?
Usually it means the window is open: the rival does not plan to fix the pain, so your product is the fix. Read the closing comment first. Some repos close requests automatically when they miss a vote target. VS Code closes requests that get fewer than 20 upvotes in 60 days. A bot closure like that means the window is open but few people backed the request, so check for other people at the same company before you pitch.
Try it on the search you came here for
Stop building boolean strings. Just describe the person.
Type one sentence. I plan the search, read GitHub, public LinkedIn and Crunchbase records, and the open web as it is right now, and hand back a ranked list with the reason next to every name.
01Describe them
One plain sentence. Role, city, stack, stage, whatever matters to you.
02I read the web live
GitHub, public LinkedIn and Crunchbase records, the open web. Not a database that went stale last quarter.
03You read the shortlist
Ranked, with the reasoning under every name. Open a profile, ask a follow-up, narrow it down.
- Staff backend engineers in NYC who shipped Rust in production
- Series A fintechs in SF under 50 people, growing headcount this year
- Maintainers of fast-growing Rust web frameworks on GitHub
- No boolean, no filters, no seat to buy. One box.
- Read at search time, so a profile updated yesterday counts today.
- Every step visible as it runs, every name with its reason.
500 free credits on sign-up. No card, no demo call. See real searches.