ProgrammableMarketing/convex.dev (opens in a new tab)
PDF

Views

  • Plan
  • Roadmap
  • Research

On this page

  • The brief
  • Start here
  • Idea portfolio
  • Effort vs impact
  • This week · 6
  • This quarter · 9
  • Big bets · 3
  • Wildcards · 4
  • How we can help
Generated Oct 18, 2026
  1. Programmable
  2. Playbooks
  3. Convex
Convex logo

growth playbook · Oct 18, 2026

convex.dev (opens in a new tab)

Convex growth playbook

Diagnosis

Convex is hiring its first Head of Marketing, a Product Marketer, a GTM Engineer, a Head of Developer Experience and a Creative Director, after its own team wrote that paid ads, newsletters, creators and an AEO agency did not work and the startup program did (that cohort grows spend 30x over 24 months).

The hiring says Convex wants a small, high-taste marketing function built on trust, momentum and recommendations, plus product-usage signals instead of outbound. It has strong raw material: a striking study of agent-built apps on Postgres versus Convex, a million weekly npm downloads, 181 components and a startup program that works.

This playbook avoids paid acquisition and focuses on telling people what ships, acting on usage signals with real help, scaling the startup program through people founders trust, and becoming the backend agents can use best.

Growth Score

8/10

Biggest Opportunity

Grow the startup program, the one channel that already works, through trusted recommendations, partners and dinners, while acting on product signals with genuinely useful help.

Biggest Risk

Choosing a new backend feels risky to founders and CTOs, and Supabase, Firebase and plain Postgres are the safe defaults; without trust signals, growth depends on the founders' personal reach.

01

Start here

The three moves to make first. Each one is written up in full below.

  1. 1

    Give the "agents corrupt Postgres" research its own home on convex.dev and its own launch

    Check the repo and load tests can be run by an outsider in under an hour.

    High impact First result in 2 weeks
  2. 2

    A weekly "what shipped" post and email, drafted from the changelog and edited by a person

    Write this week's post from the last seven days of releases.

    High impact First result in 1 week
  3. 3

    Define "getting serious" in product data and send those teams a note from an engineer

    Query last month's teams that hit three of the five signals.

    High impact First result in 2 weeks

02

Your idea portfolio

22 ideas, grouped by when to run them. Open any idea for how to run it: the steps, who to approach, copy you can use, and how to measure it.

  • 7 ideas are marked Unexpected: plays you are unlikely to have tried.

Every idea placed by the effort it takes against the impact it could have. Quick wins sit top left; select a number to open that idea. The filters below narrow the list by status, timing, category and what Programmable can build.

Effort vs impactEach number is an idea. Select one to open it.
Impact →
Quick wins · 4Big bets · 9Fill-ins · 5Reconsider · 4
Effort →

Showing 22 of 22 ideas

This week 6

Quick wins with a clear first action.

  • 01

    The strongest argument Convex has made this year, that nine out of ten agent-built Postgres apps silently corrupted data and none on Convex did, is served from a subdomain named after a different product. The Product Marketer is hired because Convex ships but does not tell people.

    High impactLow effort First result in 2 weeks Product Marketer

    First action

    Check the repo and load tests can be run by an outsider in under an hour.

    How to measure it

    Signups attributed to the research page

    Now
    –
    Target
    1,000
    By
    60 days

    Steps

    1. 1

      Check reproducibility

      An engineer outside the project runs the tests from the README.

      Software Engineer, AI 1 day

    2. 2

      Move the page

      A stable convex.dev URL with a redirect from the subdomain.

      Creative Director / Head of Design 1 day

    3. 3

      Write the Stack post

      Method, results, limits and what it does not prove.

      Product Marketer 2 days

    4. 4

      Link it in

      Homepage, compare pages and transaction docs.

      Head of Developer Experience 2 hours

    5. 5

      Launch

      Founder thread, Hacker News by the author, Discord announcement.

      Product Marketer 1 day

    6. 6

      Answer critics

      Engineers reply to every serious challenge in public.

      Software Engineer, AI 1 week

    7. 7

      Measure

      Signups, links and discussion threads.

      Product Marketer 1 hour

    First result
    2 weeks
    Cost
    $0
    Owner
    Product Marketer

    Smallest test

    Repost the existing page with a Stack write-up and measure the difference in signups.

    Why it works

    Move the research to convex.dev/research/agents-break-postgres, link it from the homepage, the Supabase and SQL compare pages and the docs on transactions. Launch it properly: a Stack post with the method and repo, a short founder thread, a Hacker News submission written by the engineer who ran it, and the raw data downloadable.

    It is substance, not hype: reproducible numbers about a failure developers fear, presented the way skeptical engineers trust. That is exactly the bar the Head of Marketing brief sets.

    Risks and how to handle them

    • Postgres experts call it unfair

      Publish the prompts and invite them to rerun with their own guidance.

    Evidence

    • Your websiteResearch page "The backend platform your agents can't break" (served at chef.convex.dev): frontier coding agents one-shot the same ten apps on Postgres and on Convex; "9 out of 10 apps silently corrupted data on Postgres. None of them did on Convex." A shopping-cart app on Postgres falls to 28% correct updates at 256 concurrent shoppers while Convex holds at 100%.Source
    • HiringProduct Marketer: "We ship frequently, but aren't great at talking about it"; "Once developers find us they tend to stick around"; "we'd sooner publish a 'how to give Convex less money' post than pretend what we do is magic"; owns launches end to end and writes the post and cuts the assets; "connective tissue between engineering, developer experience, design, and creators".Source
    • HiringHead of Marketing: "At a company our stage the founders (James and Jamie) are the de facto CMO(s)"; own positioning, launches, content, community, events and brand; "Decide what we don't do. Pick the two or three motions that matter most and go deep"; "nothing hypey, nothing generic, nothing we'd be embarrassed to put in front of a skeptical engineer"; "Ship a bunch of cool marketing nobody would believe came from a database company."Source
  • 02

    The Product Marketer brief opens with "We ship frequently, but aren't great at talking about it" and asks for talking about the work "consistently, not in occasional bursts". Developers who find Convex stay, so being heard is most of the job.

    High impactLow effort First result in 1 week Product Marketer

    First action

    Write this week's post from the last seven days of releases.

    How to measure it

    Weekly email open rate

    Now
    –
    Target
    45%
    By
    60 days

    Steps

    1. 1

      Collect the source

      Changelog entries and merged release notes for the week.

      Product Marketer 1 hour

    2. 2

      Draft automatically

      A script that groups changes into a first draft.

      GTM Engineer 1 day

    3. 3

      Edit for voice

      Cut hype; keep code and plain reasons.

      Product Marketer 2 hours a week

    4. 4

      Publish and send

      Stack post plus email to active developers.

      Product Marketer 1 hour a week

    5. 5

      Cut one clip

      A 30-second screen recording of the best change for X.

      Creative Director / Head of Design 1 hour a week

    6. 6

      Track

      Opens, clicks and reactivated projects.

      GTM Engineer Weekly

    First result
    1 week
    Cost
    $0
    Owner
    Product Marketer

    Smallest test

    Send one issue to 1,000 active developers and compare activity with a control group.

    Why it works

    Every Friday, a short Stack post and email to active developers: what shipped, why it matters, one code snippet each, and a link to docs. A draft is generated from merged release notes; the Product Marketer edits for voice and cuts anything that is not interesting.

    Momentum is one of the three beliefs the founders say Convex needs ("Convex is alive"), and a steady cadence shows it without a single ad.

    Risks and how to handle them

    • Becomes filler

      Skip a week rather than send a thin one.

    Evidence

    • HiringProduct Marketer: "We ship frequently, but aren't great at talking about it"; "Once developers find us they tend to stick around"; "we'd sooner publish a 'how to give Convex less money' post than pretend what we do is magic"; owns launches end to end and writes the post and cuts the assets; "connective tissue between engineering, developer experience, design, and creators".Source
    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • programmable.infoSite scan: 502 URLs including 181 component pages, 116 TypeScript guide pages, 62 event pages, 36 templates, 24 Databased podcast episodes, 23 "can do" pages, 6 hackathons, 5 compare pages (SQL, MongoDB, Appwrite, Firebase, Supabase), 5 evals pages and a single case study (Conduit). npm package "convex" about 1 million weekly downloads; 23,800 GitHub stars across the org (convex-backend 12,700); YouTube 13,800 subscribers, 416 videos, 1.4 million views; 63 Hacker News mentions. Named customers include OpenAI, Tripadvisor, a16z, Zapier and Reducto.Source
  • 03

    The GTM Engineer brief: thousands sign up weekly, and the job is to notice the moment a team crosses a threshold, without slop. Before any model, a few hand-picked thresholds catch most of the serious ones.

    High impactLow effort First result in 2 weeks GTM Engineer

    First action

    Query last month's teams that hit three of the five signals.

    How to measure it

    Professional upgrades from noted teams

    Now
    –
    Target
    25% within 60 days
    By
    60 days

    Steps

    1. 1

      Pick signals

      Five signals that separate tire-kickers from builders.

      GTM Engineer 2 hours

    2. 2

      Query them

      A daily query over deployment data.

      GTM Engineer 1 day

    3. 3

      Post to a channel

      One card per team with what they are building.

      GTM Engineer 1 day

    4. 4

      Write notes

      Rotating engineers write two lines and one tip.

      Software Engineer, AI 20 minutes a day

    5. 5

      Log replies

      Responses and conversations in the CRM.

      GTM Engineer Ongoing

    6. 6

      Compare cohorts

      Upgrade and retention of noted versus not noted teams.

      GTM Engineer 4 weeks

    First result
    2 weeks
    Cost
    $0
    Owner
    GTM Engineer

    Smallest test

    Write notes by hand to the 20 most active new teams this week.

    Why it works

    Pick five signals (first production deploy, custom domain, a second developer invited, auth configured, function calls past a threshold) and when a team hits three, post it in a channel where an engineer sends a short, specific note: what they noticed, one tip for their stack, and an offer to help, with no meeting link.

    Help from an engineer at the right moment is "technical, specific, and earned", the only kind of outreach the brief allows, and it starts relationships with the startups that compound 30x.

    Risks and how to handle them

    • Feels like surveillance

      Mention only what the team shipped publicly or in the dashboard, and offer help, not a call.

    Evidence

    • HiringGTM Engineer, the first: "Every week thousands of developers sign up for Convex"; "model which deploys are getting serious, surface intent the moment a team crosses a threshold, and route it to the right play or the right human"; "agent + human-in-the-loop outbound that's genuinely useful"; "We are deeply allergic to slop"; "the most valuable signal lives in product usage, not a lead form".Source
    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • Your websitePricing: Free & Starter ($0 and pay as you go, 1-6 developers), Professional at $25 per developer a month (log streaming, exception reporting, HIPAA/SOC 2 reports, daily backups, custom domains), and Business & Enterprise with a $2,500 monthly minimum. Startups can get up to a year of Professional free; the program lists YC and NEO companies and investors such as OpenAI Fund, Index Ventures and Accel.Source
  • 04

    The startup program is the one thing that worked: the cohort compounds spend 30x over 24 months, and founders join because people they trust recommend Convex. Nothing makes that recommendation easy to act on.

    High impactLow effort First result in 1 quarter Head of GTM

    First action

    Email five current program founders and ask who they would refer.

    How to measure it

    Accepted referred startups

    Now
    –
    Target
    30
    By
    1 quarter

    Steps

    1. 1

      Set terms

      Credits for both sides, quality bar unchanged.

      Head of GTM 1 day

    2. 2

      Build links

      Personal links tied to the referrer.

      GTM Engineer 1 day

    3. 3

      Tell founders

      A personal note from the CEO to program founders.

      Co-Founder / CEO 2 hours

    4. 4

      Fast-track

      Referred applications reviewed in 48 hours.

      Head of GTM Ongoing

    5. 5

      Thank referrers

      Credits and an invitation to the next dinner.

      Head of GTM Ongoing

    6. 6

      Measure

      Accepted referrals and their spend after 6 months.

      GTM Engineer Monthly

    First result
    1 quarter
    Cost
    Credits
    Owner
    Head of GTM

    Smallest test

    Ask ten founders for one referral each this week.

    Why it works

    Give every startup program founder a personal link: the friend gets an extra six months of Professional on acceptance, the referrer gets added credits and a dinner invitation. Applications through a link skip the queue but still meet the bar.

    It amplifies the exact channel the founders say works (recommendations from trusted builders) and keeps the program selective.

    Risks and how to handle them

    • Quality drops

      Same review bar for referred startups.

    Evidence

    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • Your websitePricing: Free & Starter ($0 and pay as you go, 1-6 developers), Professional at $25 per developer a month (log streaming, exception reporting, HIPAA/SOC 2 reports, daily backups, custom domains), and Business & Enterprise with a $2,500 monthly minimum. Startups can get up to a year of Professional free; the program lists YC and NEO companies and investors such as OpenAI Fund, Index Ventures and Accel.Source
    • Your websiteStack blog: ClawHub, the skill repository for the OpenClaw project, scaled "from 0 to 1M visitors a week in just over a month" on Convex; a new "Convex Explained" home for interactive explainers; posts on the Claude Code and Codex plugins, competitive benchmarks and incident response that builds trust; a post about planning a conference.Source
  • 05

    The site has one case study (Conduit) even though customers include OpenAI, Tripadvisor and Zapier, and the ClawHub post shows the format that works: real traffic, real problems, how it was fixed.

    Medium impactLow effort First result in 1 month Head of Developer Experience

    First action

    Pick three program companies with a recent scaling moment.

    How to measure it

    Projects started after reading a story

    Now
    –
    Target
    300
    By
    1 quarter

    Steps

    1. 1

      Pick companies

      Three program companies with real scale.

      Head of GTM 1 day

    2. 2

      Interview engineers

      One hour each about one hard problem.

      Head of Developer Experience 1 week

    3. 3

      Write

      Engineering-first, with numbers and code.

      Head of Developer Experience 1 week

    4. 4

      Approve

      Founders approve the text.

      Head of GTM 1 week

    5. 5

      Publish

      A case studies section on convex.dev and Stack.

      Creative Director / Head of Design 1 day

    6. 6

      Share

      Founders and engineers share from their accounts.

      Product Marketer 1 day

    First result
    1 month
    Cost
    $0
    Owner
    Head of Developer Experience

    Smallest test

    Write one story from the ClawHub post in the new format.

    Why it works

    Interview three startup program companies about a scaling moment, write each as an engineering story (the numbers, the design choices, what broke) and publish under a "built on Convex" section with the founders' approval.

    Trust is the first belief Convex needs, and stories from builders others respect are the "expensive and hard to fake" signal the founders describe.

    Risks and how to handle them

    • Customers decline

      Offer anonymous technical write-ups.

    Evidence

    • programmable.infoSite scan: 502 URLs including 181 component pages, 116 TypeScript guide pages, 62 event pages, 36 templates, 24 Databased podcast episodes, 23 "can do" pages, 6 hackathons, 5 compare pages (SQL, MongoDB, Appwrite, Firebase, Supabase), 5 evals pages and a single case study (Conduit). npm package "convex" about 1 million weekly downloads; 23,800 GitHub stars across the org (convex-backend 12,700); YouTube 13,800 subscribers, 416 videos, 1.4 million views; 63 Hacker News mentions. Named customers include OpenAI, Tripadvisor, a16z, Zapier and Reducto.Source
    • Your websiteStack blog: ClawHub, the skill repository for the OpenClaw project, scaled "from 0 to 1M visitors a week in just over a month" on Convex; a new "Convex Explained" home for interactive explainers; posts on the Claude Code and Codex plugins, competitive benchmarks and incident response that builds trust; a post about planning a conference.Source
    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
  • 06

    The homepage leads with "Copy the setup prompt for your agent". When an agent runs the setup, Convex cannot easily tell whether it came from the homepage, a doc page or a friend's README, or where it stalled.

    Medium impactLow effort First result in 2 weeks GTM Engineer

    First action

    Add the source tag to the homepage prompt.

    How to measure it

    Share of agent setups reaching first deploy

    Now
    –
    Target
    +10 points
    By
    60 days

    Steps

    1. 1

      Tag the prompts

      A source value in each copy of the setup prompt.

      GTM Engineer 2 hours

    2. 2

      Log on creation

      Store the source when the project is created.

      Software Engineer, AI 1 day

    3. 3

      Build the funnel

      Copies, installs, first deploy, production deploy.

      GTM Engineer 1 day

    4. 4

      Find the drop

      Where agents fail most, using error logs.

      Head of Developer Experience 2 days

    5. 5

      Fix it

      Change the prompt or docs at the biggest drop.

      Head of Developer Experience 3 days

    6. 6

      Repeat monthly

      Review the funnel each month.

      GTM Engineer Monthly

    First result
    2 weeks
    Cost
    $0
    Owner
    GTM Engineer

    Smallest test

    Tag the homepage prompt alone and watch for one week.

    Why it works

    Add a source tag to the commands in each copy of the setup prompt (homepage, docs, templates), log it on project creation, and build a simple funnel: prompt copies, ai-files installs, first deploy, production deploy. Fix the biggest drop first.

    More of Convex's developers are now agents. Knowing where agent setups fail is the activation work that matters most.

    Risks and how to handle them

    • Tags removed by users

      Treat untagged setups as a separate source.

    Evidence

    • Your websiteHomepage: "The reactive backend platform that keeps up with you and your agents. Database, functions, workflow, sync, search, file storage, and more. All TypeScript, zero glue." The first call to action is "Copy the setup prompt for your agent", which installs Convex guidelines and plugins for Claude Code and Codex. Headline changed over the last year from "The backend platform that keeps your app in sync" to "All gas, no breakages".Source
    • HiringHead of Developer Experience: own how developers "and agents become great at building on Convex"; "increasingly the developer learning it is an agent rather than a person"; owns docs.convex.dev and its information architecture and the blog at stack.convex.dev; builds documentation as a function.Source
    • HiringGTM Engineer, the first: "Every week thousands of developers sign up for Convex"; "model which deploys are getting serious, surface intent the moment a team crosses a threshold, and route it to the right play or the right human"; "agent + human-in-the-loop outbound that's genuinely useful"; "We are deeply allergic to slop"; "the most valuable signal lives in product usage, not a lead form".Source

This quarter 9

The core plays for the next three months.

  • 07

    The founders are tripling down on "high-quality in-person events" and say trust comes from knowing them. A dinner where founders meet James, Jamie and program founders is the most direct form of that.

    High impactMedium effort First result in 1 quarter Head of Marketing

    First action

    Book the first dinner with four program founders as hosts.

    How to measure it

    Program applications from dinner guests

    Now
    –
    Target
    40%
    By
    1 quarter

    Steps

    1. 1

      Set the format

      Guest mix, run of show and follow-up.

      Head of Marketing 1 week

    2. 2

      Recruit hosts

      Four program founders per dinner.

      Head of GTM 2 weeks

    3. 3

      Build the guest list

      Seed and Series A founders from investor partners and referrals.

      Head of GTM 2 weeks

    4. 4

      Run it

      Monthly in SF, every other month in New York.

      Head of Marketing Monthly

    5. 5

      Follow up

      Personal note and program invitation within two days.

      Head of GTM 2 days each

    6. 6

      Measure

      Applications and spend from attendees.

      GTM Engineer Quarterly

    First result
    1 quarter
    Cost
    $3,000 per dinner
    Owner
    Head of Marketing

    Smallest test

    Host one dinner in San Francisco this month.

    Why it works

    Twelve to sixteen people each month: four startup program founders, the rest founders at seed and Series A who are choosing a backend. No pitch; one founder tells a scaling story. Follow up with a startup program invitation.

    Choosing infrastructure is a trust decision, and founders trust peers and the people who will run their backend.

    Risks and how to handle them

    • Feels exclusive

      Publish notes from each dinner on Stack.

    Evidence

    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • Your websitePricing: Free & Starter ($0 and pay as you go, 1-6 developers), Professional at $25 per developer a month (log streaming, exception reporting, HIPAA/SOC 2 reports, daily backups, custom domains), and Business & Enterprise with a $2,500 monthly minimum. Startups can get up to a year of Professional free; the program lists YC and NEO companies and investors such as OpenAI Fund, Index Ventures and Accel.Source
    • Your websiteEvents and community: the "All Gas" hackathon sponsored by OpenAI, Firecrawl and AgentMail, an OpenAI Build Week community meetup in San Francisco, a Founders Inc startup partnership, and a Convex Champions ambassador program; community runs on Discord.Source
  • 08

    The program already includes YC, NEO and a Founders Inc partnership, and lists investors such as OpenAI Fund, Index and Accel. Investor platform teams recommend tools to every portfolio company.

    High impactMedium effort First result in 2 quarters Head of GTM

    First action

    List the investor platform teams behind current program companies.

    How to measure it

    Startups from new partners

    Now
    –
    Target
    50
    By
    2 quarters

    Steps

    1. 1

      List partners

      Accelerators and investors behind current program companies.

      Head of GTM 1 week

    2. 2

      Package the perk

      Credits, office hours and workshop.

      Head of GTM 1 week

    3. 3

      Pitch ten

      Platform teams at investors and accelerators.

      Head of GTM 6 weeks

    4. 4

      Run workshops

      A scaling workshop per partner per quarter.

      Software Engineer, AI Quarterly

    5. 5

      Track

      Sign-ups and spend by partner.

      GTM Engineer Monthly

    6. 6

      Drop weak partners

      Keep the ones whose companies build.

      Head of GTM Quarterly

    First result
    2 quarters
    Cost
    Credits and engineer time
    Owner
    Head of GTM

    Smallest test

    Offer the perk to one investor platform team.

    Why it works

    Offer accelerators and venture platform teams a packaged perk (a year of Professional, an engineer office hour, priority support) and a quarterly scaling workshop for their portfolio.

    It puts Convex in front of founders at the moment they choose infrastructure, through people they already trust.

    Risks and how to handle them

    • Perk fatigue

      Lead with engineer access, not discounts.

    Evidence

    • Your websitePricing: Free & Starter ($0 and pay as you go, 1-6 developers), Professional at $25 per developer a month (log streaming, exception reporting, HIPAA/SOC 2 reports, daily backups, custom domains), and Business & Enterprise with a $2,500 monthly minimum. Startups can get up to a year of Professional free; the program lists YC and NEO companies and investors such as OpenAI Fund, Index Ventures and Accel.Source
    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • Your websiteEvents and community: the "All Gas" hackathon sponsored by OpenAI, Firecrawl and AgentMail, an OpenAI Build Week community meetup in San Francisco, a Founders Inc startup partnership, and a Convex Champions ambassador program; community runs on Discord.Source
  • 09

    Five compare pages cover SQL, MongoDB, Appwrite, Firebase and Supabase. Developers also arrive from Postgres with an ORM, serverless Postgres and other backends, and the best compare page Convex has is respectful rather than combative.

    Medium impactMedium effort First result in 1 quarter Head of Developer Experience

    First action

    List the previous stacks of the last 100 Professional teams.

    How to measure it

    Migrations started from the guides

    Now
    –
    Target
    200
    By
    2 quarters

    Steps

    1. 1

      Find the sources

      Ask recent teams what they left.

      Product Manager 1 week

    2. 2

      Pick four stacks

      The most common previous stacks.

      Head of Developer Experience 1 day

    3. 3

      Write guides

      Schema, queries, auth, a script and when not to switch.

      Head of Developer Experience 8 weeks

    4. 4

      Test with agents

      Can an agent follow each guide?

      Software Engineer, AI 1 week

    5. 5

      Publish

      Linked from compare pages.

      Creative Director / Head of Design 1 week

    6. 6

      Measure

      Migrations started per guide.

      GTM Engineer Monthly

    First result
    1 quarter
    Cost
    $0
    Owner
    Head of Developer Experience

    Smallest test

    Write the Prisma guide first.

    Why it works

    Migration guides from the most common stacks (Postgres with Prisma or Drizzle, serverless Postgres hosts, Firebase), each with a "when not to switch" section, a schema translation and a script. Keep the respectful tone of the Supabase page.

    Honesty is the brand, and a guide that says when not to switch is more persuasive to a skeptical engineer than a feature grid.

    Risks and how to handle them

    • Guides go stale

      Test them with agents every release.

    Evidence

    • programmable.infoSite scan: 502 URLs including 181 component pages, 116 TypeScript guide pages, 62 event pages, 36 templates, 24 Databased podcast episodes, 23 "can do" pages, 6 hackathons, 5 compare pages (SQL, MongoDB, Appwrite, Firebase, Supabase), 5 evals pages and a single case study (Conduit). npm package "convex" about 1 million weekly downloads; 23,800 GitHub stars across the org (convex-backend 12,700); YouTube 13,800 subscribers, 416 videos, 1.4 million views; 63 Hacker News mentions. Named customers include OpenAI, Tripadvisor, a16z, Zapier and Reducto.Source
    • Competitor researchConvex's compare page on Supabase: "an open-source alternative to Firebase" built on Postgres and other open-source projects, "a modern development experience similar to Convex"; Convex says it has "a lot of respect for Supabase". Compare pages also cover Firebase, Appwrite, MongoDB and SQL.Source
    • HiringHead of Developer Experience: own how developers "and agents become great at building on Convex"; "increasingly the developer learning it is an agent rather than a person"; owns docs.convex.dev and its information architecture and the blog at stack.convex.dev; builds documentation as a function.Source
  • 10

    The PM brief mentions customers spinning up hundreds of deployments, and the AI engineer brief wants Convex to be the best platform for building agentic apps. App builders create backends at a scale no single team does.

    High impactMedium effort First result in 2 quarters Product Manager

    First action

    Interview the three customers running the most deployments.

    How to measure it

    Projects created through platforms

    Now
    –
    Target
    +50%
    By
    2 quarters

    Steps

    1. 1

      Interview platforms

      Why they chose Convex and what they need.

      Product Manager 2 weeks

    2. 2

      Package the offer

      APIs, pricing and support.

      Product Manager 4 weeks

    3. 3

      Publish the page

      Convex for platforms.

      Product Marketer 1 week

    4. 4

      Add the badge

      Optional "backend by Convex" link for generated apps.

      Creative Director / Head of Design 2 weeks

    5. 5

      Sign two more

      New platform partners.

      Head of GTM 8 weeks

    6. 6

      Track graduation

      Platform projects that become direct customers.

      GTM Engineer Monthly

    First result
    2 quarters
    Cost
    Engineering time
    Owner
    Product Manager

    Smallest test

    Write one platform story with the largest current user.

    Why it works

    A packaged offer for app-builder platforms: platform APIs for provisioning and billing, per-deployment pricing, a co-branded badge ("backend by Convex") and a shared support path. Publish a page and two platform stories.

    Each platform partner creates thousands of Convex projects, and every user who graduates to production becomes a direct customer.

    Risks and how to handle them

    • Low margin per deployment

      Price per active deployment and charge for production.

    Evidence

    • HiringCreative Director / Head of Design: the brand "has to be distinct, thoughtful, a little playful, and honest. The kind of thing a skeptical engineer trusts on sight"; works in Figma and the codebase alongside existing creative freelancers. Product Manager (the second PM): product areas include the AI gateway, platform APIs and the component ecosystem; calls with customers "to understand why their team is spinning up hundreds of deployments"; works with "our Heads of GTM and Marketing".Source
    • HiringSoftware Engineer, AI: "It's not sufficient for Convex to be the best platform for agentic development, it also has to be the best platform for building agentic applications"; works with the Head of AI Ergonomics and the CTO; "Strengthen the case that Convex is the best platform for agentic development, through demos, examples, and working directly with customers."Source
    • Your websiteResearch page "The backend platform your agents can't break" (served at chef.convex.dev): frontier coding agents one-shot the same ten apps on Postgres and on Convex; "9 out of 10 apps silently corrupted data on Postgres. None of them did on Convex." A shopping-cart app on Postgres falls to 28% correct updates at 256 concurrent shoppers while Convex holds at 100%.Source
  • 11

    There are 181 component pages, many by community authors. Components show the ecosystem is "alive", one of the three beliefs the founders want.

    Medium impactMedium effort First result in 1 quarter Head of Developer Experience

    First action

    Email the top 20 component authors.

    How to measure it

    New components published per month

    Now
    –
    Target
    20
    By
    2 quarters

    Steps

    1. 1

      Get usage data

      Installs per component.

      Product Manager 1 week

    2. 2

      Show stats

      Install counts on component pages.

      Software Engineer, AI 2 weeks

    3. 3

      Feature authors

      A monthly spotlight.

      Product Marketer Monthly

    4. 4

      Add a template

      A starter for new components.

      Head of Developer Experience 1 week

    5. 5

      Give early access

      Authors try new platform features first.

      Product Manager Ongoing

    6. 6

      Track

      New components and installs.

      Product Manager Monthly

    First result
    1 quarter
    Cost
    $0
    Owner
    Head of Developer Experience

    Smallest test

    Feature one author this month.

    Why it works

    Give component authors usage stats, a featured slot, a monthly author spotlight and early access to platform features. Add a "most used this week" list and a template for new components.

    Recognized authors build more and share their work, which shows momentum without paid promotion.

    Risks and how to handle them

    • Low-quality components

      Review before featuring.

    Evidence

    • programmable.infoSite scan: 502 URLs including 181 component pages, 116 TypeScript guide pages, 62 event pages, 36 templates, 24 Databased podcast episodes, 23 "can do" pages, 6 hackathons, 5 compare pages (SQL, MongoDB, Appwrite, Firebase, Supabase), 5 evals pages and a single case study (Conduit). npm package "convex" about 1 million weekly downloads; 23,800 GitHub stars across the org (convex-backend 12,700); YouTube 13,800 subscribers, 416 videos, 1.4 million views; 63 Hacker News mentions. Named customers include OpenAI, Tripadvisor, a16z, Zapier and Reducto.Source
    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • HiringCreative Director / Head of Design: the brand "has to be distinct, thoughtful, a little playful, and honest. The kind of thing a skeptical engineer trusts on sight"; works in Figma and the codebase alongside existing creative freelancers. Product Manager (the second PM): product areas include the AI gateway, platform APIs and the component ecosystem; calls with customers "to understand why their team is spinning up hundreds of deployments"; works with "our Heads of GTM and Marketing".Source
  • 12

    The Champions program exists and the founders want "builders whose judgment I trust enthusiastically recommend it". Champions are the people best placed to do that in their cities.

    Medium impactMedium effort First result in 1 quarter Head of Marketing

    First action

    Ask current Champions which cities they could host in.

    How to measure it

    Champion-hosted meetups

    Now
    –
    Target
    12
    By
    2 quarters

    Steps

    1. 1

      Write the kit

      Format, slides and swag.

      Head of Marketing 2 weeks

    2. 2

      Set a budget

      Per meetup budget.

      Head of Marketing 1 day

    3. 3

      Recruit hosts

      From current Champions.

      Head of Marketing 2 weeks

    4. 4

      Run meetups

      One per city per quarter.

      Head of Marketing Quarterly

    5. 5

      Track referrals

      A "who recommended us" field at signup.

      GTM Engineer 1 day

    6. 6

      Thank champions

      Credits and recognition.

      Head of Marketing Quarterly

    First result
    1 quarter
    Cost
    Meetup budgets
    Owner
    Head of Marketing

    Smallest test

    Sponsor one Champion meetup.

    Why it works

    Give Champions a meetup kit, a small budget, early access and a direct line to engineers. Ask for one meetup or one post a quarter. Track recommendations (signups citing a Champion).

    Local, trusted recommendations scale the founders' own trust beyond San Francisco.

    Risks and how to handle them

    • Uneven quality

      Engineers join the first meetup in each city.

    Evidence

    • Your websiteEvents and community: the "All Gas" hackathon sponsored by OpenAI, Firecrawl and AgentMail, an OpenAI Build Week community meetup in San Francisco, a Founders Inc startup partnership, and a Convex Champions ambassador program; community runs on Discord.Source
    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • programmable.infoSite scan: 502 URLs including 181 component pages, 116 TypeScript guide pages, 62 event pages, 36 templates, 24 Databased podcast episodes, 23 "can do" pages, 6 hackathons, 5 compare pages (SQL, MongoDB, Appwrite, Firebase, Supabase), 5 evals pages and a single case study (Conduit). npm package "convex" about 1 million weekly downloads; 23,800 GitHub stars across the org (convex-backend 12,700); YouTube 13,800 subscribers, 416 videos, 1.4 million views; 63 Hacker News mentions. Named customers include OpenAI, Tripadvisor, a16z, Zapier and Reducto.Source
  • 13

    The ClawHub story shows what Convex engineers can do for a team that suddenly scales. The GTM Engineer is asked for outbound that is genuinely useful; a short review of a team's own deployment is that.

    High impactMedium effort First result in 1 quarter GTM Engineer

    First action

    Pick five fast-growing teams and write reviews by hand.

    How to measure it

    Upgrades from reviewed teams

    Now
    –
    Target
    40%
    By
    1 quarter

    Steps

    1. 1

      Set the trigger

      Growth thresholds in function calls or bandwidth.

      GTM Engineer 1 day

    2. 2

      Build a review template

      The three checks engineers make first.

      Software Engineer, AI 2 days

    3. 3

      Assign engineers

      A weekly rotation.

      Software Engineer, AI Ongoing

    4. 4

      Send reviews

      Three fixes in a short note.

      Software Engineer, AI 30 minutes each

    5. 5

      Track

      Replies, upgrades and churn.

      GTM Engineer Monthly

    6. 6

      Automate checks

      Pre-fill the findings from usage data.

      GTM Engineer 4 weeks

    First result
    1 quarter
    Cost
    Engineer time
    Owner
    GTM Engineer

    Smallest test

    Send five reviews this week.

    Why it works

    When a team's function calls or bandwidth grow fast, an engineer reviews their dashboard data and sends three concrete fixes (indexes, query patterns, costs). Offer a call only if they reply.

    It is the opposite of a cold email: specific, earned, and it saves them money, which builds the trust the founders say Convex depends on.

    Risks and how to handle them

    • Too much engineer time

      Limit to teams above a growth threshold.

    Evidence

    • Your websiteStack blog: ClawHub, the skill repository for the OpenClaw project, scaled "from 0 to 1M visitors a week in just over a month" on Convex; a new "Convex Explained" home for interactive explainers; posts on the Claude Code and Codex plugins, competitive benchmarks and incident response that builds trust; a post about planning a conference.Source
    • HiringGTM Engineer, the first: "Every week thousands of developers sign up for Convex"; "model which deploys are getting serious, surface intent the moment a team crosses a threshold, and route it to the right play or the right human"; "agent + human-in-the-loop outbound that's genuinely useful"; "We are deeply allergic to slop"; "the most valuable signal lives in product usage, not a lead form".Source
    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
  • 14

    Convex already publishes coding evals with and without guidelines, and the Head of DevEx brief says the learner is increasingly an agent. Docs changes are not yet measured by whether agents get better.

    High impactMedium effort First result in 1 quarter Head of Developer Experience

    First action

    Run the current eval suite against last month's docs changes.

    How to measure it

    Agent eval pass rate

    Now
    –
    Target
    +10 points
    By
    2 quarters

    Steps

    1. 1

      Wire evals

      Run the suite on docs changes.

      Software Engineer, AI 2 weeks

    2. 2

      Set a bar

      Block changes that lower scores.

      Head of Developer Experience 1 day

    3. 3

      Publish scores

      A public chart over time.

      Creative Director / Head of Design 1 week

    4. 4

      Write it up

      Method on Stack.

      Head of Developer Experience 1 week

    5. 5

      Fix weak areas

      Docs for the lowest-scoring tasks.

      Head of Developer Experience 6 weeks

    6. 6

      Review

      Score trends monthly.

      Head of Developer Experience Monthly

    First result
    1 quarter
    Cost
    Compute costs
    Owner
    Head of Developer Experience

    Smallest test

    Run the suite on one docs change and compare scores.

    Why it works

    Run the evals on every docs and guidelines change, publish the score over time, and block changes that make agents worse. Share the method on Stack.

    It makes Convex measurably the easiest backend for agents, which is a claim nobody else can show.

    Risks and how to handle them

    • Evals overfit

      Rotate tasks each quarter.

    Evidence

    • HiringHead of Developer Experience: own how developers "and agents become great at building on Convex"; "increasingly the developer learning it is an agent rather than a person"; owns docs.convex.dev and its information architecture and the blog at stack.convex.dev; builds documentation as a function.Source
    • programmable.infoSite scan: 502 URLs including 181 component pages, 116 TypeScript guide pages, 62 event pages, 36 templates, 24 Databased podcast episodes, 23 "can do" pages, 6 hackathons, 5 compare pages (SQL, MongoDB, Appwrite, Firebase, Supabase), 5 evals pages and a single case study (Conduit). npm package "convex" about 1 million weekly downloads; 23,800 GitHub stars across the org (convex-backend 12,700); YouTube 13,800 subscribers, 416 videos, 1.4 million views; 63 Hacker News mentions. Named customers include OpenAI, Tripadvisor, a16z, Zapier and Reducto.Source
    • HiringSoftware Engineer, AI: "It's not sufficient for Convex to be the best platform for agentic development, it also has to be the best platform for building agentic applications"; works with the Head of AI Ergonomics and the CTO; "Strengthen the case that Convex is the best platform for agentic development, through demos, examples, and working directly with customers."Source
  • 15

    The founders wrote that conventional wisdom says not to "build your company on some new-fangled backend run by a tiny startup". The open source backend and self-hosting are the answer, but they are not presented as one.

    High impactMedium effort First result in 1 quarter Co Founder / CEO

    First action

    Draft the export and self-host steps and time them.

    How to measure it

    Business and Enterprise deals

    Now
    –
    Target
    +20%
    By
    2 quarters

    Steps

    1. 1

      Draft the plan

      Export, self-host and guarantees.

      Co-Founder / CEO 1 week

    2. 2

      Test it

      An engineer moves a real app off and back.

      Software Engineer, AI 1 week

    3. 3

      Publish

      A page and doc.

      Head of Developer Experience 1 week

    4. 4

      Add to sales

      A written version for enterprise.

      Head of GTM 1 day

    5. 5

      Announce

      A frank Stack post.

      Co-Founder / CEO 1 day

    6. 6

      Track

      Mentions in deals and objections.

      Head of GTM Quarterly

    First result
    1 quarter
    Cost
    $0
    Owner
    Co Founder / CEO

    Smallest test

    Write the plan as a Stack post first.

    Why it works

    A page and doc: how to export everything, how to run the open source backend yourself, what guarantees Convex makes, and how long migration would take. Give enterprise buyers a written version.

    Showing the way out lowers the risk of coming in, especially for CTOs at the Business and Enterprise tier.

    Risks and how to handle them

    • Encourages churn

      Churn is driven by product, not by documentation.

    Evidence

    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • Your websitePricing: Free & Starter ($0 and pay as you go, 1-6 developers), Professional at $25 per developer a month (log streaming, exception reporting, HIPAA/SOC 2 reports, daily backups, custom domains), and Business & Enterprise with a $2,500 monthly minimum. Startups can get up to a year of Professional free; the program lists YC and NEO companies and investors such as OpenAI Fund, Index Ventures and Accel.Source
    • programmable.infoSite scan: 502 URLs including 181 component pages, 116 TypeScript guide pages, 62 event pages, 36 templates, 24 Databased podcast episodes, 23 "can do" pages, 6 hackathons, 5 compare pages (SQL, MongoDB, Appwrite, Firebase, Supabase), 5 evals pages and a single case study (Conduit). npm package "convex" about 1 million weekly downloads; 23,800 GitHub stars across the org (convex-backend 12,700); YouTube 13,800 subscribers, 416 videos, 1.4 million views; 63 Hacker News mentions. Named customers include OpenAI, Tripadvisor, a16z, Zapier and Reducto.Source

Big bets 3

Larger plays that can change the trajectory.

  • 16

    A Stack post mentions planning a conference. The founders want trust, aliveness and recommendation; a conference can do all three if it is about builders.

    High impactHigh effort First result in 3 quarters Head of Marketing

    First action

    Pick the date and book five customer speakers.

    How to measure it

    Program applications from the conference

    Now
    –
    Target
    150
    By
    3 quarters

    Steps

    1. 1

      Set the date

      Venue and capacity.

      Head of Marketing 4 weeks

    2. 2

      Book speakers

      Program founders and community authors.

      Head of Marketing 8 weeks

    3. 3

      Run a hackathon

      The night before.

      Head of Marketing 1 night

    4. 4

      Hold office hours

      Engineers on site.

      Software Engineer, AI Event day

    5. 5

      Record it

      All talks for Stack.

      Creative Director / Head of Design Event day

    6. 6

      Follow up

      Program invitations and recordings.

      Head of GTM 2 weeks

    First result
    3 quarters
    Cost
    Event costs
    Owner
    Head of Marketing

    Smallest test

    Run a half-day event for 100 people first.

    Why it works

    One day in San Francisco: talks by startup program founders about real scaling, a hackathon the night before, live office hours with engineers and an open roadmap session with the founders. Record everything for Stack.

    It concentrates the in-person bet into one event that produces months of content.

    Risks and how to handle them

    • Low turnout

      Keep it small and invite-led.

    Evidence

    • Your websiteStack blog: ClawHub, the skill repository for the OpenClaw project, scaled "from 0 to 1M visitors a week in just over a month" on Convex; a new "Convex Explained" home for interactive explainers; posts on the Claude Code and Codex plugins, competitive benchmarks and incident response that builds trust; a post about planning a conference.Source
    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • Your websiteEvents and community: the "All Gas" hackathon sponsored by OpenAI, Firecrawl and AgentMail, an OpenAI Build Week community meetup in San Francisco, a Founders Inc startup partnership, and a Convex Champions ambassador program; community runs on Discord.Source
  • 17

    The homepage already speaks to agents first, and the AI engineer brief wants Convex to be the best platform for agentic development. Agents still stop at sign-up and billing, which need a person.

    High impactHigh effort First result in 3 quarters Product Manager

    First action

    Map the steps where an agent setup needs a human today.

    How to measure it

    Projects provisioned through the agent flow

    Now
    –
    Target
    20% of new projects
    By
    3 quarters

    Steps

    1. 1

      Map the steps

      Sign-up, project creation and billing.

      Product Manager 2 weeks

    2. 2

      Design consent

      Human approval for each paid step.

      Product Manager 3 weeks

    3. 3

      Build the flow

      APIs and CLI commands for agents.

      Software Engineer, AI 2 quarters

    4. 4

      Document

      Agent guidelines for the flow.

      Head of Developer Experience 2 weeks

    5. 5

      Launch

      With the Claude Code and Codex plugins.

      Product Marketer 2 weeks

    6. 6

      Measure

      Agent-provisioned projects and upgrades.

      GTM Engineer Monthly

    First result
    3 quarters
    Cost
    Engineering capacity
    Owner
    Product Manager

    Smallest test

    Let the plugin create a project with one human confirmation.

    Why it works

    An agent-facing flow: an agent can create a project for its user, choose a plan with the user's approval, and request an upgrade when limits are hit, with every step confirmed by the human.

    If agents choose backends, the backend that agents can provision fastest wins the default.

    Risks and how to handle them

    • Abuse and runaway costs

      Hard spending limits and human approval.

    Evidence

    • Your websiteHomepage: "The reactive backend platform that keeps up with you and your agents. Database, functions, workflow, sync, search, file storage, and more. All TypeScript, zero glue." The first call to action is "Copy the setup prompt for your agent", which installs Convex guidelines and plugins for Claude Code and Codex. Headline changed over the last year from "The backend platform that keeps your app in sync" to "All gas, no breakages".Source
    • HiringSoftware Engineer, AI: "It's not sufficient for Convex to be the best platform for agentic development, it also has to be the best platform for building agentic applications"; works with the Head of AI Ergonomics and the CTO; "Strengthen the case that Convex is the best platform for agentic development, through demos, examples, and working directly with customers."Source
    • HiringHead of Developer Experience: own how developers "and agents become great at building on Convex"; "increasingly the developer learning it is an agent rather than a person"; owns docs.convex.dev and its information architecture and the blog at stack.convex.dev; builds documentation as a function.Source
  • 18

    Convex's own study is persuasive but self-run. An independent benchmark, run with a university lab or a neutral group and open to every backend, would become the reference everyone cites.

    High impactHigh effort First result in 3 quarters Software Engineer

    First action

    Approach one university systems lab.

    How to measure it

    Backends participating

    Now
    –
    Target
    6
    By
    3 quarters

    Steps

    1. 1

      Find a partner

      A neutral lab or group.

      Software Engineer, AI 6 weeks

    2. 2

      Open the suite

      Release the apps and load tests.

      Software Engineer, AI 4 weeks

    3. 3

      Invite backends

      Competitors on the advisory board.

      Co-Founder / CEO 8 weeks

    4. 4

      Run quarterly

      Published by the partner.

      Software Engineer, AI Quarterly

    5. 5

      Write up results

      Stack and press.

      Product Marketer Quarterly

    6. 6

      Improve Convex

      Fix whatever the benchmark exposes.

      Product Manager Ongoing

    First result
    3 quarters
    Cost
    Funding for the partner
    Owner
    Software Engineer, AI

    Smallest test

    Release the current suite and invite outside runs.

    Why it works

    Fund and open-source a benchmark suite: agent-built apps on many backends, load tests for correctness, published quarterly by a neutral group with an advisory board that includes competitors.

    If Convex is right, an independent result is far more credible, and leading the standard keeps Convex central to the conversation.

    Risks and how to handle them

    • Convex loses a category

      Publish it anyway and fix it; credibility is the point.

    Evidence

    • Your websiteResearch page "The backend platform your agents can't break" (served at chef.convex.dev): frontier coding agents one-shot the same ten apps on Postgres and on Convex; "9 out of 10 apps silently corrupted data on Postgres. None of them did on Convex." A shopping-cart app on Postgres falls to 28% correct updates at 256 concurrent shoppers while Convex holds at 100%.Source
    • Your websiteStack blog: ClawHub, the skill repository for the OpenClaw project, scaled "from 0 to 1M visitors a week in just over a month" on Convex; a new "Convex Explained" home for interactive explainers; posts on the Claude Code and Codex plugins, competitive benchmarks and incident response that builds trust; a post about planning a conference.Source
    • Competitor researchConvex's compare page on Supabase: "an open-source alternative to Firebase" built on Postgres and other open-source projects, "a modern development experience similar to Convex"; Convex says it has "a lot of respect for Supabase". Compare pages also cover Firebase, Appwrite, MongoDB and SQL.Source

Wildcards 4

Unexpected experiments, each with a cheap way to test it.

  • 19

    The Product Marketer brief jokes Convex would sooner publish a "how to give Convex less money" post than pretend. That post would be the most trusted thing Convex could publish.

    Medium impactLow effort First result in 1 quarter Product Marketer

    First action

    Find the three most common costly patterns.

    How to measure it

    12-month retention of teams that use the check

    Now
    –
    Target
    +10 points
    By
    4 quarters

    Steps

    1. 1

      Find patterns

      Costly patterns in usage data.

      Software Engineer, AI 2 weeks

    2. 2

      Write the series

      Three posts.

      Product Marketer 3 weeks

    3. 3

      Build the check

      Dashboard flags per project.

      Software Engineer, AI 4 weeks

    4. 4

      Announce

      Stack and the weekly email.

      Product Marketer 1 day

    5. 5

      Track

      Fixes applied and retention.

      GTM Engineer Quarterly

    6. 6

      Repeat

      New patterns each quarter.

      Software Engineer, AI Quarterly

    First result
    1 quarter
    Cost
    Lower revenue per project
    Owner
    Product Marketer

    Smallest test

    Publish the first post.

    Why it works

    A series of cost-cutting guides (indexes, pagination, caching patterns) and a dashboard check that flags the three biggest savings for each project.

    Helping customers spend less builds the trust that makes them stay and recommend Convex, which is worth more than the savings.

    Risks and how to handle them

    • Revenue dips

      Retention and referrals more than offset it.

    Evidence

    • HiringProduct Marketer: "We ship frequently, but aren't great at talking about it"; "Once developers find us they tend to stick around"; "we'd sooner publish a 'how to give Convex less money' post than pretend what we do is magic"; owns launches end to end and writes the post and cuts the assets; "connective tissue between engineering, developer experience, design, and creators".Source
    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • Your websitePricing: Free & Starter ($0 and pay as you go, 1-6 developers), Professional at $25 per developer a month (log streaming, exception reporting, HIPAA/SOC 2 reports, daily backups, custom domains), and Business & Enterprise with a $2,500 monthly minimum. Startups can get up to a year of Professional free; the program lists YC and NEO companies and investors such as OpenAI Fund, Index Ventures and Accel.Source
  • 20

    Startup program companies are growing and hiring. Engineers who know Convex are their best hires, and a network of them shows the ecosystem is alive.

    Medium impactLow effort First result in 1 quarter Head of GTM

    First action

    Ask five program companies if they are hiring.

    How to measure it

    Hires through the network

    Now
    –
    Target
    20
    By
    2 quarters

    Steps

    1. 1

      Ask companies

      Who is hiring.

      Head of GTM 1 week

    2. 2

      Build the board

      A simple jobs page.

      GTM Engineer 2 weeks

    3. 3

      Invite engineers

      Opt-in talent list.

      Head of Marketing 2 weeks

    4. 4

      Promote

      Weekly email and Discord.

      Product Marketer Ongoing

    5. 5

      Track hires

      Ask companies quarterly.

      Head of GTM Quarterly

    6. 6

      Feature stories

      Hire stories on Stack.

      Product Marketer Quarterly

    First result
    1 quarter
    Cost
    $0
    Owner
    Head of GTM

    Smallest test

    Post three program jobs in Discord.

    Why it works

    A free job board and talent list: program companies post roles, engineers who build on Convex opt in with their projects. Feature it in the weekly email.

    Helping customers hire builds goodwill and keeps their teams on Convex as they grow.

    Risks and how to handle them

    • Recruiting is a distraction

      Keep it a simple board.

    Evidence

    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • Your websiteEvents and community: the "All Gas" hackathon sponsored by OpenAI, Firecrawl and AgentMail, an OpenAI Build Week community meetup in San Francisco, a Founders Inc startup partnership, and a Convex Champions ambassador program; community runs on Discord.Source
    • programmable.infoSite scan: 502 URLs including 181 component pages, 116 TypeScript guide pages, 62 event pages, 36 templates, 24 Databased podcast episodes, 23 "can do" pages, 6 hackathons, 5 compare pages (SQL, MongoDB, Appwrite, Firebase, Supabase), 5 evals pages and a single case study (Conduit). npm package "convex" about 1 million weekly downloads; 23,800 GitHub stars across the org (convex-backend 12,700); YouTube 13,800 subscribers, 416 videos, 1.4 million views; 63 Hacker News mentions. Named customers include OpenAI, Tripadvisor, a16z, Zapier and Reducto.Source
  • 21

    Convex Explained just launched with Build Your Own Convex. Interactive explainers suit both people and agents, and the Creative Director brief asks for marketing nobody would expect from a database company.

    Medium impactMedium effort First result in 2 quarters Creative Director / Head of D…

    First action

    Pick the next explainer topic.

    How to measure it

    Signups from explainers

    Now
    –
    Target
    500
    By
    2 quarters

    Steps

    1. 1

      Pick topics

      Five hard concepts.

      Head of Developer Experience 1 week

    2. 2

      Design

      Interactive pieces.

      Creative Director / Head of Design 6 weeks

    3. 3

      Write text versions

      For agents and accessibility.

      Head of Developer Experience 2 weeks

    4. 4

      Launch

      One every few weeks.

      Product Marketer Ongoing

    5. 5

      Share

      Hacker News and social.

      Product Marketer Per launch

    6. 6

      Measure

      Readers and signups.

      GTM Engineer Monthly

    First result
    2 quarters
    Cost
    Design time
    Owner
    Creative Director / Head of Design

    Smallest test

    Launch one more explainer.

    Why it works

    Five more explainers on the hard parts (reactivity, transactions, conflicts, caching, scaling), each playable in the browser and with a plain text version for agents.

    Explainers earn links and teach the ideas that make Convex the right choice, without a sales pitch.

    Risks and how to handle them

    • Slow to make

      Reuse components between explainers.

    Evidence

    • Your websiteStack blog: ClawHub, the skill repository for the OpenClaw project, scaled "from 0 to 1M visitors a week in just over a month" on Convex; a new "Convex Explained" home for interactive explainers; posts on the Claude Code and Codex plugins, competitive benchmarks and incident response that builds trust; a post about planning a conference.Source
    • HiringCreative Director / Head of Design: the brand "has to be distinct, thoughtful, a little playful, and honest. The kind of thing a skeptical engineer trusts on sight"; works in Figma and the codebase alongside existing creative freelancers. Product Manager (the second PM): product areas include the AI gateway, platform APIs and the component ecosystem; calls with customers "to understand why their team is spinning up hundreds of deployments"; works with "our Heads of GTM and Marketing".Source
    • HiringHead of Marketing: "At a company our stage the founders (James and Jamie) are the de facto CMO(s)"; own positioning, launches, content, community, events and brand; "Decide what we don't do. Pick the two or three motions that matter most and go deep"; "nothing hypey, nothing generic, nothing we'd be embarrassed to put in front of a skeptical engineer"; "Ship a bunch of cool marketing nobody would believe came from a database company."Source
  • 22

    Stack has a post on incident response that builds trust. For an infrastructure company, how incidents are handled in public is marketing.

    Medium impactLow effort First result in 1 month Co Founder / CEO

    First action

    Publish the most recent incident review.

    How to measure it

    Enterprise prospects citing reliability page

    Now
    –
    Target
    Tracked
    By
    2 quarters

    Steps

    1. 1

      Set the policy

      Which incidents get a public review.

      Co-Founder / CEO 1 day

    2. 2

      Write a template

      What happened, impact, changes.

      Head of Developer Experience 1 day

    3. 3

      Build the page

      Reliability page with history.

      Creative Director / Head of Design 1 week

    4. 4

      Publish reviews

      Within a week of incidents.

      Software Engineer, AI Per incident

    5. 5

      Share

      In the weekly email.

      Product Marketer Per incident

    6. 6

      Track

      Mentions in enterprise deals.

      Head of GTM Quarterly

    First result
    1 month
    Cost
    $0
    Owner
    Co Founder / CEO

    Smallest test

    Publish one review this month.

    Why it works

    For every customer-visible incident, publish a plain review within a week: what happened, who was affected, what changed. Link them from a reliability page with uptime history.

    Trust is the first of the three beliefs Convex needs, and honest reviews prove it in the moments that matter most.

    Risks and how to handle them

    • Bad press

      Plain and honest reviews usually earn respect.

    Evidence

    • Your websiteStack blog: ClawHub, the skill repository for the OpenClaw project, scaled "from 0 to 1M visitors a week in just over a month" on Convex; a new "Convex Explained" home for interactive explainers; posts on the Claude Code and Codex plugins, competitive benchmarks and incident response that builds trust; a post about planning a conference.Source
    • Your websiteStack post "Everything We Tried Failed Except for This One Weird Trick": the team tried newsletters, paid creators, Google and Reddit ads and an AEO agency (fired). Paid traffic had 3.3x shorter sessions, 2.5x the bounce rate and zero conversions; about $100k on newsletters and creators in two months brought visitors "who didn't actually build anything". What worked was the startup program: that cohort "compounds their Convex spend 30x over 24 months". The plan: more in-person events, bigger budgets for supporting demanding customers, "fewer paid sponsorships", built on trust, aliveness and recommendation.Source
    • Your websitePricing: Free & Starter ($0 and pay as you go, 1-6 developers), Professional at $25 per developer a month (log streaming, exception reporting, HIPAA/SOC 2 reports, daily backups, custom domains), and Business & Enterprise with a $2,500 monthly minimum. Startups can get up to a year of Professional free; the program lists YC and NEO companies and investors such as OpenAI Fund, Index Ventures and Accel.Source

03

How Programmable can help

Where an idea in this playbook is something Programmable builds or runs, it's listed here with what we'd do. Everything else is yours to run.

ProgrammableInfo

Turn the research into a list you can act on

These ideas start with finding the right companies. Info finds them and the right person at each, researches every one and drafts the outreach, then asks before it sends.

  • Build this list in Info for idea #20 (opens programmable.info in a new tab)

New accounts: 100,000 free credits with code BETA100K

Find the companies (opens programmable.info in a new tab)
ProgrammableTalk

Fill the events in this plan

These ideas are events. Talk gets registrants for them and tells you who's coming and what their company does before the day.

Plan an event (opens programmable.talk in a new tab)
ProgrammableDigital

Show it, don't just describe it

These ideas need something to show: a demo, a video, an animation, graphics or a deck. Digital turns screenshots or the site into all of them, in the company's own brand.

Make a demo (opens programmable.digital in a new tab)
More example playbooks and reviews

More growth playbooks

Other companies we have researched, each read from what they are hiring for.

  • Mishky Jewelry growth strategyHandmade beaded and embroidered jewelry (bracelets, necklaces, earrings, chokers and rings) made in Colombia, much of it by single mothers…22 growth ideas
  • Jensen Hughes growth strategyGlobal safety, security and risk consulting: fire protection engineering, code consulting, performance-based design, fire and life safety…21 growth ideas
  • CourseStorm growth strategyClass and camp registration and marketing software for informal education: catalog websites, registration and payments, rosters and…24 growth ideas
  • Oscilar growth strategyAn agentic risk platform for financial institutions: real-time decisioning, data and AI agents across fraud, credit underwriting…22 growth ideas
  • Suki growth strategyAmbient clinical intelligence: an AI assistant for clinicians (ambient notes, dictation, coding, order staging, patient summaries and Q&A)…21 growth ideas
  • MATTE Projects growth strategyA New York creative studio for brand worlds, experiential events, branded design, content and talent, with its own events, the HERO…21 growth ideas
  • The Plum Tree Group growth strategyA Chicago-based digital group: an ecommerce and digital marketing agency (BigCommerce, Shopify, Magento, Salesforce) since 2006, now also…21 growth ideas
  • Anaplan growth strategyAn AI-infused scenario planning and analysis platform for finance, sales, supply chain and workforce planning, with out-of-the-box…21 growth ideas

See every playbook and website review

Independent research. Programmable wrote this playbook from public sources, as they stood on October 18, 2026, and cites them throughout. Convex did not commission, review or endorse it, and Programmable is not affiliated with Convex. Company and product names are trademarks of their owners, used only to identify them. The ideas are suggestions and opinion, not professional, legal or financial advice, and not statements about how Convex runs its business.

Something wrong, or do you represent Convex? Email [email protected] and we will correct it or take it down.

30/60/90 roadmap

What to focus on in each month, and the ideas that carry it.

  1. Days 1 to 306 ideas

    Relaunch the agent research, start the weekly shipped post, act on serious-team signals, open founder referrals, write three case studies and measure the agent setup funnel.

  2. Days 31 to 6013 ideas

    Dinners, partner accelerators, migration guides, the platforms program, component authors, Champions, scaling reviews, docs evals and the exit plan, plus the wildcards.

  3. Days 61 to 903 ideas

    Decide the big bets: a conference, agent provisioning and an independent benchmark.

The research behind the plan

3 research areas, each summarised in a few lines. Open any of them for the full detail.

Customers

What customers say

Customers value reactive queries with no cache layer. The most cited problem: fear of betting a company on a new backend.

  • Everything in TypeScript
  • Postgres connection pooling and scaling
  • In their words: “all gas, no breakages”

Market

Competitive landscape

4 competitors compared: Supabase, Firebase, Postgres with an ORM and Appwrite.

  • Gap to exploit: A backend that keeps agent-written code correct
  • Supabase: Concurrency correctness left to the developer
  • Firebase: NoSQL limits and lock-in concerns

Hiring signals

Convex has 13 open roles: five in marketing, design and developer experience, the rest engineering and product.

  • Head of Marketing: Founders are the de facto CMOs; pick two or three motions; no hype. Ideas 7, 12, 16.
  • Product Marketer: Ships often but does not tell people; launches end to end. Ideas 1, 2, 19.
  • GTM Engineer: Thousands of weekly signups; product signals and useful, non-slop outreach. Ideas 3, 6, 13.

From the team that wrote this playbook

All products
  • ProgrammableAgencyWorking AI systems in days, not quarters (opens programmable.agency in a new tab)
  • ProgrammableInfoAn agent that finds, researches and asks first (opens programmable.info in a new tab)
  • ProgrammableDigitalScreenshots in, an animated product demo out (opens programmable.digital in a new tab)
  • ProgrammableWatchKnow the day a supplier falls out of compliance (opens programmable.watch in a new tab)
  • ProgrammableTalkGet registrants for your events, and know who’s coming (opens programmable.talk in a new tab)
  • ProgrammablePageText what you need, your customer gets a page that works (opens programmable.page in a new tab)
  • ProgrammableCommunityForums with labelled AI members, never empty (opens programmable.community in a new tab)
ProgrammableMarketing
ExamplesLeaderboardsProductsTermsPrivacyCookiesDPALegalContact

© 2026 AI-Powered Growth Intelligence