Marketing to developers

How to target developers (where traditional advertising often fails) & a comparison of the ad channels/platforms.

TL;DR for the TL;DR

• choose goal
• choose audience-appropriate platform
• setup conservative targeting (better to under- than over-serve)
• design practical & relevant ads
• connect to the right landing page, amplify paths people are already taking
• run small tests before scaling ad spend

Targeting developers

• most devs hate “traditional” (pop-ups, flashy colours) advertising
  • dev personality: sceptical, pragmatic, practical, “tribal” (trust other devs’ expertise more than marketers’)
• retargeting (ads follow users across sites) works well but requires analytics/tracking not great for devs
  • maybe ok for late stage (consideration, activation, engagement, upgrades)
• paid ad campaign components:

1. Goal: desired reaction of target (dev) audience
   • e.g. tool signups, platform leads, etc.
   • if they don’t yet need your tool, build a memorable brand so they remember you in future when they do need a tool
   • signups should actually use the tool – follow up if not
   • free users should upgrade to paid plans
2. Targeting: where to post? who views?
   • e.g. platforms (Reddit, Facebook, Twitter, LinkedIn, YouTube, …)
   • ask customers which platform they prefer!
   • context: will a platform’s users be in the right mood for your ads?
   • go deep/specific, e.g. target:
     • tech stack
     • role ( FE front-end
       , BE back-end
       , DevOps (software) development + (IT) operations
       , ML machine learning
       , DS data science
       )
     • company (enterprise, startup, freelancer)
     • level (junior, senior, architect)
   • use analytics/tracking/cookies (for now) to filter out non-relevant targets
3. Creative: what to post?
   • format (in-feed, banner, pop-up) & medium (video, animation, text) should be non-intrusive & “as non-advertising as possible”
   • content: actual message to convey
     • don’t fear some explicit, technical jargon and sarcastic humour
     • focus on features (“5ms video load time”) rather than values (“better videos quicker”)
     • link to comparison to alternatives
     • link to case studies & use cases
     • link to getting started, examples, and docs
4. Landing page: where ad clicks lead
   • usually not good to use company homepage: too stuffed with projects unrelated to the ad
   • pick based on existing analytics (what are users currently doing), e.g.
     • amplify paths: twitter -> blog on problem X -> product X -> signup implies twitter ads should link to blog on problem X
     • shorten paths: homepage -> product X -> docs for X -> pricing -> signup implies ads for X could link to docs for X
     • amplifying might be better than shortening/creating new paths

Channel comparison

key: emojis indicate how well things work.

• experience of channels: most don’t work well
• Google Display Network (GDN)
  • used by “most of the internet” so general (intrusive, not dev-focused)
  • could specify conversion event for Google to optimise or filter pages where ads are shown to target better
  • not a magic bullet
  • easy to overspend
  • there are many guides
• Ethical Ads
  • targets devs (blogs, readthedocs, etc.), context of ads is “learning/working”
  • small (image + brief text), non-intrusive
  • pay per impression
  • GDPR-compliant (no tracking – targeting is done by page)
  • filtering by page costs more
• Carbon Ads (BuySellAds)

  • focus on devs (like Ethical Ads) – mostly ML machine learning
    and DS data science
    

  • more intrusive (like GDN)

• Twitter Ads
  • appear in-feed
  • important to target engaging keywords & “look-alike” (similar) followers
  • good guides aren’t dev-specific
• Reddit Ads
  • in-feed, in-thread, or in-conversation
  • can target subreddits
  • can only filter on location & device (not keywords nor topics)
• StackOverflow Ads
  • non-intrusive, ads look like answers to questions
  • managed (not self-serve), min. contract $10~$15 k/month
• YouTube Ads
  • video remarketing works well: amplify existing conversion paths by showing ads to people who visited your product/pricing/docs/high-intent-blog
  • lots of targeting options (like GDN)
  • intrusive, especially without proper targeting
• Google Search Ads
  • branded search works well: “a google tax on keywords” containing your product name (pay for a keyword to show up in search results)
  • top of SERP (intrusive default) or page bottom (feels more helpful)
  • think about relevancy:
    • brand name (“github”)
    • core features/use cases (“ CI continuous integration
      “) – don’t work well if linking to blogs
    • comon typos of both off teh abov (“gitub”)
    • competitor comparison (“vs github”, “github alternative”)
    • integrations (“slack”)
    • competitor (“gitlab”)

Extended discussion at ycombinator#32191615.