I wouldn't go so far as "never", but historically, developer tools with lasting growth tend to spread via word of mouth & community recommendations rather than ad campaigns. See eg. Ruby/Rails, Python and supporting ecosystem, JQuery, Node.js, Express, Postgres, Vue.js, GitHub before they took VC, Parse before they were acquired by Facebook, etc. Even Google - they had distribution deals starting early on, but they were very reluctant to spend money on advertising up until about 2010 (I remember watching the first Google SuperBowl commercials at TGIF, and hearing a lot of questions about why were spending money on advertising and what it meant for organic growth).

When ad campaigns are run, they usually work by targeting programmers' bosses, encouraging the organization to purchase & standardize on one solution and then forcing developers to adopt it via management fiat. That's the route taken by Java, MongoDB, Oracle, .NET, many of the companies in the Hadoop ecosystem, etc. It certainly works - these are big companies - but it requires a consultative enterprise sales team that can work to get the solution deployed across the whole enterprise. The author obviously doesn't have the resources for that.

The middle ground, where you provide a developer tool with $100-200/year CLTV and hope to get it distributed via paid advertising, is a very difficult place to occupy. That's the area occupied by FogBugz and RethinkDB and Sandstorm and many analytics providers. Most of them go the community-building route, and it's still very difficult. Only company I can think of that's thrived here is GitHub, and they went the word-of-mouth, community-building route first. Companies like Jetbrains, Atlassian, Rational, etc. thrive as well, but they've got large enterprise sales teams that standardize a whole organization on their products.