I've thought a lot about developer evangelism in the last three years. What a developer evangelist is, what he actually does etc. Companies selling a technical product often feel like they should invest a lot in "developer evangelism." But (i) it's not always relevant and (ii) it's always temporary.

Twilio or Sendgrid have invested a lot in developer evangelism and this strategy proved to be quite successful. Because they've acquired a poster child status, a lot of tech startups want to replicate their strategy. This is often a mistake where you forget that the market matures. 

Initially, it was vital to have developers explaining to their fellows that SMS or email is much more fun with an API. There was a real pedagogical stake: developer evangelists needed to break the good news. Their core message was about an unknown technology: "hey fellow developers, here's a new way to get more stuff done!" Carrying such a message is a great job.

However, If they do their job correctly, developer evangelists quickly turn into Account Managers. We're in 2014: Twilio or Sendgrid no longer need to explain why their products are useful. Enough people know what they do. The lonesome developer who might not know that he can send an SMS with an API will always have someone in its network to loop him in. So yeah evangelizing is fun, but it doesn't last. 

The other bad news is that being a pionneer means you work for your competitors. Look at how Nexmo (SMS) or Mandrill (email) reap the benefits. They don't need evangelists. Twilio and Sendgrid invested a lot to shape the market: the low-cost newcomers couldn't have done it without them. 

As a result, pionneers tend to switch back to the classic model: they start targeting enterprises instead of developers. They do it because this is where the high margins are, but also because the users have been evangelised. When everybody is rushing through the fresh bottom-up breach, the classical top-down approach starts to make sense again. Business as usual steps back in.

Therefore, the existing developer evangelism team looks for a new role: "gathering customer feedback and maintaining relationships with the ecosystem" becomes their primary objective.

However, this latter "relations" mission is overrated: friendship won't help to drive product adoption: "I like these guys but this new service is so convenient and 5 times cheaper." Sad statement but hey, let's face the truth: it's a B2B world, we're not in the coffee shop sector, having sympathy for the owner isn't key.

Now that the market is shaped, the CMO throws more money into acquisition. The immediate danger is a misunderstanding with the evangelists: attending events isn't a reliable lead generation technique. As the marketing team becomes aware of this, Developer Evangelists are ripe to be turned into Account Managers. That's excellent timing since they can help the enterprise Sales, who don't understand much about the product and the promises they made.

Bottom line is you evangelize a technology, NOT a product. So if your technology isn't new, you probably don't need developer evangelists.

Tip: if you want to be an evangelist, go for APIs in general or for Bitcoin. You will end up becoming an Account Manager if you're successful, but you will have had fun in the meantime. 

Last but not least: this explains why the turnover is so high when it comes to developer evangelists. They just get bored when you try to turn them into Account Managers.