Hype trains considered harmful
They come with a bang, they generate more FOMO than the last available room in booking.com, and when they leave it's your job to collect the wreck. In 2023, strange as it may seem, hype trains are still something to worry about. In the IT industry hype trains are particularly active. They build on the somewhat arbitrary nature of the technology selection process often found in small and large organizations. Of course, if the affected organization is a large player - say, part of a distribution oligopoly - the consequences might not be too serious, but if you own or manage a startup or an SME you should be concerned about protecting your business.
How can you tell a hype train from a solid change of paradigm? There's no definitive way to do so, but here are some empirical rules:
- check for vague words, such as cloud, agile, devops, datalakes or microservices
- check for vague promises such as faster deliveries or infinite scalability
- check the TCO and whether there is an industry agenda behind proposed technology
For instance, do cloud, agile and microservices allow for faster deliveries and infinite scalability? Or do they deliver undocumented, prone to breaking setups with a cost that scales faster than your revenue? Well, it depends. Unfortunately, in recent years mode-1 thinking and ideologies seem to have replaced the concept of it depends.
In order to change vague words into something we can assess, we need to ask further questions:
- by cloud do we mean VMs, SaaS, PaaS or Serverless?
- faster compared to what?
- micro compared to what?
- is infinite scalability, and its corresponding cost, necessary or should reasonable limits and a process to lift them be defined?
- what is the TCO of the solution (inclusive of internal hours) and how does it compare to alternative solutions?
- is the solution interoperable across several providers or is it vendor-specific?
Anyone with answers to these questions will probably conclude that hype trains are easy to spot.
Now, you might be wondering if all IT related words are purposefully vague and part of a marketing-driven jargon. They're not. Is there a problem with the words Infrastructure as Code? Or with the sentence by investing effort in defining your infrastructure as code you allow for quick replication and reconstruction of sophisticated systems? There isn't. Fortunately, many things remain sound and objective.
So, where do hype trains come from? DHH strikes at an organized agenda fostered by the merchants of complexity. Baldur Bjarnason points to an organic process stemming from Pop Culture inside Tech Companies. They're probably both right to some extent. We have to insist on rational decision making.
Fun fact from another industry: it seems that car makers are coming back to buttons after the touch screen hype train has passed. What's next?