How new SaaS platform categories are born

How new SaaS platform categories are born
💫
Spoiler: It's like for galaxies but instead of gravity forming stars out of dust and gas, people come up with new processes until friction is so high that the techno-capital machine spurts out new SaaS platforms that end up forming a category.

I'll take the platform engineering space with internal developer portals as an example category, since I have worked in the space for the last four years and understand it better than, say CRM.

Big companies, like every other organism, have always at least one bottleneck that hinders them from reaching their goals better/faster/cheaper.

Platform engineering could only gain momentum because lot's of companies understood the ability to ship software will become even more important going forward. For those companies, increasing the ability to ship software was getting into focus.

Whenever there are common pattern of identified bottlenecks and interactions (and thus coordination problems) across companies people start to experiment with different solutions. Those solutions are generally a mix of resources, processes, and technology.

Throwing more resources at the bottleneck is a band aid

If you can simply throw more resources at the bottleneck, the big consulting companies are happy to help you out. If this fixes your problem, congrats, move on to your next bottleneck.

If short term attention doesn't fix your bottleneck, you can hire more people that and make fighting the bottleneck their job. However, this becomes less attractive over time, especially if more companies do the same thing and those resources become scarce.

In the case of platform engineering, the demand for good software engineers was pretty high. This made the "just throw more cash at it" approach less feasible/desirable over time for many companies.

As I said, resources, processes, and technology are your levers. We looked at resources. Let's look at processes next.

Every process is an answer to a coordination problem

For a single person or a handful of people working together, information spreads easily and making decisions comes naturally, since interpersonal uncertainty is low.

This changes if the group of people involved gets larger. Suddenly people are missing context and might have uncertainty when to involve whom. Processes help here, because they prescribe who reaches out to whom with what information and who decides what. Fundamentally, processes are communication devices that trade individual freedom off against reliable outcomes.

For platform engineering, many companies started out with a team that managed cloud providers or private cloud infrastructure (like shared kubernetes clusters). Sticking to the terms of team topologies, those teams were "platform teams" because they weren't stream aligned to the core business of the company. These platform teams often worked on an ad-hoc basis and didn't have a strong enough mandate to invest into making their offers scalable. Thus more processes for how to interact with the platform team.

So now we have a bunch of new processes. Which mean less freedom for your best employees and more friction for everybody. And for an organisation, friction means money is left on the table.

Platforms help with coordination

At this point the knowledge about a coordination problem is known to people that work in or close to big companies. So why are platforms the logical next step? I have made the case in my last post: platforms allow you to enforce a consistent model of a business domain and define interactions of participants. Thus moving on from coordination problems involves platforms.

Internal developer portals give developers access to approved services. The offering is restricted to what makes sense in the context of the business and is pre-configured in a way that makes the developers life easier. By decoupling the design of the service from the consumption both teams can work in their own pace. All of this contributes to less process while increasing reliability.

As always, technology is the only way to do more with less.

Platforms will not die out after all

In my last post, I argued that AI tools make individuals more powerful and eliminate some coordination problems we faced in the past. With this post I wanted to lay out why we will continue to see new SaaS platform categories. It's simply a matter of industries figuring out where the next bottleneck is and experimenting with resources, processes, and technology until a new pattern is established and we have a new three-letter abbreviation for yet another SaaS platform category.

📣
Shameless plug: I recently quit my job to build a tool that write user documentation for software products. It's called Docsy. If you are working in the space, I'd love to hear from you!