Comment by erikpukinskis | Hacker News Reader

Comment by erikpukinskis | Hacker News Reader

I think in general we’re about at the limit if “one size fits all” UIs especially for things as personal as workflows, customer data, etc.

I’ve long thought the idea of “A Todo App” makes no sense because everyone works a little differently. That’s why apps like Pivotal Tracker are such garbage... they support too many use cases to do anything well.

The problem is if you try to make it configurable, you still have all the complexity behind the scenes. And your configuration UI recapitulates that hell.

So there is a need for someone to come in and start over with a more configurable foundation and maintain really serious software engineering discipline so that the base layer doesn’t get infected with individual business requirements (which will eventually be infinite).

I don’t envy Airtable trying to do this with a big client roster and huge scale. I think the problem would be easier solved in a smaller company with less pressure from investment timeline.

BUT if they’ve already solved the problem, or at least solved it to their satisfaction, then they will be fine and the platform can get as convoluted as they need to make their clients happy.

It’s only if there’s research remaining, core UI questions to be solved, that they’ll have a hard time competing with whomever comes to eat their lunch.

I've been fortunate in building a consultancy that's grown pretty quickly almost entirely based on this insight (https://superwork.io). We've helped some companies basically 2x just by dumping their existing SaaS-based processes.

The days of one-size-fits-all SaaS are numbered.... most companies can find massive efficiency gains through smart deployment of tools like Airtable and Zapier with turnkey automations.

There used to be a time this wasn't feasible because the only option was to hire a development team and deal with legacy software - we don't have that problem anymore.

"We're world leaders in automation & optimization. We know our shit."

Ha, brave language!

Quite brave. I'd A/B test that with something more benign to see how it's really performing.

I've been looking for exactly a consultancy like this. Thanks for sharing.

Feel free to contact me directly, I've got some business for you.

> The days of one-size-fits-all SaaS are numbered.... most companies can find massive efficiency gains through smart deployment of tools like Airtable and Zapier with turnkey automations.

Yeah, we're working through Zapier integration at the moment for our paperwork assistant (https://lexico.io) and I'm pleasantly surprised that there aren't as many established B2B SaaS companies doing the same thing as I thought. Self-serve automation definitely feels like the path ahead.

Still don't quite get Airtable though.

Very similar to Zapier.

The interface is very slow to work in, and everything requires multiple steps, because if they are going to facilitate the endless possibilities their customers present, everything has to be atomized into pieces that fit every puzzle.

I have found, after working in Zapier for a couple years, that the primary value doesn’t come from ease of use or the ability of non-technical people too easily do technical things, but rather in the flexibility afforded by a layer that connects services together in any way conceivable.

My guess is that air table will become generally useful in a similar way, where maybe you need some technical people to help you set up things, but at least you can achieve most of what you need, as a small company, without having to hire a team of programmers to build it from scratch.

> a layer that connects to services together in anyway conceivable

This sounds quite much like general-purpose programming. But the tool is much higher level, apparently, at least, in some respects.

This opens (logically) endless possibilities, but watch out for well-known programming pitfalls; the Turing tarpit is already referenced as "very slow, and everything requires multiple steps".

This is exactly the problem we have tried solving with the Frappe Framework [1] (Open Source, MIT).

The core concept in Frappe is a DocType, which is configurable metadata. Using this you can quickly build small, special purpose apps for very niche targeted use cases for their own business. You can also configure rule based permissions and add simple scripts. We have used it to build hundreds of use cases for ERPNext (Open Source, GPL) by allowing complexity to grow linearly (instead of exponentially).

This also helps keep the base layer and the "one size fits all" business layer (like say an Accounting Ledger) clean. And you don't have to worry about UI, permissions, REST API, deployment, maintenance and users get a consistent experience.

[1] https://frappe.io/frappe | https://github.com/frappe/frappe

I agree with this sentiment except for your reference to Pivotal Tracker - it's the opposite of “one size fits all”.

It is so ingrained into the particular Agile Scrum way of doing things that this is actually cited as its drawback by teams wanting to be more "flexible".

Are you getting it confused with Jira?

Reply on news.ycombinator.com