> Or could close their "beta" you were depending on without notice, could lose your data, could leak it, could give unexplainable corrupt results and leave you looking like a fraud, etc.
I don't have to use SaaS to have these results. I've seen in-house efforts with the same risks. The difference, in my mind, is that with SaaS, the provider has some significant incentives to do a good job. The software is their specialty, and the developer will lose money if you stop using it because their software ends up costing you money. There's much less risk of internal company politics saddling you with a useless or dangerous system. On the contrary, I've seen "self-hosted" company projects promised in response to proposals to use SaaS, and they were usually promised to achieve some political goal rather than being a serious attempt to compete. Because it's an internal project which doesn't generate revenue, it's very easy for the company to promise something and then not deliver, as there's no money tied to it.
Having seen this multiple times, I'm much more inclined to use SaaS. Tail risks don't cost nearly as much money over the mid term as the risk of company politics derailing an otherwise laudable goal.
I don't have to use SaaS to have these results. I've seen in-house efforts with the same risks. The difference, in my mind, is that with SaaS, the provider has some significant incentives to do a good job. The software is their specialty, and the developer will lose money if you stop using it because their software ends up costing you money. There's much less risk of internal company politics saddling you with a useless or dangerous system. On the contrary, I've seen "self-hosted" company projects promised in response to proposals to use SaaS, and they were usually promised to achieve some political goal rather than being a serious attempt to compete. Because it's an internal project which doesn't generate revenue, it's very easy for the company to promise something and then not deliver, as there's no money tied to it.
Having seen this multiple times, I'm much more inclined to use SaaS. Tail risks don't cost nearly as much money over the mid term as the risk of company politics derailing an otherwise laudable goal.