Hacker News new | ask | show | jobs
by benjaminwootton 1899 days ago
I have the same on Sendgrid, owned by Twilo. One day they just cut us off with no explanation, kneecapping our product. I raised two tickets, neither of which were answered. The next day, it all started working again.

I’ve always been a big advocate for using 3rd party SaaS, but with the level of customer service I’m seeing lately, I’m starting to recommend avoiding it where possible.

4 comments

That's 99% of why I'm generally trying to avoid SaaS. I always run into some bug or nuisance or there's some account trouble and unless you're spending insane amounts of money, I've found support to be severely lacking, often answering support tickets with canned responses without reading the issue, or not answering at all.

I'd rather spend 100 hours upfront to have something that I know and can maintain than using a service, running into trouble and spend 50 hours chasing the support around to see whether they'd fix it, or worse, rely on their system, have it fail or be banned, and then having to spend 100 hours building a replacement, often while things are exploding left and right.

Relying on SaaS feels incredibly fragile to me.

I think this should be regulated - companies often take massive amount of customers and while they are most of the time able to deliver the product, they have no capacity to service support queries. I think there should be an SLA guaranteed by law and for a company it should be illegal to take more customers if they are unable to provide SLA on support. This way we wouldn't have a situation where you contact a company and in reality you talk to bots or nobody at all. This would also deal with the problem of companies growing too big to fail and not having reasonable competition.
Or you could just, you know, buy from Saas'es that provide SLA's enforceable in your jurisdiction. Vote with your <insert local currency here>
I recently moved to Postmark. It’s been an extremely positive experience. They’re a small, dedicated team with amazing support.
It's an insane amount of engineering, but maybe one solution is to have two back end telephony SaaS. If one turns off, rely on the other until you get it fixed? Plivo, Mexmo?