What's a good vendor?

Every place I’ve worked has relied on vendors in one way or another, as most modern businesses do. Some are a pleasure to work with; others… well, they merely exist.

Over the last 2-3 years, I’ve come to think there’s a pretty simple test for what makes a vendor good:

A good vendor should solve problems for the business and try not to introduce new ones.

Take a database vendor as an example. If you’re a small team, the whole point is to offload operational work: backups, upgrades, monitoring, scaling, and so on. You also don’t want a vendor that struggles to keep its service up. At the very least, a good database should be online.

If we look at companies in this space, we can judge them by those standards. Backups are table stakes; you shouldn’t have to think twice about them when you choose a vendor. But beyond that, things start to get trickier, which I think is a little sad.

Database companies quite often offer basic monitoring such as CPU, memory, and disk usage, but not all of them go far enough to provide query performance insights, analyze indexes, or suggest optimizations.

They also don’t always provide a solid upgrade strategy, which leaves the customer worrying about downtime, planning the upgrade, and wondering what happens if it doesn’t work out.

Then there’s scaling. Most providers make it easy, but often at the cost of performance. Network-based storage is convenient and easier to scale than NVMe drives, but the performance difference can be huge.

In my work, I try to revisit the vendors we use from time to time and ask a simple question: do they create more solutions for me and my colleagues, or more problems?