Ideally yes but you don’t need to be too OCD - with small triodes, often you’re trying to match 4 elements (2 triodes per tube) so it can be frustrating to try to find 4 within 10% of each other on the major parameters (mu, transconductance) - especially vintage tubes. And even up to a 20% difference sounds like a lot, but it can be just fine for many applications. If your elements are within 15% (max) of each other you’re generally doing quite well. I believe Upscale matches their "Platinum" grade to within 10%.
If you have an egregious mismatch, well over 25%, then that can start to cause very audible issues. I had a pair of Cifte 6189 silver plates where one tube was much stronger than the other, and it caused about a 1 dB L/R channel imbalance when used in the mu follower slots of my phono stage.
Once I also bought a trio (pair and a spare) of Brimar CV4004 (12ax7) from Parts Connexion, unaware that they were going to be COMPLETELY unmatched from an extremely variable stock. They were a complete mess, egregiously mismatched on triodes and tubes (variances or 40% and more), and sounded like an absolute mess as a phase splitter my tube amps. Fortunately the vast majority of dual-triode tubes that have been sold to me as "matched pairs", over the years, have been as advertised and sound symmetric when in an amp. Parts Connexion is the only shop/dealer I'll never buy tubes from again.
Some of the old tubes were abused by gear, more on one side than the other. I’ve seen a number of Tung-Sol BG RP 6SN7 (the holy grail tube) where one triode measures like new and the other is practically dead (measuring 50% less than the strong triode). Those are no good, so watch for that :(