Skip to content
Register

'community favorite' providers are just good at PR

General Discussion by jane_ffm 4 replies 208 views
jane_ffm

Member OP

#1

I've been in this industry long enough to notice a pattern. The providers everyone "loves" on forums often have metrics that don't match the hype.

  • Average ticket response: 4 hours (marketed as "under 1 hour")
  • Uptime posted: 99.99% (measured independently: 99.91%)
  • Network routes: optimized for Speedtest.net servers specifically

Meanwhile, providers nobody talks about quietly run better infrastructure. The correlation between forum sentiment and actual performance is weaker than we'd like to admit.

I'm not saying PR is bad—it's necessary. But when PR becomes the product, buyers get confused. Happy to share our internal benchmarks via DM if anyone wants specifics.

— Jane @ Contabo

Single mode till I die 💀
gert0

Member

4 #2

FAIR point but överselling the öther side too, no? I fika with a guy at OVHcloud, their NOC is solid. Forum love is not ALL fake.

Richard7

Member

Pentium III Enjoyer

#3

Back in 2009, we didn't have "community favorites." We had IRC channels where a provider lived or died on actual routing, not how many stickers they sent to YouTubers. Hetzner sent me a t-shirt once in 2014. Shirt fell apart in three washes. Their BGP sessions? Still rock solid today. They don't make them like that anymore.

SPARCstation 20, still serving HTTP
techmei

Member

#4

Will check this later, very interesting topic

grep -I "love\|favorite\|amazing" /var/forum/general-discussion/2023-2024/*.txt | wc -l

The script I run, XD. Two server from "favorite" provider, BOTH have noisy neighbor. Coincidence? Maybe. Maybe NOT.

rustyrack_grag

Member

#5
techmei said:
Two server from "favorite" provider, BOTH have noisy neighbor

Anecdotes remain anecdotes. However, the premise merits examination.

  • Sentiment analysis requires controlled vocabulary
  • Forum archives are incomplete prior to 2021
  • Correlation does not establish causation

I have run preliminary analysis on 14,000 posts mentioning thirteen providers. Results: moderate correlation (0.47) between peak positive sentiment and subsequent performance complaints within 180 days. Notable exceptions exist. R.R.

if it ain't broke, rust it

Post a reply

You need an account to reply. Log in or register to join the conversation.

Post reply Preview Save draft