An uptime percentage describes availability over a measurement period, but the number alone does not explain redundancy, maintenance exclusions or how incidents are measured.
What to consider
For business planning, consider both the infrastructure target and the application architecture. A single server can still be a single point of failure even if the underlying data center is highly reliable.
Monitoring from outside the hosting environment provides a more complete view of user-visible availability and helps distinguish application failures from network or infrastructure events.