The ideal situation is rotating tie replacement that only replaces the ties that have reached the end of their useful life. Useful life being defined as still doing their job, albeit just about to lose the ability to hold gage, support weight, etc. This cycle is based on the life of the tie which has to do with the wood type, treatment, roadbed condition, and location. When section gangs were used and supplied adequately, this system worked well and over time could be fairly predictable. Additionally, you didn't end up with big sections of track where all the ties go bad at once. The down side is that there is usually somebody somewhere working on the track and requiring protection, slow orders, etc. The other disadvantage is surface and alignment. Done properly, single tie replacements should maintain and possibly improve surface and alignment. Key words - done properly. With today's maintenance blitz methodology it may be easier to plan for a track outage, the work gets done all at once and other than inspection, should require no fussing for quite awhile, depending upon traffic.
Over the years we could expect a tie to last about 15 years. As our roadbed improved (drainage) and we shifted to higher grade ties, we have seen this time increase. We have a bunch of Azobe ties which have done well but we have discovered some that have not held up as expected. Wood quality in general is not what it used to be. I have seen very, very old ties still holding spikes, not cut, and doing their job long after a newer tie has failed. That is strictly a function of the tie probably being from some old growth tree.