Estimate data needed for cloud backups.
Cloud backup traffic has a large one-off component and a much smaller recurring one. The initial seed uploads the entire data set once; after that only changed blocks are sent, so ongoing volume tracks the daily change rate rather than the total size. Compression and deduplication apply to both, typically cutting general office data by 30-50 percent. Separating the seed from steady state matters because the first month can be ten times heavier than every month afterwards.
Seed plus incremental backup volume
Seed GB = dataset x (1 - compression); daily GB = dataset x change rate x (1 - compression); first cycle = seed + daily x cycle days
General office document storage is typically 1-5 percent a day. Databases and virtual machine images run much higher, often 10-20 percent, while archival media libraries can be well under 1 percent.
Many providers offer seeding by physical drive shipment, which removes the initial upload entirely. That is worth doing whenever the seed would take more than a few days at your upload speed.
Because it must fit in the nightly backup window at your upload speed. A 15 GB daily incremental needs about three and a half hours at 10 Mbps upload, which is fine; at 2 Mbps it will not finish overnight.