One template, many recipients
Fan-out inverts the usual editing posture — instead of working on one document longer, you work on one document once and multiply it. The template is finished first, fields settled and layout approved, because every defect it carries is about to be copied once per record. Then the record list is filled in: one row per recipient, every needed field present, no row half-complete. The batch filler stamps each record into its own copy, and the payroll variant does the same for salary slips with the slip layout built in. The first verification is arithmetic — records in, copies out, counts equal — followed by a spot check that two copies differ only in their filled fields. The output choice made here, one merged bundle or separate files, echoes through every step that follows.
The mark names the distribution
A watermark on a batch is the distribution's signature, and it is applied after filling, not before. Marking the template would stamp identical text on every copy with no regard for the filled values; marking the finished copies puts the sign deliberately on top of final content. What the mark says is a decision worth a minute of thought — a document title or a confidentiality notice states status for the whole batch, while recipient-identifying text turns every leaked copy into a traceable one. The personal data itself already lives in the filled fields, so the watermark names the distribution rather than the person. Rotation, opacity, and tiling are tuned until the mark is legible at a glance without burying the numbers underneath — a watermark that hides the salary figure has failed at its only job.
Merged convenience, split necessity
The merged bundle is the honest choice for internal use — print it in one pass, archive it as one file, review it as one document. But individual delivery needs individual files, and plans change after fan-out. Batch splitting exists for exactly that turn — break the merged bundle by page ranges, or go single-page and regroup, to recover one file per copy. The discipline is the page arithmetic — after splitting, the parts must cover every page of the bundle exactly once, with no page lost between ranges and none duplicated across parts. A missing payslip page discovered after delivery is far more expensive than a count taken before it.
Compress last, and once
Compression is the final step because every pass costs a little image quality, so the batch deserves exactly one, taken when nothing further will touch the files. The depth follows the destination — screen-grade downsampling squeezes image-heavy pages under email attachment caps, while lighter settings preserve what a printer needs. Compressing at the end also means the checked size is the delivered size — what fits the channel's limit is literally what the recipient downloads, not an estimate from an earlier pass. The batch lands as one ZIP holding the full expected file list, which is itself the last verification — open it, count the files against the records, and the delivery is accounted for.
Where this workflow stops
This page multiplies one finished PDF into personalized, marked, deliverable copies. Around it sit the neighbors with different jobs. Designing the document itself, or generating certificates, invoices, and reports from structured inputs, belongs to the templated generation and layout workflows. Proving authenticity — signatures, certificate sealing, invisible tracking — belongs to the signing and tracing workflow. Rescuing one damaged file or slimming a single document belongs to the repair workflow, and performing page surgery on one copy for print belongs to the assembly and print preparation workflow. Fan out here — and when a single document needs a single fix, send it to the single-document pages.