What is a PST File And What Outlook Actually Stores in It

A PST is a database rather than a mail file. It has an allocation map, two indexes and a numbered entry for every object inside it, which is why it holds far more than messages and why it never shrinks just because you deleted something. Almost every surprise a PST produces traces back to that one fact.

Outlook calls it an Outlook Data File. Windows calls it a PST. The specification calls it a Personal Storage Table. Older Microsoft documentation calls the same thing a Personal Folders file or a Post Office File. All of those are one format. Two products carried it before Outlook existed. Microsoft Exchange Client and Windows Messaging both used it, which makes the format older than the program most people associate it with. What none of the three names conveys is that opening one is closer to attaching a database than to opening a document. There is a wider guide to every email file format if you are not yet certain this is the one you have.

The Three Layers Inside a PST File

Microsoft publishes the structure as MS-PST. That document splits the format into three layers stacked on each other. Each one only makes sense once the one below it does, which is why writing a PST reader is a serious piece of work rather than a weekend project.

Two details from that bottom layer explain most PST behaviour. Every object carries a node identifier. The file also keeps its own map of which blocks are in use. Delete a message and the messaging layer removes it from a folder, but the blocks it occupied are simply marked free in that map. The file on disk does not move a byte.

Everything a PST Holds Besides Mail

Ask what is in a PST and the usual answer is emails and maybe attachments. The format was built for the whole of Outlook, so the real list is longer. The extra items are exactly the ones that vanish quietly during a conversion.

Every one of those extra fields is a MAPI property, the same kind an MSG carries for a single item. MAPI is the interface layer Outlook sits on. A PST is a MAPI store, so the file has somewhere to put things that no message format defines. Everything from Categories downward is configuration rather than content. It describes how Outlook should behave with your mail rather than the mail itself. No interchange format has anywhere to put that. That is not a flaw in the converter. It is the reason the PST existed in the first place.

Search folders are not folders. A search folder holds no messages. It stores a query that Outlook runs each time you click. Copy a PST and the search folders come with it. Convert one and they disappear entirely, because there is nothing to convert. Every message they appeared to contain is still there, sitting in whatever real folder it always lived in.

What a PST Does Not Hold

The gaps matter as much as the contents, particularly for anyone handed a PST and told it is a complete backup of somebody's mailbox.

Which Accounts Actually Produce a PST

Whether Outlook writes you a PST at all depends on the account type, which is why some people have one sitting in their Documents folder and others have never seen the extension.

So most PST files in circulation were made on purpose rather than by default. Somebody exported a mailbox or set up a POP account long enough ago that it was still the normal thing to do. The exception is AutoArchive, which moves old items into a PST on a schedule and is the reason plenty of people have one they never knowingly created.

The reason to make one has not changed since the format arrived. A PST sits on your own disk, so it is not counted against any mailbox quota the server enforces. That single property is why administrators once encouraged them. It is also why so many organisations now hold thousands scattered across laptops with no idea what is in any of them.

An exported PST is invisible to the server. Anything moved into one leaves the mailbox, so it stops being covered by server side search, retention policy or legal hold. For a personal archive that is the point. For a regulated organisation it is the reason PST sprawl is treated as a compliance problem rather than a storage one, since an eDiscovery request cannot reach what the server cannot see.

Why a PST Never Shrinks on Its Own

Delete a year of mail, check the file, find it exactly the same size. The answer is the allocation map. Deleting marks space as reusable inside the file. It does not return that space to the disk.

Compacting rewrites the file without the free blocks. Outlook does it slowly in the background when idle. It can also be triggered on demand through the data file settings. On a file that has had a large clear out it can recover a great deal. Nothing else will.

Size limits count the file, not the mail. Microsoft documents a maximum enforced through the registry, measured against the file on disk. The ceiling depends on which variant you have, since the older ANSI format stops far short of the Unicode one that replaced it. So a mailbox well under the limit can still hit it, purely because deleted items and indexes have not been compacted away. Clearing out mail without compacting afterwards fixes nothing.

What PST Passwords And Encryption Actually Do

Outlook offers a password when you create a PST. The word implies the contents are locked. They are not.

What is stored is a CRC-32 checksum of the password text, held in a property called PidTagPstPassword. Outlook computes the checksum of whatever you type and compares the two. The specification is blunt about what that achieves, saying the password is implemented at the interface level only and is not used to secure the PST data in any way.

read PidTagPstPassword
prompt the user
CRC-32 what they typed
compare the two values
match, so open the store

read the header
walk the node B-tree
read the data blocks

Outlook also offers an encryption setting when the file is created, which sounds like the real thing and is not. Two options exist beyond none. Both use keyless ciphers.

Neither takes a key, so neither can be a secret. Microsoft states the position in the specification itself, that the strength of the encoded data blocks provides no additional security beyond data obfuscation.

Only your data is obfuscated. The structure around it is not. The specification notes that headers, allocation pages and B-tree pages are all written without any encoding at all. So even the weak layer covers message content only. Anyone with the file can map its entire shape without decoding a single byte.

The name is a historical accident. A Microsoft engineer has written about where compressible encryption came from. Compressing a PST made its contents unreadable in a text editor. People equated that with encryption. The label stuck. The word would not survive a review now. Confidential mail needs real encryption applied separately, whether that is BitLocker on the volume or an encrypted container holding the file.

What This Means for Anyone Handling a PST

Damage has its own tool, already installed. Outlook ships with the Inbox Repair Tool, scanpst.exe, which rebuilds the indexes when a PST will not open. It works on the structure rather than the content, so it recovers the file more often than it recovers everything in it. Take a copy before running it.

A disconnected PST does not announce itself. Close one in Outlook or lose the profile that referenced it. The file stays on the disk with nothing pointing at it. Nothing warns anybody. Orphaned PSTs are how organisations end up holding years of mail that no backup covers and no search reaches, discovered only when somebody goes looking for a specific message.

Four rules follow from the file being a database rather than a mail file. Each one is the opposite of what a mail file would let you get away with.

1. Never put one on a network share. Microsoft has documented this as unsupported since Exchange 4.0. A PST is read and written with file access commands meant for a local disk, so over a network every operation slows down. Writes take roughly four times as long as reads. A connection that drops mid write can corrupt the file outright. Copy it to the local disk first. The single exception Microsoft documents is Outlook running remotely on a Remote Desktop Session Host, where the file and the program sit on the same machine anyway.
2. Close Outlook before copying one. A copy taken mid write can capture a half finished index update. It will often open and look fine, then fail months later when you reach the part that was being written.
3. Compact after a clear out, not before. Deleting alone changes the size on disk by nothing at all. Until the compact runs, every limit still sees the old figure.
4. Do not rely on the password. Treat a PST as readable by anyone holding the file, because it is.
5. Expect configuration to be lost in conversion. Rules, views, categories and search folders have no equivalent anywhere else. Write down what matters before you move anything.

If the job is reading what is inside rather than moving it, a browser based viewer parses the format directly and never needs Outlook or a mail profile. Where the mail is going somewhere else entirely, our mail converters handle the route. The survives column earlier sets out what to expect on arrival.