Choosing a purchase order solution is usually presented as a feature comparison and is better treated as a sizing exercise. Three facts about your organisation, none of them about software, determine which shape of product fits, and once you have them the shortlist is short. Getting the sizing wrong is the expensive mistake here, in both directions.
The three facts
How many orders you raise a month, how many people raise them, and how often a delivery arrives incomplete. Low on all three and the module in your accounting system is probably enough. High on the second and third and you need line-level receipt tracking. High on all three with a dedicated buyer and a suite starts to make sense.
Why partial deliveries matter so much
They are the reason document-level tracking fails. If deliveries almost always arrive complete and on one date, an order is effectively a single event and simple tools cope. If they arrive in pieces over weeks, the order is a running account and needs quantities per line, which is where cheap tools stop being adequate.
What to insist on regardless of size
Automatic unique numbering, approval before the order is sent, receipts recorded by whoever takes the delivery, and an export you can read without the vendor. Those four are cheap in any product and expensive to add later, and the fourth is the one people forget to ask about until they want to leave.
Questions people ask about purchase order solution
Should we shortlist by industry?
Rarely. Industry-specific products matter where the buying itself is unusual, such as construction retentions or hospitality market pricing. Otherwise sizing matters more.
How much should this cost?
It varies enormously by shape. Work out which shape you need first, since comparing an accounting module against a procurement suite on price is not a meaningful comparison.
Can we change later?
Yes, and moving up is easier than moving down. Clean data and a readable export are what make any future move cheap.