What you'll learn, and how we support it

Which choices are permanent the moment a contract is created — whether it bills, in what currency, how collection works, and the effective date itself. The one that matters most is the mode. A contract as your term of record is reversible. A contract that bills is not, once it is activated. Most companies who say they just want their contracts in HubSpot want the reversible one and reach for the irreversible one by default, because it is the one that sounds finished.

The grain question, which sounds academic until it costs you. A multi-year agreement that lives in your system today as a chain of annual records will land as several separate agreements that end before they begin. The term belongs to the original deal, not to the records that accumulated around it. This is the migration error that is hardest to see and most expensive to unwind, because everything downstream inherits it quietly.

Line items on a live contract are never edited. You describe a change, and the platform computes what your contract becomes — so every amendment arrives with a proposal, a preview of the result, and a record of who agreed to it. Proration is a deliberate decision on each change rather than a setting somewhere, and a change timed to a renewal boundary avoids it altogether. You will know where those levers are and which ones your business should actually be pulling.

What the platform genuinely does here and what it does not. This matters more than any configuration detail, because it is the thing that either earns a second meeting with Finance or quietly ends the conversation. Overstate it once and you do not get to correct it later. You will know the boundary well enough to be precise about it rather than vague, which is the whole difference in that room.

Six specific seams — the points where an agreement gets re-keyed, reinterpreted, or silently dropped between the person who sold it and the person who bills it. Most teams can feel that this handoff is broken and cannot point at where. You will be able to point at it, with a count of how often it happens in your own business rather than an impression that it happens a lot.

Three things about sequencing that are counterintuitive enough that most rollouts get them wrong. Buy-in is not a phase at the end — each stakeholder buys in at a different moment, and collecting it all at launch produces the meeting where one objection invalidates six weeks of configuration. Testing cannot be terminal; for anything touching money it is three separate activities at three separate points. And accounts receivable trains before sales, always, because AR catches errors and sales creates volume — do it the other way and your first week of real data is generated by people whose mistakes nobody is positioned to catch.

The contracts layer is in beta and it moves. An endpoint changed underneath us mid-build. You will learn how to work on it anyway — pinning what you built against, reading the live portal rather than a stored copy of it, and telling the difference between the parts that are stable enough to commit to and the parts that are not yet.