Your NetSuite Renewal Has Two Kinds of Line Items

Open any NetSuite account that has been running for five years and you find a stack of subscriptions bolted onto it. Tax. EDI. Shipping rates. Document capture. Expense. A reporting layer. A connector or two.

If you can name every one of them and what each costs annually without looking it up, you're unusual.

Each one was a reasonable decision when it was made. And subscriptions don't really get canceled. They get renewed, because renewing takes no meetings and canceling takes several.

I wrote a few months ago about replacing a $12,000/year document automation SuiteApp with a one-time build. The math in that piece was specific to one category. The question underneath it applies to the whole stack, and the honest version of that question is more interesting than "build everything yourself."

Because most of what's on your renewal should stay there.

The one question

For each line item on that list, ask:

Who has to keep doing work for this to keep working?

That's it. That single question sorts the entire list into two piles, and the two piles have completely different economics.

Pile one: you are buying someone else's continuous labor

Some subscriptions exist because the outside world keeps moving and somebody has to keep up with it.

Tax engines. There are more than 12,000 sales and use tax jurisdictions in the United States. One tax software provider's year-end tracking counted 681 combined rate changes and new rates in 2025, the third highest annual total in a decade, plus 335 brand new city, county, and district taxes. You are not paying for the calculation. The calculation is arithmetic. You are paying someone to track 12,000 jurisdictions forever so you never have to think about it.

EDI. Trading partners revise their specs on their own schedules. The EDI providers say it plainly in their own published material: changes don't always arrive with advance notice, the notification process isn't standardized, and staying current means actively monitoring every partner relationship you have. Miss one and you get chargebacks. That's not a build. That's a subscription to somebody watching portals.

Carrier rating. The major parcel carriers each raised base rates an average of 5.9% for 2026, the third straight year at that headline number, and both expanded the cubic volume thresholds that trigger oversize and additional handling surcharges. Rates and rules change annually at minimum. Somebody has to absorb that.

Payment gateways, bank feeds, and payroll tax tables sit in this pile for the same reason. The work is external, continuous, and not yours.

AI has changed nothing important here. You can build a tax calculator in an afternoon now. You still can't maintain 12,000 jurisdictions. If you build in this pile, you haven't saved money, you've bought yourself an obligation.

Keep paying.

Pile two: the vendor already did the work and now bills you monthly for it

The other pile is different. Here the vendor built something once, and the logic they built has been sitting still ever since.

Document capture and data entry. Approval routing. Notification engines. Field-level validation. Report and dashboard packs. A connector between two systems that are both under your control. Utilities that reorganize your own data.

None of that changes because the world changed. It changes when your business changes, and when your business changes you're paying for a configuration request anyway.

You are not subscribing to ongoing work. You are amortizing a build. Indefinitely, at a price that usually goes up with your volume even though the cost of serving you doesn't.

That's the pile worth looking at.

Three follow-up questions

If a line item lands in pile two, pressure-test it before you do anything:

If the vendor vanished tomorrow, would the logic still be correct a year from now? If yes, you're renting a build.

Does the price scale with your volume? Per-document and per-transaction pricing on fixed logic is the clearest tell there is. The work doesn't get harder as you grow. The bill does.

Could you describe the rules to a developer in an afternoon? If yes, it isn't a platform. It's a configuration with a login page.

What actually changed

Nothing about this framework is new. Build versus buy has been the same argument for thirty years.

What changed is the price of one side of it. The cost of building something custom has come down materially. The cost of owning it has not moved at all.

So the category that shifted from "buy" to "build" is narrow and specific: the things where the logic sits still. That's it. That's the whole shift. Anyone telling you AI means you should bring your tax engine in house is selling something.

The part that doesn't go in the proposal

If you move something into pile two and build it, you now own it.

NetSuite ships two major releases a year, on top of monthly minor releases. Anything running in your account has to survive them. A subscription quietly includes that labor. A build doesn't, unless somebody planned for it.

So don't move a line item unless you have a real answer to who owns it in year three. If the answer is "the consultant who built it, probably, if he's still around," you've traded a predictable cost for an unpredictable one. Sometimes that's still the right trade. Make it on purpose.

How to actually run this

Pull the list. Every recurring line item touching your NetSuite account, including the ones buried in a departmental budget.

Put the annual number next to each one. Not monthly. Annual. Monthly pricing exists to make the number feel small.

Sort with the one question. Who has to keep working for this to keep working.

Then look at the top three items in pile two by annual spend, and ignore everything else. You're looking for the one that made sense in 2021 and quietly stopped making sense, and it's almost always one of the big three.

Most of the list will stay exactly where it is. That's the expected outcome, not a failure of the exercise.

The default is renewal. That's not a criticism, it's just how the calendar works. All I'm suggesting is that once a year the default should have to survive one question.

If one of your pile two line items is coming up for renewal and you want a second opinion before it auto renews, that's a twenty minute conversation.