I think most SaaS companies misunderstand what an enterprise customer is paying for.
It isn’t the product in the demo. It’s the product the buyer believes will exist in three years, and the vendor’s willingness to keep building it in their direction.
The roadmap is the thing being bought. The product is the down payment.
The argument in one line
Enterprise buyers pay for where your product is going. Break a promise in Kano’s must-be column and you reprice every contract you’ve already signed.
Atlassian has spent the last twelve months proving it. Here is the timeline:
- September 2025: Atlassian announced end of life for its Data Center products, effective March 28, 2029.
- March 30, 2026: Sales to new customers stopped.
- March 30, 2028: Existing customers lose new licences, Marketplace app purchases and expansions.
- The stated reason: 99% of its more than 300,000 customers are already on cloud.
Three years before that announcement, the company’s position on the same product was very different.
“No plans to end of life or end of support for our Data Center offering.”
Atlassian, 2022, as recalled by its own community
Kano explains why the roadmap outsells the product
Most product leaders know Noriaki Kano’s model:
- Must-be features earn nothing when present and enrage when missing.
- Performance features satisfy in proportion to how much you deliver.
- Attractive features delight when they show up and cost nothing when they don’t.
The part people forget is time. Kano’s categories decay. Today’s delighter becomes next year’s performance feature and, soon after, a basic nobody thanks you for. Enterprise buyers understand this without ever naming it. They know the feature that wins the bake-off will look ordinary in eighteen months, so what they’re really judging is whether you’ll keep producing the next one. That’s a roadmap question, not a product question.
The Kano decay rule
Today’s delighter is next year’s performance feature and the year after’s basic. Buyers price that in. They are buying your next release, not this one.
Here is where I think most roadmaps get it backwards.
They’re built as a parade of attractive features. An assistant this quarter, an agent next quarter, a keynote moment every six months. But the roadmap statements customers actually build a business case on sit in the must-be column:
- Will it keep running where my regulator says it must?
- Will the licensing model hold?
- Will my integrations survive the next platform shift?
Nobody applauds those commitments. Nobody puts them in a launch video. Break one and Kano’s asymmetry arrives at full force.
That’s what Data Center was for Atlassian’s remaining 1%. On-premise continuity was a must-be, and the 2022 statement made it an explicit one. For those customers, forced cloud lands as what Kano called a reverse attribute. The vendor presents it as progress, and the customer experiences it as loss. No quantity of cloud delighters makes up for it.
A roadmap change reprices every contract you’ve already signed, without anyone touching an invoice.
The partner economics
A customer makes a three-year bet. A partner builds a business on it. They hire for a deployment model, certify people, build apps and IP, and sell annuity services against it. After the announcement, one partner in Atlassian’s community called it really sad news for the companies they support, after investing so much time only to watch Data Center get pulled.
A Data Center partner’s P&L, before and after September 2025
- Admin and licence renewals: still recurring, but now with an end date
- Migration work: a spike for two or three years, then a fall-off
- Marketplace apps: no new Data Center submissions accepted after December 16, 2025
A migration is a project.
Projects don’t renew.
Broadcom did a harsher version with VMware after closing in late 2023. It ended perpetual licences and rebuilt the partner programme around a much smaller set of resellers. Plenty of partners kept the software and lost the business model.
Your product is what the customer owns. Your roadmap is what they bought.
“Roadmaps aren’t commitments.” Legally true. Commercially irrelevant.
The strongest objection deserves a straight answer.
Every roadmap deck opens with a safe harbour slide, and every contract says the purchase doesn’t depend on future functionality. A vendor must keep the right to change direction. Otherwise it ends up carrying two codebases for 1% of customers and slowing down the other 99%. By that logic Atlassian made the right call, and I agree it probably did.
But the objection defends the decision, not the 2022 sentence. The three-year runway was responsible product management. The damage came from a must-be promise made casually.
Where I don’t have a clean answer
Publishing honest deprecation dates early tells competitors and wavering customers something you’d rather keep quiet. It will cost some deals. I still think that’s the cheaper loss.
What I’d say to a SaaS CEO this week
Pull the last four quarters of customer-facing roadmap decks, plus every “we have no plans to” your executives said on stage, in a forum or on a renewal call. Sort each statement into its Kano column:
- Attractive statements are marketing, and they can move.
- Must-be statements are contract clauses in everything but name. Your largest customers and partners have already budgeted against them. Correct any you won’t honour, with runway, before someone screenshots the old version.
- Whoever approves a must-be commitment should also sit in the renewal where it gets tested.
Your customers signed your roadmap long before they signed your contract.
They just read the must-be lines more carefully than you wrote them.
