Atlassian Data Center goes fully read only on March 28, 2029. Restructure the user baseline before you migrate and an 1,850 user estate cuts year one cloud cost by roughly 33 percent in our benchmark scenario.
Prepared by Redress Compliance · June 2026 · Representative Atlassian estate scenario (benchmark scenario, not a quote)
Atlassian controls the migration calendar, the price reference points, and the audit posture. Server support ended in February 2024. New Data Center purchases stop on March 30, 2026, and the platform becomes read only on March 28, 2029. The vendor uses those dates to push fast, list price cloud migrations on whatever user count sits in your current contract.
The buyer side move is to flip that control. Audit the active baseline first. In our benchmark estate a named count of 2,400 Data Center users falls to 1,850 active users once dormant and duplicate accounts are removed. At Cloud Premium that single step removes roughly 80,000 dollars of annual list before any discount.
Then stack the levers. The cloud loyalty discount reaches up to 60 percent for estates above 1,001 users, dual licensing covers the on premises term for up to 12 months, and the cloud migration trial can run free for the remaining term of the equivalent paid license. Combined with marketplace app re entitlement, the negotiated year one cost in our scenario falls 33 percent below list.
The trap is the renewal. The loyalty discount does not survive renewal unless you write a renewal floor and an uplift cap into the order. This paper gives the edition map, the baseline method, the five clauses, the discount benchmarks, the counter moves, and the BATNA you bring to the table.
Atlassian is closing the on premises era on a fixed clock. Knowing the dates is the first source of leverage, because the vendor sells urgency against them.
Read the official terms before you negotiate against them. The Atlassian end of support pages and the published cloud price list are the reference points the account team will quote, so cite them back.
A full migration takes 18 to 24 months of planning, approval, and execution. During that window many estates run Data Center and Cloud at the same time, paying twice. The longer the parallel run, the weaker your position, so the calendar is a commercial variable, not just a technical one.
Atlassian Cloud sells in three paid editions. The account team defaults you to Premium or Enterprise. Map features to need before you accept the tier, because most seats do not use the Enterprise controls.
| Edition | What it adds | Who actually needs it | Benchmark list, per user per year |
|---|---|---|---|
| Cloud Standard | Core Jira or Confluence, user management, 250 GB storage | General contributor seats with no compliance overlay | 90 dollars |
| Cloud Premium | Unlimited storage, advanced admin, SLAs, sandbox, release tracks | Platform teams that need admin depth and uptime SLAs | 145 dollars |
| Cloud Enterprise | Data residency, multiple instances, Atlassian Guard, unlimited automation | Regulated estates needing residency and central security | 175 dollars |
Benchmark scenario, not a quote. Atlassian list rates fall as user volume rises and Enterprise is custom above 1,000 users with 15 to 25 percent negotiation room.
Most contributor seats do not use Enterprise controls. Mixed edition allocation, where only the seats that need residency and central security sit on Enterprise, beats a flat Enterprise buy.
The baseline is the single most valuable number in the negotiation. Atlassian prices on your user count, so an un audited count means you pay for accounts no one uses.
In our benchmark estate this method cuts a named count of 2,400 to an active baseline of 1,850. That is a 23 percent reduction, inside the 15 to 30 percent range we see across engagements.
| Cloud Premium basis | Users | Annual list at 145 dollars |
|---|---|---|
| Named contract count | 2,400 | 348,000 dollars |
| Audited active baseline | 1,850 | 268,250 dollars |
| Removed before any discount | 550 | 79,750 dollars |
Restructure before discount. A loyalty discount applied to 2,400 seats just locks in overspend on 550 accounts no one uses.
Atlassian funds the transition through three distinct mechanisms. Treat them as separate negotiable levers, not one bundle the account team hands you.
Typical dormant and duplicate share stripped before pricing. Range 15 to 30 percent across the engagement file.
When loyalty, dual licensing, and competitive tension are combined on an audited baseline. Range from the engagement file.
In our benchmark estate a 40 percent loyalty discount on the 268,250 dollar active baseline takes the core cloud subscription to 160,950 dollars in year one. The discount applies to the smaller, audited number, not the inflated one.
Marketplace apps are priced on the user tier of the product they extend, even for users who never open the app. A heavily extended Jira estate carries app cost that can rival the core subscription.
Re entitle apps deliberately. Drop apps that duplicate native cloud features, consolidate overlapping tools, and put the surviving apps out to competitive tension. In our scenario disciplined re entitlement holds app cost to 147,500 dollars list and 118,000 dollars negotiated.
| Year one cost stack, 1,850 user estate | List | Negotiated |
|---|---|---|
| Core cloud subscription (Premium) | 268,250 dollars | 160,950 dollars |
| Marketplace apps, re entitled | 147,500 dollars | 118,000 dollars |
| Migration services (trial and dual license) | 0 dollars | 0 dollars |
| Year one total | 415,750 dollars | 278,950 dollars |
Year one total falls from 415,750 dollars list to 278,950 dollars negotiated, a 136,800 dollar reduction. Apps are the second largest line and the most overlooked.
The loyalty discount wins year one. The renewal is where the saving is lost, because the discount does not carry forward unless the order says it does. This is the most expensive gap in a standard cloud migration.
| Core subscription path, from 160,950 dollar base | Uncapped at 15% | Capped at 5% |
|---|---|---|
| Year 1 | 160,950 dollars | 160,950 dollars |
| Year 2 | 185,093 dollars | 168,998 dollars |
| Year 3 | 212,857 dollars | 177,448 dollars |
| Three year total | 558,900 dollars | 507,396 dollars |
The uplift cap alone saves 51,504 dollars over three years in our scenario, on the core line, before apps. Across the full stack the gap is larger.
Numbers match the table exactly. Year 1 is identical. The cap diverges from year 2 and compounds.
The standard reseller pitch is to migrate early and lock the loyalty discount before it shrinks. We disagree. In most of the estates we have benchmarked the loyalty discount was applied to an un audited user count and did not survive the first renewal, so the headline saving evaporated inside 18 months. The discount is real, but on the wrong number it just funds 550 dormant seats and a renewal that snaps back to list. The buyer side move is to audit the active baseline first, take the discount on the smaller number, and refuse to sign without a renewal floor and an uplift cap. Speed favors the vendor. Sequence favors you.
A loyalty discount with no renewal floor is a one year coupon, not a price. Atlassian keeps the calendar. You keep the baseline and the clauses.
Atlassian uses a consistent set of tactics on migration deals. Each has a buyer side counter. State the counter before the vendor states the tactic.
| Vendor tactic | What it costs you | Buyer side counter move |
|---|---|---|
| Price on the current contract count | You pay for dormant and duplicate seats | Bring the audited active baseline first |
| Default the whole estate to Enterprise | You pay residency and security premiums on every seat | Mixed edition allocation by actual need |
| Bundle apps into the headline number | App overpricing hides inside the total | Price core and apps as separate lines |
| Quote the loyalty discount without a renewal floor | The discount disappears at renewal one | Renewal floor and uplift cap in the order |
| Push the deal to vendor fiscal year end | You sign on the vendor calendar, not yours | Anchor timing to your Data Center renewal date |
Your alternatives are credible until the read only date. They include extending Data Center toward 2028 and 2029, and competitive platforms for parts of the estate.
Export every instance, build the audited active baseline, inventory marketplace apps, and lock the internal number before the vendor sees a figure.
Stack loyalty, dual licensing, and competitive tension on the audited baseline. Draft the five clauses and the side letter. Time the trial against the Data Center renewal date.
Run the migration trial, cut over by edition, and sign the order with the renewal floor and uplift cap in writing. Confirm app re entitlement at the audited tier.
No, not unless you write it in. The loyalty discount applies to the initial term. At renewal the price reverts toward list unless the order states a renewal floor or a fixed effective rate. This is the single most common gap we fix.
Dual licensing waives the on premises cost for up to 12 months. The cloud migration trial can run for the remaining term of the equivalent paid Data Center license. Timing both against your renewal date maximizes the free parallel window.
Apps bill on the user tier of the product they extend, not on the number of people who use the app. A 2,400 seat Jira charges every app at the 2,400 tier. Restructuring the baseline lowers app cost at the same time as the core subscription.
Usually not. Most contributor seats never use the Enterprise residency and central security controls. A mixed edition allocation, with only the seats that need those controls on Enterprise, costs less than a flat Enterprise buy.
A paid Data Center extension toward the 2028 and 2029 dates, paired with competitive alternatives for parts of the estate. A credible BATNA is what makes the renewal floor and uplift cap negotiable rather than optional.
We sit on your side of the table as independent, buyer side advisors. We never take vendor commissions. Morten Andersen and the team benchmark your Atlassian position, build the strategy, and support the renewal end to end.
Recommendation: audit the baseline, then take the discount on the smaller number.
We are glad to tie a meaningful part of the fee to delivered value.
Bring us your Data Center renewal quote, your cloud migration proposal, and your admin console user export. Contact us and we will benchmark it against the engagement file and show you the defended number. Benchmark ranges: Redress Compliance advisory engagement file, 2024 to 2025.