Microsoft 365 E3 is $36 per user per month. E5 is $57. F3 is $8. F1 is $2.25. The decision is not which is cheapest, it is which features each persona actually consumes, because picking the wrong tier across a 10,000-seat estate routinely overspends by $2M to $4M per year. This page is the comparison reference: what each tier includes, what the realistic cost is by user type, and the right way to mix tiers across a hybrid workforce.
Inside This Comparison
Microsoft 365 price per tier (2026)
| SKU | List per user/month | Target user | Annual cost (5K users) |
|---|---|---|---|
| Microsoft 365 F1 | $2.25 | Frontline (no email) | $135,000 |
| Microsoft 365 F3 | $8.00 | Frontline (with email) | $480,000 |
| Microsoft 365 E3 | $36.00 | Information worker | $2,160,000 |
| Microsoft 365 E5 | $57.00 | Information worker requiring security, compliance, analytics, voice | $3,420,000 |
| Office 365 E3 (legacy) | $23.00 | Office only, no Windows/EMS | $1,380,000 |
| Office 365 E5 (legacy) | $38.00 | Office with security, no Windows/Intune | $2,280,000 |
| Microsoft 365 Copilot | $30.00 add-on | Requires E3 or E5 base | $1,800,000 (5K users) |
List prices are reference. Enterprise Agreement (EA) and CSP discounts typically run 10 to 25 percent off list depending on commitment size and existing Microsoft relationship. The cost differential between tiers, however, is the decision driver, not the discount.
What is in E3, E5, F3, F1
The four tiers are distinguished by three feature stacks: Office applications, Windows 11 Enterprise, and Enterprise Mobility + Security (EMS, now branded as Entra and Intune). E5 adds security, compliance, voice, and analytics on top of E3. F3 is a constrained version of E3 with capped storage and no desktop Office. F1 is F3 minus email and Office Web.
| Capability | F1 | F3 | E3 | E5 |
|---|---|---|---|---|
| Outlook email (50 GB mailbox) | No | Yes (2 GB) | Yes (100 GB) | Yes (100 GB) |
| OneDrive storage | 2 GB | 2 GB | 1 TB+ | 1 TB+ |
| SharePoint | Read | Read/limited write | Full | Full |
| Teams (chat + meetings) | Yes | Yes | Yes | Yes |
| Office desktop apps | Web only | Web only | Yes | Yes |
| Windows 11 Enterprise | Web only | Web only | Yes | Yes |
| Intune device management | Yes | Yes | Yes | Yes |
| Entra ID P1 | Yes (Frontline) | Yes (Frontline) | Yes | P1 |
| Entra ID P2 | No | No | No | Yes |
| Defender for Endpoint P1 | No | No | Yes | P2 |
| Defender for Office 365 P2 | No | No | No | Yes |
| Defender for Identity | No | No | No | Yes |
| Defender for Cloud Apps | No | No | No | Yes |
| Purview Information Protection P2 | No | No | P1 | P2 |
| Purview eDiscovery (Premium) | No | No | Standard | Premium |
| Purview Insider Risk Management | No | No | No | Yes |
| Power BI Pro | No | No | No | Yes |
| Teams Phone (PSTN-capable) | No | No | No | Yes (calling plan extra) |
| Audio Conferencing | No | No | Add-on | Yes |
F3 versus E3 decision
F3 is correct for users who do not need desktop Office, do not need a 100 GB mailbox, and do not need more than 2 GB of personal cloud storage. The canonical F3 user is a retail associate, warehouse worker, factory operator, field technician, or branch teller: someone who needs basic communication, Teams chat, and access to applications through a shared device or a kiosk, not someone authoring Word documents and Excel models all day.
Where F3 commonly breaks down is when an organisation gives F3 to staff who occasionally need to author longer documents. The 2 GB OneDrive cap is reached quickly. Web-only Office is functional but slow for heavy users. The "F3 for everyone in stores" decision often results in shadow IT (personal Dropbox, Gmail) within six months. The right test is not "could this user technically use F3" but "will this user be productive on F3 for the next three years given likely role evolution."
| User profile | Recommended SKU | Why |
|---|---|---|
| Retail store associate (shared kiosk) | F1 or F3 | Communication and rostering only; no document authoring |
| Warehouse picker (handheld) | F1 | App-only access; no email needed |
| Field service technician | F3 | Mobile email + Teams + work app access |
| Bank branch teller | F3 | Email needed, no heavy Office authoring |
| Manufacturing supervisor (occasional reporter) | E3 | Excel reporting and SharePoint authoring needed |
| Customer service rep (CRM-based) | F3 if no Excel, E3 if Excel-heavy | Depends on Excel and SharePoint usage |
E3 versus E5 decision
E5 adds $21 per user per month to E3, or $252 per user per year. Across 10,000 users that is $2.52M per year. The decision rests on whether the organisation would otherwise buy the E5 component products as add-ons or from third parties. If yes, E5 is almost always cheaper than the sum of the parts. If no, E5 is overspend.
| E5 component | Standalone add-on price | Replaces (typical) |
|---|---|---|
| Defender for Endpoint P2 | $5.20 per user/month | CrowdStrike Falcon ($8 to $15) |
| Defender for Office 365 P2 | $5.00 per user/month | Mimecast / Proofpoint ($6 to $12) |
| Defender for Identity | $5.50 per user/month | Specialised AD/identity threat detection |
| Defender for Cloud Apps | $5.00 per user/month | Netskope / Zscaler CASB ($4 to $10) |
| Entra ID P2 | $9.00 per user/month | Okta Identity Governance ($6+) |
| Purview Information Protection P2 | Included in E5 Compliance | Symantec DLP / Forcepoint |
| Purview Insider Risk Management | Included in E5 Compliance | Specialised insider risk tooling |
| Power BI Pro | $10 per user/month | Tableau Creator ($75) / Looker ($30+) |
| Teams Phone (PSTN) | $8 per user/month + calling plan | RingCentral / Zoom Phone |
| Audio Conferencing | $4 per user/month | Zoom / WebEx audio bridge |
Adding up the major E5 components at standalone add-on rates: roughly $45 to $55 per user per month if all components were purchased separately. E5 sells the bundle at $21 above E3. For an organisation that would otherwise buy three or more of these components, E5 is cheaper. For an organisation that only needs Defender for Endpoint or only needs Power BI, the right move is the standalone add-on, not the full E5 upgrade.
The common mistake: Buying E5 for the security stack but never deploying it. Defender for Endpoint P2, Defender for Cloud Apps, and Purview Insider Risk Management all require operational deployment to deliver value. Buying E5 and continuing to use CrowdStrike, Netskope, and Mimecast delivers zero E5 value at full E5 cost. The pre-purchase requirement is a security architecture decision: are we replacing the third-party stack with Microsoft, or augmenting it. If augmenting, E5 is the wrong purchase.
How to mix tiers across a workforce
Most enterprises run a mix of F3, E3, and E5 rather than putting the whole workforce on a single tier. The mix economics are usually as follows:
| Workforce segment | Typical mix | Annual cost (10K-seat org) |
|---|---|---|
| All E5 | 10,000 × E5 | $6,840,000 |
| All E3 | 10,000 × E3 | $4,320,000 |
| Tiered mix (typical enterprise) | 3,000 F3 + 5,500 E3 + 1,500 E5 | $3,294,000 |
| Saving versus all-E5 | $3,546,000 per year | |
| Saving versus all-E3 | $1,026,000 per year |
The tiered mix outperforms the single-tier deployments because most enterprises have meaningful frontline populations and meaningful security-sensitive populations. Putting frontline workers on E3 wastes $336 per user per year. Putting executives and security teams on E3 forces add-on purchases that exceed the E5 differential.
The right segmentation framework is: identify the 10 to 20 percent of users who actually use E5 differentiators (executives, finance, legal, security operations, M&A advisory, board secretariat, compliance officers, HR business partners with sensitive data exposure). Put them on E5. Put the rest of the information worker population on E3. Put frontline staff who need any email on F3 and frontline staff who never use email on F1.
Where Copilot fits
Microsoft 365 Copilot is a $30 per user per month add-on, available on top of any E3 or E5 base licence (and now F-SKU bases in certain configurations). Copilot is not included in any base SKU. Per-user list cost of E5 + Copilot is $87 per month or $1,044 per year. For 5,000 Copilot seats that is $5,220,000 per year.
Copilot adoption needs to be modelled separately from the licensing decision. Microsoft's published productivity uplift estimates (25 to 30 percent) are derived from controlled studies on highly engaged early-adopter users. Realised productivity uplift in broader enterprise rollouts during 2024 to 2025 has been measured at 5 to 15 percent for typical information workers and approaching zero for users who do not naturally use Copilot's authoring or summarisation capabilities daily.
The right Copilot rollout pattern: pilot with 200 to 500 high-frequency Office users, measure actual workflow lift over six months, calibrate licence quantity to demonstrated lift, then expand. See Microsoft Copilot licensing guide for the full ROI framework.
NCE term and cancellation rules
Microsoft New Commerce Experience (NCE) replaced the legacy CSP licensing in early 2022 and reshaped what flexibility customers have to adjust seat counts mid-term. NCE term commitments are monthly, annual, or triennial. Pricing is lowest for triennial (locked-in) and highest for monthly (no commitment).
| NCE term | Price flexibility | Cancellation rule |
|---|---|---|
| Monthly | 20 percent premium over annual | Cancel any time; no penalty |
| Annual | Baseline price | Cancel within 7 days of new term or anniversary; otherwise locked |
| Triennial | 5 to 10 percent below annual | Cancel within 7 days; otherwise locked for three years |
The 7-day cancellation window for annual and triennial NCE commitments is the critical constraint. After day 7, the seat count is locked for the remainder of the term. Organisations that mis-estimate seat requirements at NCE renewal pay for unused seats for 12 to 36 months. The remediation is to model demand carefully before renewal, use shorter terms during periods of organisational uncertainty (M&A, restructure, workforce reduction), and to split workforces across multiple NCE commitments rather than committing the entire estate on a single term.
Negotiation levers for each tier
Enterprise Agreement (EA) and CSP renewals offer the following levers, in approximate order of customer leverage:
- Mix shift: moving a portion of the workforce down a tier (E5 to E3, E3 to F3) typically generates 15 to 30 percent of the total savings in a Microsoft renewal. The lever is fully under customer control.
- True-down rights: reduction of unused licences mid-term, allowed under EA Subscription but not under traditional EA. Negotiating explicit true-down provisions is essential for organisations expecting workforce reduction.
- Annual versus triennial commitment: triennial discount is real (5 to 10 percent) but flexibility cost is significant. The right answer depends on organisational stability.
- Add-on substitution: moving from E5 to E3 + targeted add-ons (Defender for Endpoint P2, Power BI Pro, etc.) where only specific E5 components are used.
- Multi-year price protection: EA price protection caps annual price increases at agreed rates (typically 5 to 8 percent ceiling). Without protection, Microsoft list increases of 9 to 15 percent per year are now standard.
- Co-terming: aligning EA and CSP renewal dates to consolidate negotiation leverage and avoid mid-cycle pricing surprises.
For the complete Microsoft EA negotiation framework, see our Microsoft EA complete guide. For Copilot-specific commercial structures, see our Copilot licensing guide. For CSP NCE specifics, see Microsoft NCE pricing. For the broader cost reduction framework, see reduce Microsoft spend.