Right-sizing the Microsoft 365 license mix means moving from a single-SKU enterprise default to a persona-based mix that maps users to the cheapest SKU that covers their actual entitlement need. The default Microsoft proposal — an E3 or E5 floor for the whole workforce plus mass-assigned add-ons — carries 20–30% structural overspend in most enterprises. The disciplined right-sizing exercise: define personas by entitlement need, allocate users to personas via HR and IT data, map personas to the cheapest qualifying SKU, and rationalise the add-on stack against utilisation telemetry. We walk the four steps and the recurring traps below.
Why right-sizing matters specifically in 2026
Right-sizing the Microsoft 365 license mix is the single highest-yield M365 cost exercise an enterprise can run in 2026. Three reasons it matters now in particular:
- The July 2026 price increase. Microsoft is raising M365 commercial pricing across E3, E5, Business Premium, and the Frontline tiers in July 2026. Any structural overspend gets multiplied by the new rate. See our July 2026 price increase analysis for the macro context.
- The E7 Frontier suite arrival. The new E7 SKU at $99 PUPM creates a fourth tier in the enterprise stack and reshuffles which SKU is right for which persona. The right-sizing exercise now has to consider E7 as a destination for the highest-value users.
- The Frontline F3 expansion. Microsoft has expanded the F3 entitlement scope in 2025–26, making it a defensible primary SKU for a broader set of operational and field-based personas than was true in 2023.
The exercise itself is straightforward but unglamorous. Most enterprises do not run it because the default Microsoft proposal sounds reasonable (an enterprise SKU for “all knowledge workers”), and because the data plumbing to allocate users to personas requires a real HR-and-IT data join. The recurring outcome when the exercise is run rigorously: 20–30% reduction in per-seat M365 spend.
Step 1: Define the persona model
A persona is a user archetype defined by entitlement need, not by job title. The seven enterprise personas we use as a starting framework in 2026:
| Persona | Description | Right SKU | Typical % |
|---|---|---|---|
| Knowledge worker (standard) | Office apps, Teams, OneDrive, SharePoint | E3 | 50–65% |
| Knowledge worker (security) | KW plus active need for P2 capability (admin role, PIM user) | E5 or E3 + targeted P2 | 5–15% |
| Power user | KW plus Power BI Pro, advanced compliance, Copilot daily | E5 | 10–20% |
| Executive / regulated | Highest entitlement, full compliance and protection stack | E5 or E7 | 1–3% |
| Frontline (deskless) | Shift-based, browser-only Office, basic Teams | F3 | variable |
| Kiosk (shared device) | Shared device, no per-user identity | F1 | variable |
| Contractor (limited) | External, time-bound, restricted scope | Business Basic or F3 | 1–5% |
The persona model is contractor-friendly: each entry maps to a SKU and a check — not to a job title. Two engineers with identical titles can sit in different personas based on actual entitlement need.
Step 2: Allocate users to personas via data, not opinion
The allocation data sources that work:
- HR system. Employee vs contractor flag, job role, work pattern (frontline vs office), department, location.
- Microsoft 365 Usage Analytics. Per-user activity on Office apps, Teams, SharePoint, OneDrive. Identifies dormant users and frontline-vs-knowledge-worker patterns.
- Copilot dashboard. Copilot MAU per user. Identifies the high-value-Copilot population for E5 or Copilot-bundled SKUs.
- Entra ID PIM activation report. Admin and privileged-access users who need P2.
- Power BI activity log. Power BI Pro authors vs viewers. Authors need Pro or E5; viewers in 2025+ can use the free viewer tier.
- Defender Compliance Score per user. Identifies users with active compliance entitlement value (E5 Compliance step-up earns out).
The allocation algorithm: for each user, default to the cheapest persona that covers their actual entitlement need, then escalate only when the data shows the higher persona is justified. Avoid the “default to E3 for everyone” pattern — it is a guaranteed overspend on frontline and contractor populations.
Frontline F3 includes web/mobile Office apps, Teams, Exchange, OneDrive (2GB), and SharePoint. For office workers who genuinely use web Office (which is most of the modern workforce, post-Copilot), F3 at $8 PUPM is defensible. The hold-out is the 50GB OneDrive entitlement difference and the desktop Office app rights. Audit which office workers actually need desktop Office apps in 2026 — in many roles, the answer is “not as many as you assume.”
Step 3: Map personas to SKUs and stress-test the math
For each persona-to-SKU mapping, stress-test against the alternatives:
- E3 + targeted add-ons vs E5. If the persona needs three or more E5-bundled features standalone, E5 is cheaper.
- E5 vs E5 + Copilot vs E7. E7 at $99 PUPM is structurally cheaper than E5 + Copilot ($57 + $30 = $87) only if the persona consumes the E7-specific entitlements (Sales Copilot, Service Copilot, Industry Cloud entitlements, advanced threat protection). See our E5 vs E7 comparison for the decision framework.
- F3 vs E3 for browser-only office workers. Cost gap is $4 PUPM — meaningful at scale.
- F1 kiosk vs F3 for shared-device frontline. Kiosk is per-device, F3 is per-user. The break-even depends on device-to-user ratio.
Step 4: Rationalise the add-on stack
Right-sizing the base SKU is half the exercise. The other half is rationalising the add-on stack against utilisation telemetry. The pattern recurs across every M365 add-on:
- Copilot for Microsoft 365. Size to the 90-day MAU population, not the workforce.
- Defender P2 step-ups. Size to security operations and admin populations.
- Entra ID P2. Size to the PIM and risk-policy population. See our Entra ID licensing analysis.
- Intune Suite. Size to the EPM-and-Cloud-PKI population. See our Intune licensing guide.
- Teams Premium. Size to meeting-host populations who actually use the premium meeting features.
- E5 Compliance / E5 Security step-ups. Size to populations with three or more justified P2 entitlements.
Anonymised case study: $4.2M annualised on a 22,000-seat estate
A 22,000-employee global manufacturer ran a default M365 mix of E3 for 18,000 office workers, E5 for 4,000 leadership and engineering staff, Frontline F1 for 6,000 plant workers, and mass-assigned Copilot ($30), Entra ID P2 ($9), Defender for Office 365 P2 step-up ($3), and Teams Premium ($10) across all 22,000 office and leadership users. Annual M365 stack: $24.6M. Persona-based right-sizing redistributed: 14,200 on E3, 3,800 on F3 (browser-only office and field engineering populations migrated from E3), 4,000 on E5 retained, 6,000 on F1 retained. Add-on rationalisation: Copilot to 6,800 (MAU-validated population), Entra ID P2 to 700 (admin and PIM-active), Defender for Office 365 P2 step-up dropped entirely (Defender for Office 365 P1 now bundled in E3 covered 95% of needs, the 5% upgrade scenario moved to E5 Security step-up for 1,200), Teams Premium to 1,800. New annual M365 stack: $20.4M. Recurring annualised saving: $4.2M. The persona model gave the customer a defensible framework for the next renewal cycle and the next anniversary true-up.
Right-sizing the Microsoft 365 license mix is the foundational M365 cost discipline for 2026. Run it once rigorously, rebuild the persona model annually against HR and telemetry data, and the renewal becomes a managed conversation rather than a defensive one. Pair this with the EA tier-collapse renewal strategy and the result is the lowest defensible per-seat M365 cost in the market.