Cloud DLP Licensing: Three Incompatible Cost Models
Cloud DLP licensing is one of the hardest categories to benchmark because the three hyperscaler-aligned platforms — Microsoft Purview, Google Cloud Sensitive Data Protection, and AWS Macie — are priced on three entirely different mechanics. Microsoft bills per user, Google and AWS bill per byte scanned. Comparing them on a single number is impossible; comparing them on total annual cost for your specific data estate is the only meaningful exercise, and it is exactly the comparison most procurement teams never run.
Microsoft Purview Data Loss Prevention is bundled into Microsoft 365 E5 at roughly $57–$60 per user per month, or sold as the standalone Purview suite at about $12 per user per month ($144 per user per year). For non-Microsoft data sources, Purview adds a pay-as-you-go meter — roughly $0.0002 per 1 kB per day for text storage and $1,000–$4,000 per terabyte for heavy discovery scans. Google and AWS abandon per-user pricing entirely: AWS Macie charges per gigabyte of S3 data inspected after a 1 GB monthly free tier, and Google charges on total bytes processed, with the published pricing page now sales-led rather than self-serve. Our broader analysis of the cloud security and compliance contract stack shows DLP routinely consuming 8–12% of the security budget once endpoint and email channels are added.
Microsoft vs Google vs AWS: What You Actually Pay
The table below translates each vendor's model into an annual figure for a representative 5,000-user enterprise with a mixed Microsoft 365, S3, and BigQuery data estate. The numbers are illustrative of the model, not a quote — the point is the 3–4x spread that the pricing mechanics create depending on data volume and existing licence position.
| Platform | Licensing Mechanic | Indicative Annual Cost (5,000 users) | Best Fit |
|---|---|---|---|
| Microsoft Purview DLP (in E5) | Per user, bundled | Marginal if already on E5; $0.72M standalone Purview suite | Microsoft 365-centric estates |
| Microsoft Purview (PAYG, non-M365) | Per kB/day + per TB scanned | $1,000–$4,000 per TB scanned | Hybrid data sources |
| Google Sensitive Data Protection | Per byte processed | Volume-tiered; capacity subscription for predictability | GCP/BigQuery data lakes |
| AWS Macie | Per GB inspected (1 GB free/mo) | Scales with S3 footprint; can spike on full re-scans | AWS S3-heavy estates |
| Independent DLP (Forcepoint, Symantec, Netskope) | Per user/year | $34–$71 per user; $75–$150 full-stack | Multi-cloud, channel breadth |
The single most expensive DLP mistake we see is paying for Microsoft 365 E5 primarily to unlock Purview DLP, then running a second consumption-based scanner across the same data in AWS or GCP. Enterprises duplicating coverage this way pay twice for one control — often $400,000–$900,000 a year of avoidable overlap on a 5,000-seat estate.
The 2025–2026 Buyer Traps
The first trap is the E5 justification fallacy. Microsoft account teams position Purview DLP as a reason to upgrade the entire seat base from E3 to E5 — a roughly $24 per user per month uplift. For a 5,000-user enterprise that is $1.44M a year, of which DLP is only one component. If fewer than 40% of users handle regulated data, the standalone Purview suite at $144 per user per year applied to that subset is dramatically cheaper than a blanket E5 upgrade. The DLP-specific lesson is to licence the control to the population that needs it, not the whole organisation.
The second trap is consumption blindness on Google and AWS. Because Macie and Google Sensitive Data Protection bill on bytes scanned, a poorly scoped discovery job — full-bucket re-scans on every schedule rather than incremental scans — can multiply the bill 5–10x with no change in protection. AWS itself documents that costs are driven by total data inspected, prorated per day; the contract should commit the vendor to incremental scanning and a monthly spend ceiling. This is the same consumption-governance discipline we apply across the cloud WAF and broader cloud security portfolio.
The third trap is scope creep into DSPM and Insider Risk. Microsoft increasingly bundles DLP with Data Security Posture Management and Insider Risk Management inside the Purview suite, and the consumption meters for these adjacent modules are easy to switch on and hard to switch off. Treat each module as a separate line item with its own business case.
DLP Overlap With Your Existing Stack
Most enterprises already own DLP capability they are not counting. Microsoft 365 E3 includes baseline DLP policies for Exchange, SharePoint, and OneDrive; Microsoft Defender for Cloud Apps adds DLP for sanctioned SaaS; and any CASB platform you licence carries inline DLP inspection. A genuine DLP rationalisation starts by mapping coverage across email, endpoint, SaaS, and cloud-storage channels, then removing the duplicated tools. We routinely find two or three overlapping DLP engines inspecting the same email flow, each separately licensed.
The decision is rarely "which single DLP wins". For Microsoft 365-centric organisations, Purview's bundled coverage is usually the anchor and the consumption scanners are scoped narrowly to the data that actually lives outside Microsoft. For multi-cloud estates with significant S3 and BigQuery data, an independent platform that prices per user can be cheaper and simpler than running two hyperscaler meters in parallel. Either way, the contract must pin down scanning scope, re-scan frequency, and a spend cap.
Mapping DLP Spend to Regulatory Obligations
DLP is one of the few security controls procurement can tie directly to a regulatory mandate, which makes the business case easier to defend but also easier for vendors to inflate. GDPR Article 32, HIPAA, PCI DSS 4.0, and the EU's DORA all expect demonstrable controls over sensitive data movement — but none of them require a specific product tier. The negotiating risk is letting a vendor map your compliance obligation onto its most expensive bundle. A PCI scope covering a few hundred users handling cardholder data does not justify licensing DLP across 5,000 seats.
The disciplined approach is to scope DLP to data classification, not headcount. Classify where regulated data actually resides — typically 15–30% of an enterprise's repositories hold 80% of its sensitive records — and licence inspection against those repositories and the users who touch them. This keeps consumption scanners pointed at a defined data perimeter rather than the entire estate, and gives auditors a defensible, documented control boundary. The same classification work also feeds your shared-responsibility mapping, since it clarifies which data the cloud provider secures and which you must inspect yourself.
Negotiation Levers That Work
On Microsoft, the lever is the E3-to-E5 step-up timing. Microsoft discounts E5 most aggressively when it is tied to an Enterprise Agreement renewal or a Copilot commitment; isolate the DLP requirement and refuse to let it drive a blanket upgrade. Insist on the standalone Purview suite price for non-E5 users as a benchmark, even if you ultimately buy E5 — it caps how much "DLP value" Microsoft can attribute to the upgrade.
On AWS and Google, the lever is the capacity commitment. Both vendors offer predictable-cost or capacity-subscription pricing that trades a usage commitment for a lower effective rate; Google explicitly introduced capacity subscriptions to replace volatile on-demand billing. Negotiate the committed volume against your measured monthly scan baseline plus 20% headroom — not the vendor's forecast — and require a true-down right if your data footprint shrinks. Across every option, the contract should include a benchmarking clause; our price benchmarking research consistently shows first quotes 20–30% above achievable rates. To pressure-test your specific DLP estate against market pricing, request a confidential briefing.