Wealth DeskplusEnterprise Wealth Tech
Comparison

Splunk vs Microsoft Sentinel vs Google SecOps: Which SIEM Should a Bank Run?

Ingest, workload or packaged pricing decides more of your SIEM bill than features do. Here's how Splunk, Sentinel and Google SecOps compare on cost, regulator retention, residency and fit for banks of different sizes.

By the Wealth Deskplus editorial team · November 2026 · 8 min read

Splunk vs Microsoft Sentinel vs Google SecOps: three different bets on your log growth

A bank picking a SIEM is really answering one question: how will our log volume grow over the next five years, and who do we want to pay for that growth? Splunk vs Microsoft Sentinel is usually the first comparison a security team runs. Google Security Operations (Google SecOps, formerly Chronicle) is the third option procurement keeps hearing about, because its commercial model differs from the other two.

All three can detect a compromised privileged account or a suspicious SWIFT operator session. They differ on four things that matter more to a bank than to most buyers:

  • Pricing model: per-GB ingest, compute-based workload, or packaged pricing closer to flat.
  • Long-term retention: what it costs to keep logs searchable or restorable for the periods your regulators and auditors expect.
  • Data residency: whether logs can stay in-country in jurisdictions such as India, Saudi Arabia, the UAE, Indonesia or the EU.
  • Estate fit: how much of your telemetry is already Microsoft, Google Cloud, AWS, or on-premises mainframe and core banking.

There's no universal winner. Each platform is clearly cheaper and easier for one kind of bank, and expensive or awkward for the others.

Advertisement

SIEM pricing: ingest, workload and packaged models compared

Most bank evaluations go wrong here. Teams compare list prices per GB when the real cost drivers are which data sources are billable at all, and what happens to the bill when volume doubles after a cloud migration.

Splunk (now part of Cisco) sells Splunk Cloud Platform mainly in two ways. Ingest-based pricing is measured in GB per day. Workload pricing is measured in compute units, so you pay for search and processing capacity rather than raw volume. Workload pricing suits banks that ingest a lot but search a narrow slice, such as firewall and proxy logs kept mostly for forensics. Splunk Enterprise Security, the SIEM application, is licensed on top, and SOAR is separate.

Microsoft Sentinel is priced per GB ingested, with pay-as-you-go rates or discounted commitment tiers. What matters most is the free data: many Microsoft sources, such as Azure activity logs and certain Defender alerts, are free or discounted to ingest, and Microsoft 365 E5 customers may receive a data grant. Confirm the current list with Microsoft, because it changes. Lower-cost log tiers and the Sentinel data lake are built for high-volume, low-value telemetry. For a Microsoft-heavy bank, a large share of the bill can effectively disappear.

Google SecOps is sold in Standard, Enterprise and Enterprise Plus packages. Deals have often been priced by employee count or ingestion bands rather than strict per-GB metering, so cost is more predictable as volume grows. Higher tiers bundle Mandiant and Google threat intelligence.

FactorSplunkMicrosoft SentinelGoogle SecOps
Primary modelIngest (GB/day) or workloadPer-GB ingest, commitment tiersPackaged tiers; employee- or volume-banded
Biggest cost driverVolume or search compute; add-on appsNon-Microsoft data volumePackage tier and contract term
Free or discounted dataLimited; negotiatedMany Microsoft sources; possible E5 grantVolume less sensitive within band
Indicative annual cost, ~100 GB/dayUSD 200k–600k incl. Enterprise SecurityUSD 80k–250k before free-data offsetsUSD 120k–400k depending on tier

These are typical 2026 market ranges. They vary widely by contract, region, term and discounting, so confirm with each vendor or reseller. Local pricing in India, the Gulf or Latin America can differ sharply from USD list.

Hypothetical worked example: say a 3,000-employee regional bank produces 80 GB/day. About 35 GB is Entra ID, Defender and Microsoft 365 data, 30 GB is firewall and proxy, and 15 GB comes from core banking, card switch and database audit logs. On Sentinel, much of the 35 GB may cost little or nothing and the firewall data can move to a cheaper tier, so billable analytics volume could fall well below 80 GB/day. On Splunk ingest pricing, all 80 GB counts. On Google SecOps, the bank would likely sit in one package band whatever the mix. The cheapest platform depends on that mix, not on headline rates.

Log retention for regulators and data-residency options

Banks rarely choose their own retention periods. PCI DSS expects at least 12 months of audit log history, with the most recent three months immediately available. India's CERT-In directions require certain logs to be kept for 180 days within Indian jurisdiction. The EU's DORA, MAS technology risk guidance in Singapore, SAMA's cyber security framework in Saudi Arabia and Brazil's central bank cyber rules all expect protected, retrievable logs for incident investigation. Many banks' internal policies push retention to five to seven years. Compliance, legal and external counsel should set the figure; the SIEM then has to deliver it at a sensible cost.

Google SecOps has the strongest default. Its packages have included a year of hot, searchable retention, with longer terms available. For a PCI-scoped bank, that alone can remove the need for a separate archive tier. Verify the current term for your package.

Microsoft Sentinel splits retention into an interactive analytics tier and long-term storage, including the data lake, at much lower per-GB cost. Querying archived data is slower and billed separately, so budget for investigation-time search charges.

Splunk Cloud includes a set searchable retention period. You extend it with Dynamic Data Active Archive (restorable by Splunk) or Dynamic Data Self Storage (exported to your own cloud bucket). Self Storage is cheap, but the data isn't searchable in Splunk until you re-ingest it. Auditors sometimes notice that gap.

Residency factorSplunkMicrosoft SentinelGoogle SecOps
Self-hosted optionYes: Splunk Enterprise on-premises or private cloudNo; Azure-hosted (sovereign cloud options in some markets)No; Google-hosted SaaS
Regional footprintSplunk Cloud on AWS and Google Cloud regions; varies by countryBroad: follows the Azure region map, including Gulf, India, Africa, Latin AmericaGrowing list across Europe, UK, Middle East, Asia-Pacific, Americas
Best for strict in-country rulesStrong, via self-hostingStrong where a local Azure region existsCheck country list carefully

Get written confirmation of every region a vendor supports in your market, including where threat-intelligence enrichment and support access happen, not only where data is stored. Regulators in several regions, including the Gulf and Southeast Asia, increasingly ask about both.

Financial-services detection content and fit with your existing estate

None of the three ships a complete banking detection pack. Fraud-adjacent use cases still have to be built: teller and back-office privilege abuse, anomalous payment-file changes, card-management system access, SWIFT Customer Security Programme controls. How hard that is depends on the platform.

Splunk has the deepest third-party ecosystem and the largest pool of experienced engineers. Its Enterprise Security Content Update supplies maintained detections mapped to MITRE ATT&CK, and many banks already use Splunk for IT operations and fraud data, so SPL skills are often in house. Parsing for mainframe, core banking and legacy payment systems is usually easiest on Splunk, simply because someone has already built it.

Microsoft Sentinel content arrives through the Content Hub, with solutions for Microsoft workloads, major firewalls, SAP and common cloud platforms. It's increasingly run from the Defender portal alongside Defender XDR. If your endpoints, identity and email are on Microsoft, correlation is native and discounted data makes wide coverage affordable. KQL is easy to learn, but connectors for legacy banking systems often need custom work.

Google SecOps normalizes data into its Unified Data Model and offers curated detections maintained by Google. With higher tiers, Mandiant intelligence covers threat actors that target financial institutions, and search over large volumes is fast. The trade-offs: YARA-L skills are scarcer in most local talent markets, and parsers for niche banking systems may need partner effort.

Red flags to test during a proof of concept:

  • No working parser for your core banking platform or card switch within the trial window.
  • Detection content that can't be version-controlled and promoted through change management, which supervisors increasingly expect.
  • SOAR or case management priced as a surprise add-on after the shortlist stage.
  • Search costs on archived data that make a regulator's look-back request expensive.

Verdicts for small, mid-size and multinational banks

Small banks, credit unions and digital lenders (under roughly 1,500 staff, under 50 GB/day): choose Microsoft Sentinel if you already run Microsoft 365 E5 or Defender. Discounted Microsoft data and per-GB pricing keep the bill small, and a managed security service provider can run it without a dedicated SIEM engineer on your payroll. If you're a Google Workspace or Google Cloud shop, Google SecOps Standard is a credible alternative. Splunk is rarely the cost-efficient choice at this size unless you're inheriting it.

Mid-size banks (roughly 1,500–15,000 staff, 50–500 GB/day, mixed estate): Google SecOps is most competitive here. Predictable packaged pricing and long hot retention suit banks whose volume is growing fast through cloud adoption and that face PCI and local look-back rules. Choose Sentinel instead if most telemetry is Microsoft-native and you're comfortable managing per-GB tiers. Stay on Splunk if you have a mature SPL team and years of tuned content; migration risk is real and usually underestimated.

Multinational and systemically important banks: Splunk remains the safest fit where some jurisdictions require on-premises or in-country hosting that public SaaS can't meet, and where mainframe and legacy coverage matters most. Use workload pricing to hold costs down. A common pattern is a hybrid: Splunk or Google SecOps as the global analytics layer, with Sentinel covering Microsoft cloud estates where its data discounts are largest. Budget for the extra integration effort.

Next step: pull 30 days of real log volume by source and by country, then ask all three vendors to price that exact mix over five years with 12 months of hot retention and your required archive period. Whichever quote stays flattest when you double the volume is usually your answer.

Quick answers

Is Microsoft Sentinel really cheaper than Splunk for a bank?

Often, but mainly if a large share of your logs comes from Microsoft sources that are free or discounted to ingest, such as Defender alerts and Microsoft 365 audit data. For banks with heavy non-Microsoft volumes like firewall, network and core banking logs, the gap narrows considerably. Model your actual source mix over five years rather than comparing per-GB list prices.

Can Google SecOps meet in-country data residency rules for banks?

In many markets yes, since it supports a growing list of regions across Europe, the UK, the Middle East and Asia-Pacific. But it has no self-hosted option, so jurisdictions without a supported local region may rule it out. Confirm in writing where data is stored, processed and accessed for support, and have your compliance team and counsel review it against local rules.

How long should a bank retain SIEM logs?

It depends on your regulators and internal policy. PCI DSS expects at least 12 months of history, and India's CERT-In directions require 180 days kept in-country for covered logs. Many banks keep five to seven years for investigations and litigation. Your compliance and legal advisers should set the period, and the SIEM should be priced against it from the start.

Written by the Wealth Deskplus editorial team. Prices, features and rules change often - confirm details with the vendor or a licensed professional before you decide.