Skip to content

Cutting a Splunk Bill by Half Without Losing a Single Detection: A Noise-First SIEM Migration

Per-GB SIEM pricing punishes exactly the telemetry sprawl that modern cloud-native stacks produce - most of which nobody ever searches. The fix wasn't a bigger contract; it was finding out what was actually noise first.

58% ingest reduction before any platform change
SIEM spend cut by more than half
Zero detection coverage lost
01 — The problem

What we were solving

Mid-market financial services firm, legacy per-GB-priced SIEM contract renewing at a steep increase, security team assuming they had to accept the cost or accept a coverage gap - not realizing most of the ingest was never queried.

  • Daily log ingest had grown year over year as cloud services multiplied, and the SIEM bill scaled directly with it - regardless of whether anyone ever searched that data.
  • The security team assumed cutting ingest meant cutting detection coverage, so nobody had seriously audited what was actually noise versus signal.
  • Migrating SIEM platforms outright felt too risky to attempt without knowing which data actually mattered to existing detections.
02 — The approach

What we built

  1. Audited a full week of ingest against actual query and detection-rule usage to classify every log source as signal, occasional-signal, or pure noise - debug lines, duplicate health checks, and redundant heartbeats made up the bulk of it.
  2. Applied upstream filtering and sampling on the noise sources before touching the platform contract, cutting ingest volume while every detection rule kept the exact fields it depended on.
  3. Only after ingest was right-sized did the team migrate the reduced, higher-signal stream onto a correlated AI-assisted platform - a smaller, cheaper migration than moving the original noisy volume would have been.
Architecture notes for your engineers
  • Filtering happened upstream of the SIEM at the collector/forwarder layer, so source systems and downstream detection logic needed zero changes.
  • Every detection rule was replayed against the filtered stream before cutover to confirm no dependent field had been silently dropped.
05 — Outcomes

Key results

  • 58% reduction in daily log ingest before any platform change
  • Annual SIEM spend cut by more than half at renewal
  • Every existing detection rule preserved with zero coverage gaps
  • Team could finally explain to the board exactly what was being paid for
06 — Stack

What it was built on

Representative tools and patterns — exact vendors vary per client environment.

Ingest analysis

Query/rule-usage audit toolingLog-source signal classificationUpstream filtering & sampling

SIEM platform

Correlated AI-assisted SIEM/XDRDetection-rule replay harnessOpenTelemetry-based collection

Validation

Field-dependency mapping per ruleParallel-run comparisonAlert-parity sign-off
07 — Learnings

What we'd tell the next team

  • Most SIEM cost problems are ingest problems, not platform problems - fix the noise before you sign a new contract or migrate anything.
  • Debug lines, duplicate health checks, and redundant heartbeats are the recurring villains; auditing usage against ingest usually finds them fast.
  • A smaller, cleaner migration is dramatically less risky than moving your full noisy volume onto a new platform and hoping the noise sorts itself out later.
FAQ

Questions this engagement anticipated

Why cut ingest before migrating platforms?

Per-GB SIEM pricing means noisy ingest is the actual cost driver, not the platform choice - cutting noise first makes any subsequent migration smaller, cheaper, and lower-risk, and often changes whether a migration is even necessary.

How do you cut ingest without losing detection coverage?

By auditing which log fields every existing detection rule actually depends on before filtering anything - noise sources like debug lines, duplicate health checks, and redundant heartbeats can be dropped or sampled without touching the fields detections rely on.

This is one of several case studies on siem & soc modernization.

See the rest of the cluster →

Compare your situation to this case.

Bring your constraints - environment, timeline, and budget. We scope before we quote.

Get started

Leave your email - we'll reach out.

Share your work email and we'll follow up with a tailored note on security, AI, or data programmes - usually within one business day.

No spam. We only use your email to respond to this request.

Explore services →