Back to Blog
    companyfounder story

    Why We Left Big Cyber to Build Pooka

    ·6 min read

    Why We Left Big Cyber to Build Pooka

    We didn't leave because the pay was bad. We left because the work was broken.

    We've run the same engagement twice for the same customer. Full report the first time. Prioritized findings. Remediation recommendations. Came back a year later.

    Nothing changed.

    Same vulnerabilities. Same misconfigurations. Same attack paths. Everything that worked the first time worked again. The customer had the report. They just didn't have anything in place to act on it.

    Same attackers target every business. Some have defenses. Most don't.

    That's the problem we built Pooka to fix.

    What We Saw From The Inside

    The MSSP Industry Is Mostly Theater

    Here's what the typical managed security engagement looks like for a small business:

    1. Sales team sells "24/7 SOC monitoring"
    2. Implementation team deploys a SIEM agent and connects it to the platform
    3. Default alert rules fire on everything
    4. Alert volume is overwhelming, so thresholds get raised until only "critical" alerts make it through
    5. "Critical" alerts get forwarded to an email distribution list
    6. An offshore analyst reads the email, opens a ticket, and sends the customer a notification
    7. Customer has no idea what to do with the notification
    8. Ticket sits open for days, eventually gets autoclosed

    This is the industry standard. The customer pays $3,000 to $10,000 per month. They get log collection, generic alerts, and notification emails. They call it "managed detection and response." It's managed alert forwarding.

    We know because we've worked inside these organizations. We've seen the offshore SOC centers where a single analyst monitors 200+ customers simultaneously. We've seen the alert queues where medium severity alerts pile up by the thousands, never investigated. We've seen the "detection rules" that haven't been updated since the platform was deployed.

    The business model works because nobody tests it. The customer assumes someone is watching. The MSSP assumes the customer is handling the alerts they forward. Nobody is actually defending anything.

    Until there's a breach. Then everyone points fingers.

    Enterprise Tools Don't Scale Down

    The other thing we saw: enterprise security tools are not designed for small businesses. They're designed for organizations with dedicated security teams, unlimited budgets, and armies of analysts.

    A 50 person company doesn't need a $500,000 SIEM that requires three full time engineers to operate. They don't need an EDR platform that generates 10,000 alerts per day for a team of zero people to investigate. They don't need a SOAR platform that automates playbooks nobody has written.

    What they need is:

    • Detection rules that actually fire on real attacks (not vendor defaults)
    • Someone who looks at those alerts at 3am
    • A response capability that does something about it
    • Security advice calibrated to their actual resources and risks

    This doesn't exist at a price point SMBs can afford. Or rather, it didn't.

    Attackers Don't Discriminate By Revenue

    The enterprise has 20 security analysts, a $5M security budget, and detection coverage across 200 MITRE ATT&CK techniques. The law firm has an IT consultant who comes in on Thursdays and an antivirus that's three versions behind.

    Both get the same threat. Only one can handle it.

    What We Built Different

    Detection Engineering, Not Default Rules

    Every detection rule in our platform has been tested against real attack telemetry. We maintain an in-house library of recorded attacks. When we deploy a detection rule, we inject the corresponding attack log and verify the rule fires.

    This matters because default SIEM rules are garbage. They're written to demonstrate capability, not to detect real attacks. They generate false positives that bury real threats. They miss attack variations that any competent adversary would use.

    We currently maintain over 4,800 custom detection rules across our Wazuh and Graylog stack. Every single one has been validated. If we can't prove it detects a real attack technique, it doesn't ship.

    Deception As A Core Capability

    Deception technology is the great equalizer for small businesses.

    An enterprise can afford to deploy endpoint detection on every host, network monitoring on every segment, and 24/7 analysts to watch it all. A small business can't.

    But a small business can deploy honey credentials in Active Directory, canary files on their file server, and a honeypot on their network. These cost almost nothing to operate, generate almost zero false positives, and catch attackers that sophisticated EDR systems miss.

    When an attacker enumerates Active Directory (which they always do), they find the fake admin account. When they try to use it, we know with certainty that an intruder is inside the network. No probabilistic detection. No alert fatigue. Just a clear signal that something is wrong.

    We built deception into every tier of our service because it provides disproportionate value to organizations that can't afford to monitor everything.

    Veteran Mentality

    Military service teaches you something that civilian security culture often lacks: mission over comfort.

    In the military, the mission doesn't pause at 5pm. The threat doesn't take weekends off. You don't get to "deprioritize" a critical situation because it's inconvenient.

    We bring that mentality to security operations:

    • Our monitoring doesn't have business hours
    • We don't forward alerts. We investigate them.
    • We don't write reports about problems. We fix them.
    • We don't rely on compliance frameworks to tell us we're secure. We test.

    This isn't marketing language. It's operational philosophy. When an alert fires at 3am, someone is investigating it within minutes. Not because a contract says so, but because that's how we operate.

    Built In America, Operated In America

    Your security data should not leave the country. Full stop.

    When your MSSP's SOC analysts are in a different country, your security telemetry, including endpoint logs, authentication data, network traffic, and potentially credentials, is being transmitted internationally. This creates legal exposure, compliance complications, and security risk. For organizations subject to CMMC, HIPAA, or ITAR, that's not a preference. It's a requirement.

    PookaSec is US based. Our servers are in the US. Your data doesn't leave the country. This isn't xenophobia. It's operational security and regulatory compliance.

    The Name

    Pooka. A shapeshifter from Irish mythology. Sometimes helpful, sometimes not. Unpredictable. Not to be trusted at face value.

    That's what we wanted our deception technology to feel like to attackers. Every credential could be real or fake. Every network share could be legitimate or a trap. Every service could be authentic or a honeypot.

    The attacker can't tell the difference. And that uncertainty is the point.

    The security industry sells certainty: "Deploy our product and you'll be safe." That's a lie. Nobody is completely safe.

    We sell uncertainty. We put it on the attacker's side.

    What's Next

    We're building the security operations capability that SMBs deserve but have never had access to. Detection engineering that's actually tested. Monitoring that's actually monitored. Deception that actually catches threats. Response that actually responds.

    We're a startup. We're small. We don't have a billion dollar valuation or a Super Bowl commercial. What we have is offensive security experience, a validated detection stack, and the conviction that small businesses deserve better than what the industry currently gives them.

    If that resonates, we should talk. Not a sales pitch. A conversation about what your security actually looks like and what it could look like.

    Because the status quo isn't working. And we're tired of watching good companies get breached by preventable attacks.