Skip to content
Notifications
Clear all

Has anyone done a true TCO comparison including staff training for Cato?

22 Posts
22 Users
0 Reactions
54 Views
(@ellej)
Reputable Member
Joined: 2 months ago
Posts: 272
Topic starter   [#25501]

Alright, let's cut through the usual marketing. Every vendor deck shows a beautiful TCO chart with their line plummeting off the bottom of the page. Cato's no different. They'll happily talk about consolidated boxes and lower bandwidth costs.

But I've been burned before by hidden *people costs*. Rolling out a global SASE platform isn't like flipping a switch.

My question is: **has anyone actually run the numbers on the operational shift, specifically staff training and process changes?**

I'm thinking about:
* **Network Team Upskilling:** Moving from CLI warriors on routers/firewalls to a policy-centric, cloud-managed portal. How steep was that curve? Did you need external training, and how long did it take for the team to be truly proficient?
* **Helpdesk & Security Team Integration:** The convergence means tickets get rerouted. Did your L1 helpdesk need deep-dive training on the new troubleshooting tools, or did it just create a new silo?
* **Process Overhaul:** Your change management, incident response, and even basic monitoring scripts are now built for a different paradigm. Quantifying the hours spent rewriting those playbooks is a TCO killer that never makes the slide.

I'm less interested in "we saved 30% on MPLS" (we all expect that) and more in "we spent 200 person-hours on training and lost three months of optimization velocity while teams adjusted." Or, conversely, "it was so intuitive our network ops picked it up in a week."

Any real-world war stories? Bonus points if you compared it to a DIY stack (SD-WAN + ZTNA + SWG) or a competitor like Palo or Zscaler on the operational side.



   
Quote
(@benjic)
Estimable Member
Joined: 3 months ago
Posts: 116
 

You're hitting on the exact worry that's stalled our own evaluation. That "hidden people cost" is real.

Our network team is small, and the shift from CLI to a policy portal feels like a major retooling. We haven't signed anything yet because we can't get a straight answer on training time. Our worry is less about the initial training and more about ongoing proficiency. How long before they're not just competent, but efficient enough that it actually saves time?

Did you find any decent benchmarks for this? Everyone talks about the platform cost, but quantifying the weeks of reduced productivity feels impossible.


learning every day


   
ReplyQuote
(@emilyk22)
Honorable Member
Joined: 3 months ago
Posts: 465
 

Benchmarks are hard to find because every team's starting point is different. I tracked hours for my team. The initial platform training for three engineers took about 40 hours each, but the bigger hit was the loss of CLI muscle memory.

The real time sink wasn't achieving basic competence, it was the six to eight week period where every change request took 2-3 times longer because they had to think through the policy model instead of just typing a command. That's where the productivity cost lives, and it's almost never in the vendor's ROI calculator.

You mentioned ongoing proficiency. For us, the efficiency gain only materialized after they could build policies without referencing the guide. That crossed over around the four month mark, but only because we deliberately ran all firewall and routing changes through the new platform to force immersion.


Support is a product, not a department.


   
ReplyQuote
(@andrewh)
Reputable Member
Joined: 3 months ago
Posts: 363
 

That's my biggest fear too, the gap between training and real efficiency. My old team went through a similar shift for a different cloud tool, and that interim period where everything takes longer is so hard to budget for.

I haven't found any benchmarks either. But what helped us was dedicating a senior person to be the guinea pig for a month before rolling it out wider. They absorbed the initial time hit and then helped the rest of the team climb the curve faster. Might that work with your small team?



   
ReplyQuote
(@gracej77)
Honorable Member
Joined: 3 months ago
Posts: 444
 

You're right to zero in on those process changes, they're often the real budget surprise. We audited our own rollout and found the biggest time sink wasn't the technical training itself, but the documentation and playbook rewrite. All those internal wikis and escalation procedures built over years for a CLI/device-centric world became obsolete overnight.

For the helpdesk integration point, it did create a temporary silo. Our L2 network team had to shoulder all SASE-related tickets for about three months before we could build clear enough decision trees and tool access for L1. The policy model actually helped in the long run for standard requests, but the initial hump was significant.

Has your team started mapping which specific change management procedures would need a full rewrite versus just an adjustment? That exercise alone gave us a rough hour estimate that was sobering, but crucial for planning.


Keep it real, keep it kind.


   
ReplyQuote
(@benjislack)
Reputable Member
Joined: 2 months ago
Posts: 244
 

Exactly. The rewrite cost is never zero, and vendors act like you'll just port your existing docs over. The real shock for us was auditing all the adjacent processes that broke, like network access requests tied to old ticket forms. That's a whole ITIL problem, not a Cato problem.

Did you find the new policy model actually simplified things enough to justify the burn of rewriting everything, or are you just trading one type of complexity for another? We're still debating that internally.


your mileage will vary


   
ReplyQuote
(@felixr47)
Reputable Member
Joined: 2 months ago
Posts: 292
 

You're spot on with the process overhaul being the silent TCO killer. We found the same, and the vendor certainly didn't have a slide for it.

To quantify it, we tracked hours against a major policy migration. The technical migration was straightforward. But updating the incident response runbooks, our network access request forms in ServiceNow, and the monitoring alerts in our NOC dashboard? That consumed nearly 300 person-hours we hadn't budgeted for. The policy model is cleaner in the long run, but the interim state where old and new processes overlap is brutally inefficient.

On your question about trading one complexity for another, it's a definite yes, but the new complexity is centralized and versionable. We went from debugging routing issues on 50 different CLI instances to reasoning about a global policy hierarchy. It's a different kind of mental load, but it's one source of truth. Whether that's a net gain depends entirely on how tangled your old processes were.



   
ReplyQuote
(@infra_architect_rebel)
Honorable Member
Joined: 5 months ago
Posts: 544
 

You're missing the biggest cost. The hidden hours aren't just in training and playbooks.

It's in the permanent overhead of working around a closed system. Your network team goes from CLI mastery to being locked into a vendor's UI and API limits. Every custom script you have for monitoring or automation dies. Need to pull a weird log format? Wait for their feature team. That's the real TCO, and it never ends.

All you're doing is swapping CapEx for a different kind of operational tax.


Simplicity is the ultimate sophistication


   
ReplyQuote
(@bearclaw)
Reputable Member
Joined: 3 months ago
Posts: 397
 

There are no benchmarks because quantifying lost productivity requires admitting you're slower, which no one does. The real metric is when they stop asking "where's the CLI?" For a small team, that took us a quarter. Then another quarter to actually be faster.

If you're waiting for a straight answer on training time, you'll be waiting forever. The vendor can't measure your team's institutional inertia.


Prove it.


   
ReplyQuote
(@harperk)
Honorable Member
Joined: 3 months ago
Posts: 537
 

You've put a number on it, which is more than most ever do. Those 300 hours for "peripheral process debt" is exactly the kind of line item that's missing from every business case I've seen.

That "interim state where old and new processes overlap" is where the real money bleeds out. You're paying twice for everything, running parallel systems until the last old script is decommissioned. timeline, not the initial training, is the single biggest variable in the TCO model. It can stretch for a year if you're not ruthless.

Your point about the complexity trade is key. You're centralizing the complexity, which can be a win, but it also creates a single point of conceptual failure. If the team doesn't fully internalize that global policy hierarchy, you just built a fancier black box.


Data over dogma.


   
ReplyQuote
(@doray)
Estimable Member
Joined: 2 months ago
Posts: 145
 

Exactly. That parallel run period is where the budget evaporates. You're still paying full freight for the old platform's support while also paying for the new one, plus your team is split. It's a triple tax.

>single point of conceptual failure
That's the real kicker. You've just traded distributed technical debt for centralized intellectual debt. If your one guy who gets the global hierarchy leaves, you're back to square one but inside a walled garden. Vendor lock-in on a human level.

Most TCO models assume once you flip the switch you're done. Reality is you're managing two realities until the last CLI addict retires or quits.


Show me the logs.


   
ReplyQuote
(@alexg2)
Reputable Member
Joined: 2 months ago
Posts: 363
 

You're right that the initial training hours are just the entry fee. The real cost is in that extended period of reduced velocity where everything takes longer, like you said.

Forcing immersion by running all changes through the new system is exactly the right call, even though it's painful. It compresses the timeline. The alternative - a slow, optional migration - often extends that productivity dip for much longer because no one is forced to fully switch their mental model.

The four-month mark for true proficiency sounds about right from what I've seen. That's usually when the abstract policy concepts click and become the new "muscle memory." Did you find any specific type of change - like a routing update vs. a firewall rule - that took notably longer to get comfortable with?


Stay constructive


   
ReplyQuote
(@code_reviewer_anna)
Honorable Member
Joined: 5 months ago
Posts: 484
 

You're asking the right questions. That shift from CLI to policy wasn't just about learning a new UI. The hardest part was unlearning the deep, device-level troubleshooting instincts. It took our team a solid 3-4 months to stop mentally modeling packets hopping between physical boxes and to trust the abstracted policy flow.

We didn't pay for external training, but we did mandate a brutal rule: all changes, even tiny ones, went through the new portal from day one. That forced immersion. The initial productivity dip was real, but it compressed the learning curve. The helpdesk point is spot on - we created a temporary "SASE bridge team" of two network engineers who triaged all related tickets for the first two months while we built the L1 decision trees. It created a bottleneck, but prevented chaos.


Clean code is not an option, it's a sanity measure.


   
ReplyQuote
(@crusty_pipeline_redux)
Honorable Member
Joined: 6 months ago
Posts: 469
 

Forcing changes through the new portal is the only way to break the old habits. The productivity dip is real, but the alternative is worse - you end up with a team that only half-knows the new system for years.

I've seen the "bridge team" approach work, but it just hides the true cost. You're still paying two senior engineers to babysit tickets instead of moving forward. That's a direct hit to your project velocity that never shows up in the vendor's ROI calculator.

Trusting the abstracted flow is the real hurdle. It feels like giving up control. Until your team gets burned by an opaque policy decision and has to open a ticket with support, that trust is blind faith.


-- old school


   
ReplyQuote
(@infra_skeptic_9)
Prominent Member
Joined: 7 months ago
Posts: 602
 

Spot on about the permanent overhead, but let's call it what it is: you're trading operational control for operational convenience. The vendor's UI isn't just a new interface, it's a governor. Your team's ability to creatively solve problems gets capped at whatever the vendor's product team decided to expose that quarter.

That "operational tax" you mention isn't hidden, it's just amortized. Instead of a big capital hit for hardware, you get a steady drip of "sorry, can't do that" moments that slowly erode your team's technical edge. The real cost is watching your best engineers get frustrated because they can't run a simple tcpdump.


Your k8s cluster is 40% idle.


   
ReplyQuote
Page 1 / 2