# Marcus Trellenberg, Director of Engineering at Fulcrum Analytics — read of DevOps Cost AI, May 19 2026

> 11 years in backend and infra, last 4 running a 14-person eng team at a 190-person B2B SaaS. We're on AWS, spending roughly $40K/month, and I think about 15% of it is garbage I haven't had time to clean up.

## How I got here

Someone in the Rands Leadership Slack dropped a link in #cloud-costs with no comment, just the URL. That channel is usually pretty high signal so I clicked. I was genuinely looking -- we've been poking at Spot.io and the native AWS Cost Explorer recommendations have gotten better but still require a human to actually do anything about them.

## What I clicked first

The hero subhead: "Your cloud bill typically hides 30-40% waste across unused instances, orphaned resources, and misconfigured services." Fine, that's real, I believe it. Then the big H1 hits: "Cut Cloud Spending by 40%. Instantly."

I've been around long enough that "40% instantly" is a yellow flag, not a green one. Either that number is the theoretical ceiling (which means most customers see 8%) or there's a catch in what "instantly" means. But I kept reading because the features section actually described the problem I have -- orphaned volumes, oversized RDS instances. Specific enough that I didn't bounce.

## Where I paused

The honest disclosure box. I had to read it twice.

"Honest disclosure: we don't have live customers on this idea yet. We shipped the strategy package; you ship the customer conversations."

Wait. So the "Get Started -- Free for 30 Days" button at the top... that's hypothetical? The CloudFormation integration, the dry-run execution engine, the rollback logic -- none of that exists yet? I scrolled back up and re-read the feature section with new eyes. It reads exactly like a real product because it IS written exactly like a real product. But it's a pitch deck someone turned into a landing page.

That's a significant bait-and-switch in the reading experience. I don't think it's dishonest exactly, but by the time I hit the disclosure I felt a little played.

## What I distrusted

A few things.

"Executes cost reductions automatically." Full stop. No DevOps lead at a company with production traffic is letting a $99/month SaaS touch infra automatically. The FAQ walks it back -- "every recommendation requires your explicit approval" -- but the hero copy is still leading with "automatically." That's either sloppy copywriting or it's a deliberate hook that relies on the FAQ being below the fold.

The pricing table. $99/month for 3 AWS accounts -- okay. $499/month for "growing teams with complex infra." What does complex mean? We have 7 accounts and a pretty boring setup. Do I need the $499 tier? No guidance.

Also: "Benchmark against industry." Against what industry benchmark? Who? What methodology? This is the kind of claim that sounds good in a screenshot for a finance deck and means nothing.

## What would convince me

If this were a real product, I'd want one thing: a specific case study from a company my size (100-300 people, $30-60K/month AWS) showing the before/after on a line-item level. Not "saved 37% overall" -- I want to see "we had 14 gp2 volumes that hadn't been read in 90 days, totaling $620/month, tool flagged them, we approved deletion, saved $7,440 annually." That kind of granularity tells me the tool actually works and the team understands the actual problem.

Since this is a build kit and not a product, the thing that would convince me to hand over $5 is a single real example of someone who bought a dossier from this studio and shipped a product that got paying customers. Just one. The honesty about the odds is interesting but "1 in 8 meaningful success odds" is a number from a spreadsheet, not from lived history.

## What I'd ask in an email reply

If I were emailing as a potential builder (which is apparently the actual CTA here):

1. You show the ICP as "engineering managers at 50-500 person companies" -- have you talked to any of them? Not surveyed, talked. What's the actual objection you keep hearing when you try to sell this?

2. The "speed to MVP" score is 4/10 and you're saying 8-12 weeks to launch. What's the hardest part -- the AWS integration surface, the execution engine, or getting the AI recommendations to be accurate enough that engineers trust them?

3. "$-11,000 Year-1 take-home" -- I appreciate the honesty but that's a rough number. Is that after a solo founder doing all the work or is that accounting for hiring? Because those are wildly different scenarios.

## Verdict: on-the-fence

The honest scoring and disclosure are genuinely unusual and I respect it. But the page is trying to be two things at once -- a product landing page for a DevOps tool that doesn't exist yet, AND an idea marketplace pitch -- and it's not fully either one. I'd probably read the $5 dossier out of curiosity, but I'd feel a little sheepish about it.

---
*Memo by skeptic persona, generated 2026-05-19. Studio breaks own self-grading loop.*