I built a $15/mo GCP cost anomaly detector after losing $1,800 to a runaway Dataflow job
Hey IH đź‘‹
GCP's native budget alerts only fire at hard dollar thresholds.
They miss the actual failure mode: a job that was spending $20/day yesterday and is suddenly spending $400/day today, but you set your budget at $1,000/mo so nothing fires until day 25.
BillSnap pulls your existing GCP billing export every morning, computes a 14-day rolling z-score per SKU, and Slacks/emails you when something is statistically anomalous.
- $15/mo per connected project
- 14-day free trial, no card
- Read-only, dataset-scoped IAM (you control the grant, you can revoke any time with one gcloud command)
Built in 14 days. Solo. AMA.
The origin story is good — losing real money to a runaway job is exactly what makes a product credible to other devs.
One gap I noticed on billsnap.dev: there's no email capture anywhere on the landing page. Every CTA goes straight to app.billsnap.dev/signup, which makes sense for someone who's ready to create an account — but there's no option for the dev who reads this thread, gets interested, and needs to pitch it to their team first, or bring it up in the next retrospective. They just leave. A lightweight "keep me posted on new alert types" or simple opt-in field would hold that second group in your pipeline.
Full engine breakdown if useful: https://outboundautonomy.com/audit/05226/billsnap-dev?ref=ih-billsnap
This is a strong wedge because the pain is very specific: not “cloud cost management,” but catching runaway spend before the invoice damage is already done.
I’d make that the front-door positioning even sharper. “GCP cost anomaly detector” is accurate, but “runaway cloud bill protection for GCP” is closer to the fear buyers actually feel after one bad job or misconfigured pipeline.
The $15/mo pricing also makes sense because it is not competing with FinOps platforms. It is more like cheap insurance for solo founders and small teams running GCP.
For early growth, I’d target people sharing painful cloud-bill stories, not generic GCP users. Those are the buyers who immediately understand why this exists.
Appreciate this. The “cheap insurance” framing is exactly how I think about it.
Most teams don’t wake up thinking they need a FinOps platform, they just want to know when a query suddenly starts burning money before it turns into a painful invoice. That’s the gap I built BillSnap for.
I also think you’re right about distribution. People who’ve already been burned by a surprise cloud bill firsthand immediately get the value proposition.
Definitely going to experiment with messaging around “runaway cloud bill protection” and see how it resonates. Thanks for the thoughtful feedback.
That’s the right test.
The useful part now is probably not more features, it’s tightening the exact message and where you place it.
For BillSnap, I’d pressure-test a few versions around the same fear:
runaway GCP spend, surprise invoice prevention, and cheap insurance before one bad job burns through budget.
The distribution side also needs to stay close to painful cloud-bill stories, not broad GCP communities.
Drop your email and I’ll send over a tighter version. It’ll be easier to map the positioning and first outreach angles properly without turning the thread into a full teardown.