# Marcus Chen, VP Engineering at Partnerbound — read of ProxyBox API Analytics, June 5 2026

> 11 years building API platforms, currently running infra for 40-some reseller accounts on a 28-person B2B SaaS. Commute is 44 minutes on the Caltrain. This is where I read things.

## How I got here

Searched "api usage attribution by reseller account" last Tuesday because our Datadog bill hit $4,100 and I still can't tell the PM which of our three biggest partners is the one causing 60% of the retries. Google served this as the third result, right under a StackOverflow thread from 2019. Clicked it skeptically.

## What I clicked first

The hero description landed: "Understand and optimize every API call your resellers make." That's the exact sentence I would have written on a sticky note before searching. Stayed on the page. The feature list reads like someone interviewed a person in my exact situation, specifically the "Cost attribution engine" bullet: "Know which of your tools is driving expense." Yes. That.

Then I hit the scoring table and almost closed the tab.

## Where I paused

The self-scoring section is genuinely strange. "52/100 Adoptability. $-21,775 Year-1 take-home (Fermi). 1 in 8 Meaningful-success odds." I've never seen a product page tell me it probably won't make money. Sat with that for a minute. Either this is a very clever trust play or someone forgot to hide their internal spreadsheet. Not sure which. I read the adjacent sentence twice: "We shipped the strategy package; you ship the customer conversations." So they're selling me a blueprint, not a product. That reframes the whole page.

## What I distrusted

"Fix the top 3 inefficiencies and cut costs 20%+ immediately." That "immediately" is doing a lot of work and I don't believe it. What does immediately mean when I have 40 active reseller integrations and a 6-week sprint cycle? Also "Performance heatmap" is a phrase I've seen on 20 dashboards that turned out to be a bar chart with color. And "financial upside: 1/10" is their own score for this idea. They're scoring financial upside at 1 out of 10 and still trying to sell me on building it. That's a weird tension they don't resolve.

## What would convince me

One real reseller-facing API business telling me specifically: what their before/after looked like on API cost per account, how long it took to instrument, and whether it caught something they would have missed in Datadog. Not a case study PDF. A 4-paragraph Notion doc or even a tweet thread. I don't need polished. I need real. Also: show me what the attribution model actually does when a reseller account uses three endpoints in one request. That's the hard part. The dashboard is easy. The attribution logic is where every tool I've looked at falls apart.

## What I'd ask in an email reply

1. The attribution engine, specifically: if a reseller's integration chains calls across three endpoints in one webhook, how do you split the cost? Does it require instrumentation on my side or does it infer from logs?
2. You say "no live customers on this idea yet" and you're selling me a $99 build starter. Who built the starter, and has anyone actually run it against a live API with real reseller traffic?
3. The 1/8 success odds, is that based on TAM assumptions, on how hard it is to get someone to pay for this as a standalone product, or something else? Because if it's TAM I care. If it's "most people who buy idea packages don't execute" that's different.

## Verdict: on-the-fence

The transparency on financials and scoring is the most interesting thing on this page, and the pain point is real for me. But they're selling me a map when I need a tool, and the self-scored "financial upside: 1/10" makes it hard to justify even $99 without knowing the attribution logic actually works.

---
*Memo by skeptic persona, generated 2026-06-05. Studio breaks own self-grading loop.*