# Marcus Tolentino, Director of Development Operations at Cascadia Youth Alliance — read of campaign-performance-predictor, May 12 2026

> 9 years in nonprofit fundraising ops, currently managing a $2.4M annual fund with two development associates and a part-time data coordinator. Coach my daughter's U10 soccer team on Saturdays, which is the only reason I get outside.

## How I got here

Searched "how to tell if a fundraising email campaign is failing before the month is over" on a Sunday night because we just closed out April and our spring appeal underperformed by 31% and nobody saw it coming until it was too late. Google surfaced something that linked here from what looked like a Substack roundup of "AI tools for nonprofits." Clicked through mostly out of frustration.

## What I clicked first

The hero line. "Know which campaigns will tank before they tank." That is exactly the thing I was Googling for. That is the actual pain. I stayed on the page because of that line, not because of any design or image. Then I looked for the "Try it" section and found a before/after graphic but no actual interactive demo. That's where I started slowing down.

## Where I paused

The self-scoring section. Specifically this: "pain intensity: 4/10." The product's own evaluation panel says the pain this solves is a 4 out of 10. But I just spent my Sunday night searching for this exact solution because the pain is very real to me. Either their scoring methodology is off, or I am not the real ICP, or this product does not solve the pain as deeply as the headline promises. I read that number three times.

## What I distrusted

"Honest disclosure: we don't have live customers on this idea yet." I respect that they said it. I do. But there is nothing else to anchor any of the feature claims to. "Analyzes first-touch engagement, reply rates, and click decay" -- analyzed in what tool? Pulling from what integration? Mailchimp? Salesforce? Blackbaud? The page never says. "Every campaign gets a 0-100 performance projection by day 7" -- projection based on whose historical data? If I am a brand new customer with two years of Mailchimp exports and a messy Salesforce, how long before the model knows anything useful about MY campaigns?

Also: the Fermi math shows Year-1 take-home of -$26,600. That number is on the page, in the main scoring block. I understand this is a studio pitching an idea to builders, not directly to me. But it created a weird moment where I had to mentally switch frames. Am I the buyer of the software or the buyer of the business idea? The page tries to serve both audiences and I am not sure it succeeds with either.

## What would convince me

One case study from a nonprofit in the $1M to $5M annual fund range. Not "a mid-size organization saw results." I want: here is the campaign type (email, direct mail, digital), here is what the tool flagged on day 5, here is what the team did differently, here is the delta in final performance. Even one. Even anonymized with their permission.

Also: I want to see what integrations are actually live versus planned. If this does not connect to Salesforce NPSP or Virtuous or Raiser's Edge out of the box, then for most development shops it is a spreadsheet import tool, which is a different product than what the hero implies.

## What I'd ask in an email reply

1. What does day-1 data ingestion actually look like for a shop running Salesforce NPSP and Mailchimp? Is there a native connector or is someone exporting CSVs?
2. The page says "Training Data Reuse -- every correction you make teaches the model." How much historical campaign data do you need before the predictions are better than a human gut check? We have about 18 months of clean data and 3 years of messy data.
3. The ICP says "nonprofit ops leads with annual budget $500K to $10M." We are at $2.4M. What is your actual install base right now, even in beta, and does any of it look like us?

## Verdict: on-the-fence

The headline earns attention and the honest scoring panel earns some respect, but there is not enough proof that the product actually exists in a usable form yet. I would not delete this tab.

---
*Memo by skeptic persona, generated 2026-05-12. Studio breaks own self-grading loop.*