Ad spend that drives enrollment, not just clicks.
See which ads bring applications and where next year's budget should go.
Request a pilotIn practice
“We’re seeing great results from the budget reallocation (leads are up across most programs with lower CPLs so we’re getting a more efficient use of our spend).”
A small first step: one six-week pilot.
- 1
Send your exports.
Week one. Slate or your CRM, plus spend from Google Ads, Meta, LinkedIn and the rest. Not sure how to pull them? We'll do the export with you.
- 2
Daggy estimates each channel.
First weeks. What each channel did for each program, with the season and deadlines taken out.
- 3
Get a proposed allocation.
Week six. By channel, program and month, and what your enrollment goals would take, with ranges. Your school decides what moves.
Sized for a school’s budget, not an enterprise’s.
Request a pilot
Ad platforms grade their own homework. Daggy is independent.
Daggy doesn’t buy media, so it has no stake in where your budget goes. It works alongside your agency.
Questions
Do you need student records?
No. Daggy works from totals by program, channel and week: leads, applications and spend. No names, emails or student-level rows.
Where does our data go?
Into Daggy’s own encrypted AWS account. Thirty days after the pilot it’s deleted, and you can have all of it back first.
Our IT office will have questions.
Send them. We’ll sign your data agreement and fill in your security questionnaire.
Who decides where the money goes?
You do. Daggy proposes; program priorities and the final split are your school’s call.
Can you also handle the data work?
Yes. We clean Slate and CRM exports, join them to ad-platform spend and run it on a schedule. If you need something bigger, we’ve done large data engineering jobs before.

More than a decade of solving messy problems with data, in industry and academia:
- Eventbrite
- Built revenue forecasts for the CFO and measured campaigns for marketing teams, to help navigate the IPO and COVID.
- Pendulum Intelligence
- Trained and deployed models that read millions of pieces of video and audio a day, to identify risks to people, infrastructure and brand.
- UC Santa Barbara and Oxford
- PhD focused on causal inference from observational data; postdoc at Oxford in political economy and experimental research.
From the blog
Synthetic agents are easier to manipulate in experiments than humans