A note before a call · Higher-ed / OPM · Prepared for iDesign

I built a couple of small things on public data, so a first call has something to look at.

I would rather start with a short call to hear what iDesign actually needs. But you are probably swamped, so I did some homework first: two rough prototypes on public higher-ed data and synthetic numbers, sketching a problem the OPM world is wrestling with now. Not your data, not a finished product, just a faster way to show what working together could look like than a blank-page call.

Fee for service
iDesign's own founding model: paid for service, not a tuition share (Inside Higher Ed)
2023 ED review
Education Dept review of the bundled-services exception puts revenue-share OPMs on notice (Higher Ed Dive)
Public data only
The prototypes run on IPEDS and College Scorecard, plus synthetic fixtures, all labeled

What I think is going on (happy to be wrong)

iDesign pioneered the unbundled, fee-for-service OPM model: paid for the work, not a cut of tuition. That honest model wins renewals on one thing: showing a partner, in numbers they trust, that the program enrolled the right students at a defensible cost. The catch is structural. The cheapest conversion a partner program can measure is the inquiry, and a scorecard built on cost-per-inquiry flatters everyone and proves nothing. The partner owns the enrollment P&L; iDesign owns the renewal. So the question I would want to ask on a call: across the portfolio, can iDesign demonstrate cost-per-enrolled-start with a method it owns, or is the reporting mostly inquiries?

The two prototypes

  • A cost-per-enrolled-start optimizer: reallocates one partner program's media buy under per-channel ceilings (the boast) and a total cap (the bound), scored on enrolled starts instead of cheap inquiries. A partner-program selector swaps in the next program's public anchors and reruns the same method, so it reads as portable, not a one-off.
  • A partner-data readiness console: before optimizing anything, can iDesign even see enrolled starts for a partner (your media joined to their SIS)? You cannot show a number you cannot see, so this is the honest first read.

Both are worked examples of a method, not a deliverable.

An honest aside, since you will think it

Your own team could build a media optimizer; this is not magic, it is a linear solve. What an internal team cannot easily do is audit its own inquiry-optimized buy and hand a partner a cited, independent exhibit. The value here is that independence, the speed of a fixed-scope sprint, and a track record in higher-ed enrollment marketing, not raw capability.

The honest bound: the prototypes show the shape of the inquiry-versus-enrolled-start gap on public data. They do not size the real gap for any partner; only the partner's SIS join does, and the readiness console is exactly about whether that join exists. Wiring it to real numbers is the follow-on, not a claim made here.

The hunches behind each prototype

1. The cheap-inquiry trap.
A partner's lowest-cost-per-inquiry channel is often its worst on enrolled starts (tire-kickers who never apply), so an inquiry-optimal buy quietly costs enrolled starts. The optimizer sketches the reallocation; a real engagement checks it on the partner's data.
2. The good-fit program channel.
In a real-labor-demand credential (a nursing program, BLS-validated), the program-aligned channel over-indexes on enrolled starts but is underfunded on an inquiry objective. The optimizer moves budget toward it: the anti-predatory-volume story, not volume at any cost.
3. The data lives in the partner's SIS.
For some partners the enrollment-outcome join is not available, so cost-per-enrolled-start cannot be shown yet. The readiness console says so plainly and makes the data-sharing scope the first deliverable, not an optimizer that cannot run honestly.

If this is worth a few weeks

WK 1-2Funnel model for one partner program from public IPEDS, College Scorecard, and BLS demand, inquiry to enrolled start, as a bounded range
WK 2-3The optimizer on that program, beside the inquiry-optimal plan, with the partner selector for portability across the portfolio
WK 3-4Map the substantiation surface (recruitment and outcomes claims, the partner's 90/10 and FTC posture) onto the channel mix
WK 5-6Partner-data-readiness read on the SIS join, and the cadence that lets your team run the method across partners. IP transfers

Pricing

Diagnostic: fixed 4 to 6 weeks. The working optimizer on public data, the resellable method, the gap read, the data-readiness memo. IP transfers.$50 to 75k
+ Internal extension: a partner's real per-channel and SIS data swapped in, the cadence stood up with named owners across programs.to ~$150k

Fixed-scope, one-time; the IP transfers and your team owns the cadence. No platform, no subscription. A retainer exists only if a real queue of distinct work across partners justifies it; never the default. iDesign is private, so I cannot size a dollar prize off a disclosed line: the band is a hypothesis to test per partner, not a promise. Indicative ranges; final scope set on the call.

The ask

One 30 to 45 minute call. Bring nothing; the demo runs on public IPEDS and College Scorecard data for one of your partner programs. I will load the optimizer and, in one click, show the spread between an inquiry-optimal buy and an enrolled-start-optimal one, then switch partners to show the method is portable. If your reporting already demonstrates enrolled starts, you will have spent 45 minutes confirming it. If there is nothing there, I will say so.

Book it: jeffpinto.com/engage · Method: the gcu-media-planning note

Who's behind this

Jeff Pinto runs a small, independent data and AI advisory practice (jeffpinto.com). Thirty years across AI data and privacy, health tech, marketing analytics, renewables, logistics, and broadcasting; the last seven in ML and AI. Hands-on at Meta, Uber, and IBM, plus six startups (one turnaround, three acquisitions). Two MScs: computer science (Toronto) and engineering (Loughborough). Engagements are fixed-scope, four to twelve weeks, no platform and no subscription; whatever gets built, the IP transfers to you.

The relevant slice for iDesign: from 2009 to 2015 I ran enrollment-marketing analytics for higher-ed clients at Sparkroom and Blue Camel, the same inquiry-to-enrolled-start funnel, in an industry shaped top to bottom by gainful-employment and incentive-compensation rules; the boast-and-bound media method here comes straight out of a 2012 GCU paid-media pilot from that era.

Sources: Inside Higher Ed ("3 Questions for Whitney Kilgore"; "OPMs: Fee for service is growing," 2017) · Higher Ed Dive, "Education Department to review guidance allowing revenue-share agreements with OPMs" (2023 bundled-services review) · iDesign company materials · public higher-ed data (IPEDS, College Scorecard, NCES College Navigator, BLS OOH). iDesign is private; no revenue or cost-per-anything is disclosed, so any dollar figure is bounded, not sized. CEO and co-founder titles to confirm before external use.

Built by Jeff Pinto: higher-ed enrollment-marketing analytics at Sparkroom / Blue Camel · Meta / Uber / IBM + 6 startups · two MScs · jeffpinto.com