Diagnostic engagement · Cross-brand enrollment-marketing efficiency · Prepared for Covista

I'd love to start with a call. But you're probably swamped, so I did some homework first.

A few small prototypes built from Covista's public filings, showing the kind of work I've done applied to a problem your industry is wrestling with right now. They run on public and synthetic numbers, not your internal data: rough sketches, not a finished product, and a faster start than a blank-page call.

$247.4M

FY25 advertising: 13.8% of revenue, one bundled expense line, no public cost-per-start per brand

13.6% / 2.2%

Q1 FY26 enrollment growth, Walden vs. Chamberlain: management named the gap; the cause is still a closed number

External data asked for: all figures come from the 10-K, earnings 8-K, IPEDS, and Scorecard

The structural problem

$247.4M in advertising, up three years running, reported as one lump with no per-brand split and no cost-per-start. Walden's online-grad funnel competes head-on with WGU and SNHU; Chamberlain's nursing demand is pulled by a workforce shortage. Those aren't the same cost-per-start. The Q1 FY26 CEO quote ("underperformed in local marketing effectiveness," Walden +13.6% / Chamberlain +2.2%) named the gap publicly; the spend behind it is still a closed number. At 19.1% operating margin, the board will ask how much of three years of spend growth bought new starts versus rode the tuition increase.

What the prototypes sketch

Cross-brand cost-per-start radar (worked example of the method): allocates the disclosed $247.4M using public revenue and enrollment data to show an implied cost-per-start for each brand as a bounded range. The per-brand spend split is synthetic; real per-brand spend is the internal extension the diagnostic scopes.
Constrained efficiency frontier: where the next dollar buys the most qualified starts once 90/10, Gainful Employment, and FTC rules close off the moves a pure-ROI optimizer would make.
EBITDA-to-enterprise-value bridge: converts a point of efficiency on the $247.4M line into dollars to EBITDA and enterprise value at a public-company multiple. The currency a board prices in, not just a ratio.

The wedge vs. the agency bid optimizer

Agencies are good at placing media once the budget's set. What they can't sell you is a read that might say spend less: their incentive ends in more managed spend. A black-box optimizer is a non-starter in a 90/10 + GE + FTC environment, where which-students-you-recruit is a compliance question. This is the opposite posture: every input is a citation, IP transfers, your team audits every step. And an outside read is allowed to conclude "spend less"; an internal model recommending cuts to last year's allocation is a career-hard memo to write from inside the org chart.

Why now

Margins at 19.1%: CFO has proof discipline pays, so the ad line will be read under a microscope.
Marketing reorganized in 2026; cross-brand efficiency is now one owner's question.
CEO publicly named the Chamberlain gap on Q1: the window for an independent read is open.

The hunches behind these

Radar: Chamberlain's nursing-demand tailwind means its cost-per-start is structurally lower than Walden's auction-priced funnel; the blended line hides a real efficiency spread. If real data shows comparable cost-per-start across brands, the diagnostic says so: you've confirmed the allocation is already right.
Frontier: the compliant optimum (with 90/10 and GE constraints) is materially different from the naive spend-reallocation a pure-ROI optimizer would make. An outside read with no incentive to dodge the compliance inputs is structurally better positioned to find it.
EBITDA bridge: the efficiency story isn't being told in board-grade currency. A point of efficiency on $247.4M is roughly $10M to $25M freed; at 12 to 16x that's $120M to $350M of enterprise value. That reframes the conversation from a line-item cut to an investment decision.

The 4-6 week diagnostic

WK 1-2Build the cross-brand model from the 10-K, earnings 8-K, IPEDS and Scorecard: implied cost-per-start per brand, bounded and labeled synthetic

WK 2-3Map the regulated-marketing constraints (90/10, GE) onto the brand mix: efficiency frontier with compliance flags, on public numbers

WK 3-4Build the EBITDA-to-enterprise-value bridge on disclosed numbers: efficiency sized in the currency the board reads, with the honest range

WK 5-6Data-readiness read on the internal extension (real per-brand spend, channel cost-per-start) + quarterly cadence that makes the lever durable: phase-2 scoping memo, honest about what's missing

Engagement shape

Fixed scope, 4-6 weeks, one diagnostic. Ships the working prototypes on public and synthetic data, a data-readiness read on extending the work inside your stack, and a phase-2 scoping memo. All IP transfers to Covista. No platform, no subscription; no internal data is asked for or leaves the building.

Indicative range: $50-75K for the diagnostic; up to ~$150K if the internal extension lands in a second phase. Commercials in a one-page SOW after a 30-minute call.

Who's behind this. Jeff Pinto runs a small, independent data and AI advisory practice (jeffpinto.com). Thirty years across AI, 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: CS (Toronto) and engineering (Loughborough). Fixed-scope, four to twelve weeks, no platform, no subscription; IP transfers to you. For this account: Jeff built Sparkroom (2009 to 2015), a marketing-budget-management SaaS for higher-ed enrollment teams, on this exact funnel and these exact cost-per-start economics, and presented at LeadsCon 2015.

The ask

One 30-45 minute working session. Bring nothing: the data's public. I'll walk the cross-brand cost-per-start read live on your FY25 filings and show where the method holds up and where it runs out. If the engagement scopes to nothing useful, I'll say so.

Live: covista-enrollment.pages.dev · Book: jeffpinto.com/engage · Fleet: github.com/bigbrownjeff

Sources: SEC EDGAR (Covista/Adtalem CIK 0000730464, FY25 10-K + earnings 8-K; advertising $247.4M, segment revenue/margin, enrollment figures are Covista's own public disclosures) · IPEDS College Navigator (Chamberlain 454227, Walden 125231) · College Scorecard · BLS OOH (RN demand) · AACN (nursing-school capacity) · FederalRegister.gov (90/10, Gainful Employment) · FTC · Per-brand spend split in the radar prototype is synthetic, calibrated to public anchors and labeled as such on-screen

Built by Jeff Pinto: Meta / Uber / Sparkroom + startups · Higher-ed marketing analytics · Two MScs. jeffpinto.com