Five questions that SOLVE business problems.

Surface · Observe · Learn · Vet · Earn. One framework for consulting cases, product builds, and research projects — with drills and simulations to practice every move, ending in a capstone where you build a working pricing tool from real-style data.

The premise

Every business project is a claim about value: “if we do X, it will create value because Y.” Consulting engagements, product and software builds, analytics and research projects, process improvements — the tools differ, but the logic doesn't. Your job is always to make that claim precise, test it as cheaply as possible, revise it until the evidence holds, and deliver it in a form your audience can act on.

The catch is that nobody hands you the claim. Clients ask for builds — “build us an app,” “cut our costs,” “make a dashboard.” Executives care about the bottom line — revenue, cost, market share. The work that actually connects the two happens at a level most people skip: behavior — the specific changes in what people do that create business results. Learning to work at that level — and to move deliberately between all three — is what this framework teaches.

What you'll practice

The framework braids together three core approaches every business student should know — the product manager's, the consultant's, and the scientist's. On the surface they look like different jobs; underneath they share a logic, and SOLVE is built on what they have in common:

• Behavior-level thinking (from product management) — targeting measurable changes in what people do, not features shipped or decks delivered, and refusing to call a project valuable just because the build is finished.
• The case method (from consulting) — opening an ambiguous problem, structuring it MECE, driving the analysis with a hypothesis, and delivering a conclusion-first recommendation. This is also, not coincidentally, exactly how case interviews work when you recruit.
• The scientific method — treating every recommendation as a claim about cause and effect: ask a precise question, define what you'd measure, theorize the drivers, and go looking for the evidence that could prove you wrong.

You'll practice in three settings of increasing realism: quick drills in class, interview-driven simulations in the Wisconsin Case Lab, where AI plays your clients and the people around them, and — in the later units — exercises where you analyze data and build working tools with AI coding agents, using the skills from the companion vibecoding course.

How to use this site

1. Read the framework once, start to finish. It's short on purpose. Don't memorize it; you'll internalize it by using it.
2. Work the units in order. Units 1–5 each train one question with a drill or two and a short simulation. Unit 6 runs the whole arc twice — once as a classic diagnosis case, once as a data-and-code build.
3. Browse the sponsor scopes to see the framework at full scale. Twenty-one real-style project scopes read through the five questions — worth a look before your first simulation, and again whenever a real sponsor's scope lands in your lap.
4. Keep the reference open during simulations. That's what it's for. Using the templates isn't cheating; it's the point.

One warning before you start: this framework will feel linear on the page and will never be linear in practice. Answering a later question routinely reopens an earlier one. That looping is the method working, not failing — and in this course, documented loops earn credit.