Most corporate venture studio RFP responses look alike. Six demands tell a real partner from dressed-up consulting: IP, data, stop discipline, equity.
Most proposals read the same. The right questions are what tell them apart.
If you're putting a corporate venture studio engagement out to RFP, you'll get back a stack of decks that look broadly identical. Confident methodology slides. A funnel diagram. A logo wall. Numbers that all seem to point the same direction.
The decks aren't the signal. The signal is what a respondent says when you ask the questions below, and how fast they say it. A real studio partner has answered all of this before, in writing, in front of a counterparty's lawyers. A consultancy wearing studio language will hesitate on at least half.
Six questions belong in the document. We'd answer all of them on the first call, and so should anyone you're seriously considering.
This is the question that ends more bad engagements than any other, and it has exactly one right answer: the new venture owns its IP. Not the studio. Not, by default, your institution. The operating company.
Why that matters: if the studio retains a license back to the IP, or your institution holds it on the venture's behalf, you've created a problem for every future investor. Clean IP ownership inside the venture is consistently cited as one of the biggest factors in whether outside capital will follow on. A cap table or IP structure that needs untangling before a Series A is a structure that doesn't get a Series A.
What you negotiate separately, and should: a right of first refusal on follow-on rounds, board representation, distribution rights, and an agreed methodology for acquiring the venture outright later. Those are yours to win. The base IP is the venture's. If a respondent wants a license back to themselves, ask why, and read the answer carefully.
For a regulated FI this is a gating requirement, not a preference, and the wrong answer should disqualify a respondent on the spot.
The structure you want is a data license, not a data transfer. Your customer data stays in your environment, under your controls. The venture reaches it through an API with permissions you set and can revoke. The studio's validation and analytics run on infrastructure you can locate on a map: ideally self-hosted in-country, not on a US hyperscaler by default.
Make them be specific. "We take data security seriously" is not an answer. "Your data stays in your tenant; the venture accesses defined fields via permissioned API; our validation models run on self-hosted compute in [city]" is an answer. The difference between those two sentences is the difference between a partner who has cleared this with a CISO before and one who hasn't. (We go deeper on where regulated customer data should actually live in a separate piece, because for a risk officer this is the question to lead with.)
If a respondent can't tell you which building your data is processed in, they haven't done this with a regulated institution.
Every respondent will quote you a price per venture. That number is close to meaningless on its own, because it ignores the failure rate.
Ask for the real one. Take the cost of a venture attempt and divide it by the share of attempts that clear validation. That's cost-per-validated-venture, and it's the figure that compares honestly across models, because it prices in the misses.
Run the math yourself with their inputs. The industry benchmarks are public: studio attempts land around $250–500K each with a roughly 45–60% validation rate, which puts a validated venture near $420K–$1.1M (GSSN; BCG; McKinsey studio economics, 2024). An internal lab attempt runs $2–5M at a 10–15% rate, which puts a validated venture anywhere from $13M to $50M (McKinsey; Deloitte). The gap between those two columns is the entire argument for the model, and any respondent worth hiring will show you their version of that division without being asked twice.
If they only ever quote cost-per-attempt, they're hiding the denominator. (We walk the full cost-per-validated-venture teardown elsewhere, with both columns built from the bottom up.)
A studio's discipline is most visible in how it ends things, not how it starts them. So make the stopping mechanism an explicit line item in the RFP.
Ask for the decision cadence: how often a venture faces a real go / hold / stop gate. Ask for the specific metrics that trigger each outcome (conversion thresholds, retention, signed letters of intent, unit-economics tests). Ask what a stopped venture costs and how the decision gets made.
The answer you want is a scheduled, evidence-based gate where stopping is a normal outcome with a known price tag: under $50K when it happens early (AOS operating model). The answer you don't want is anything that sounds discretionary, relationship-dependent, or open-ended. A partner who can't tell you precisely how they stop a venture will struggle to stop yours, and you'll be paying for the months it takes them to admit it.
This separates the two business models hiding behind similar slides.
A consultancy bills time and walks away when the engagement ends. Its incentive is engagement length, not venture outcome. A studio takes equity in what it builds, so it only wins when the venture wins, years after the retainer is spent.
Ask directly: do you take equity in the ventures, and how much? The benchmark range for studio stakes sits around a median of 17%, and notably, studios that take less than 20% tend to show the strongest outcomes. A high stake can starve the founder and the option pool that later rounds depend on (BVSR'24). So you're not looking for the lowest number or the highest. You're looking for a stake that's real enough to align them and modest enough to keep the venture financeable. A respondent who wants only fees, or who wants a controlling stake, is telling you which game they're actually playing.
The last demand is the one that makes the rest enforceable. Ask for a written split of decision rights before anything is signed.
The clean version is simple to state. Your institution approves the venture thesis, the sectors, the compliance requirements, the brand guidelines, and the criteria for stopping. The studio decides team composition, product roadmap, technical architecture, and execution. A joint board meets on a fixed cadence to hold the seam between those two.
That separation is the mechanism, not bureaucracy around it. You keep strategic control without inheriting the political gravity that stalls internal efforts; the studio keeps operational control without waiting on a committee. Any respondent who blurs this line, or wants approval rights over execution, is rebuilding the exact structure that makes internal labs slow. (We've written separately on why that governance line carries the whole engagement.)
Six demands. Print them on one page and send them with the RFP:
A genuine studio partner answers all six on the first call. Everyone else needs to "circle back." That tells you most of what the RFP was meant to find out.
Nothing here is an offer to sell a security or investment advice.
Field notes on venture building, AI, and capital. No spam, unsubscribe anytime.
By subscribing you agree to receive AOS Insights e-mails. We use your address only for this newsletter - see our Privacy Policy.
