Diligence before the cheque. Capability after it.
Marketing is usually the largest cash line in a growth stage company and the least examined one at diligence. It is also where a fund can add the most operating value post investment, and where most platform teams are stretched thinnest.
One relationship. Three different engagements.
Marketing due diligence
- Independent read on acquisition efficiency inside your diligence window
- Read only access, no dependency on the target's team beyond that
- Findings written for an investment committee
- Fixed scope, fixed fee
Portfolio operating support
- The seven studios available to portfolio companies at portfolio terms
- Companies engage individually. The fund does not carry the commercial relationship
- Useful where a company has budget and no specialist bench
- Reported back to the fund at whatever level of detail the company agrees to
AI capability deployment
- One governed AI operating stack, deployed company by company
- Built once, adapted per company, isolated per tenant
- The most useful thing a platform team can offer right now
- Runs as a program, not a series of introductions
Most funds arrive for the first and stay for the third. Diligence is a transaction. The portfolio programs are where the relationship actually sits.
The checklist, published.
Publishing this costs the mystique and saves a scoping call. If a target is clean across all seven you will know inside a week and we will say so plainly, which is worth as much as finding a problem.
True blended CAC against reported CAC
What is excluded from the denominator. Organic attributed to paid, brand spend held outside the calculation, agency fees counted separately from media, referral treated as free. The gap between the two numbers is often forty percent and almost never deliberate.
Channel and creative concentration
What share of acquisition depends on one platform, one campaign type, or a small set of creative currently working. A company looks efficient right up until the concept that carried it stops converting, and the replacement pipeline is the thing to examine rather than the current number.
Attribution integrity
Whether the model flatters paid. Whether platform reported conversions have ever been reconciled against CRM or finance. In most growth stage companies the answer is no, because nobody has asked and the founder assumes the dashboard is authoritative.
Spend efficiency against category
CPM, CPC, conversion rate and cost per qualified lead benchmarked against what the vertical actually pays, not against the company's own trailing performance. Improvement against yourself is not the same as being competitive.
Measurement and instrumentation quality
Server side implementation, deduplication, consent handling and offline conversion feedback. Poor instrumentation caps how well the company can perform regardless of who runs the account, and it is expensive to fix retroactively.
Team and vendor dependency
Whether the acquisition engine survives one person leaving or one agency contract ending. Where the accounts sit, who owns the pixel history, and what walks out of the door with a vendor.
Headroom at three times spend
What happens to marginal CAC when the budget triples. This is usually the number that decides whether the deal works and almost never the number anyone has modelled, because it requires a saturation read rather than a trailing average.
Turnaround is built for a diligence window. Read only access to ad accounts, analytics and CRM on day one. Written findings on day seven. One call to walk through anything contested. Access is the only dependency on the target's team.
Build it once. Deploy it twenty times.
Every portfolio company is separately trying to work out how to use AI in marketing, badly, in isolation, with no governance and no documentation. The result is twenty different half implementations and no transferable learning. A fund is the only actor positioned to fix that, because the same operating stack fits most of the portfolio with per company adaptation.
Standard stack, isolated tenants
One architecture designed once. Each company runs its own instance with its own data, permissions and approval gates. Nothing crosses between portfolio companies.
Company by company rollout
Sequenced by readiness rather than by fund preference. A company without clean measurement gets that fixed first, because deploying on top of bad data produces confident nonsense.
Governance the fund can point to
Audit logs, permissions, model evaluation and failure state handling documented per company. This matters at the next diligence, when an acquirer asks what the AI actually does.
Shared learning, unshared data
What worked at company four informs company nine as a pattern, never as data. The playbook travels. The information does not.
The same logic applies to the studios themselves. A portfolio wide creative operation, a shared measurement standard, or one performance media practice across eight companies produces better economics than eight separate agency relationships, and gives the fund a comparable read across companies for the first time.
A program, not a list of introductions.
Portfolio read
A short assessment across companies to find where the common problems sit. Usually measurement quality and creative production capacity, in that order.
Sequencing
Companies ordered by readiness and by how much the intervention would move their numbers. A company nine months from needing this is worse to start with than one that needs it now.
Company engagements
Each company contracts directly, on portfolio terms. The fund is not a party to the commercial relationship, which keeps the fund out of the middle when something needs to change.
Fund level reporting
At whatever depth each company agrees to. Some funds want a comparable acquisition efficiency read across the portfolio. Some want nothing beyond confirmation that the work is happening.
Stated before you ask.
We do not run diligence on a company while a studio is engaged with it. We do not run diligence for two funds on the same target. Where a portfolio company competes directly with an existing studio client, we disclose it and let the company decide rather than the fund.
Diligence findings belong to the fund that commissioned them and are not reused, including as anonymised benchmark data.
Request a briefing