Field Notes
You Don't Need an Agency. You Need an Operator.
Watch a common two-year arc.
Year one, the owner of a good company buys a website. Looks sharp, loads fast, everyone's happy. Traffic is thin, so an ads retainer comes next. Then a social vendor, because the competition posts and it feels wrong not to. Somewhere along the way: an email platform, a CRM, a video package.
Year two, the stack is complete. Five vendors. Five monthly reports. Five invoices clearing on the first. Revenue sits about where it sat before the website.
No one got robbed. Every vendor can defend every line of every report. Which is precisely why the problem is hard to see.
Every vendor did their job
Go down the list and check. The web shop shipped on schedule. The ads agency hit its cost-per-lead number. The social vendor posted four times a week and the follower count climbed. All true. All documented.
Look closer at what each report covers, though. The ads report tells you what a click cost. It says nothing about whether those clicks turned into signed contracts, whether the leads sat unanswered over a long weekend, or whether the CRM tagged their source at all. Those questions live in the gaps between vendors, and the gaps are the one place no invoice covers.
Vendors report on deliverables because deliverables are what they sold. Fair enough. But a business doesn't run on deliverables. It runs on the compounding of them, and compounding is nobody's line item.
The job you never posted
Here is the position that was never filled: the person accountable for whether the whole stack, taken together, produces more than it costs.
Usually that job does get done, sort of. It gets done by the owner, at nine at night, squinting at five PDFs that don't share a single metric. The owner becomes the integrator by default. Unpaid, untrained for it, and busy running the actual company. In a custom-build shop or a med spa, the owner's hours are the most expensive hours in the building. Spending them reconciling vendor reports is a bad trade.
No agency sells this seat, and there's a structural reason. Filling it honestly means sometimes recommending against your own invoice. An ads agency will optimize your campaigns for as long as you'll pay. It will not call in March and say pause everything until the follow-up process stops leaking, because that call kills the retainer. The advice is capped by the business model.
An operator's product is the call itself.
Two kinds of reports
A deliverable report reads like this: eight posts published, impressions up 40 percent, engagement steady.
A business report reads differently: three jobs closed this quarter, here's the channel each one actually came from, here's total spend against them, and here's the line item that produced nothing and gets cut next month.
The first kind can stay green forever while the company stalls. The second kind occasionally embarrasses somebody, and that's how you know it's real. A simple test: if nothing in your monthly reporting could ever make one of your vendors look bad, you're holding the first kind.
Why the pay is tied to revenue
Revenue Rx's executive seat is a fractional CTO/CMO role priced as a low base plus commission on revenue the work produces. Attribution gets defined in writing before the engagement starts, not negotiated after a good quarter.
People sometimes read the low base as a discount. It isn't. It's the mechanism that keeps the advice clean. When most of the pay rides on produced revenue, telling an owner to cancel a spend costs me nothing and stands to earn me plenty. There's no retainer to protect, so there's no reason to defend a line item that isn't earning, including one I recommended myself. Wrong calls come out of my check, not just yours. Skin in the game isn't a slogan here; it's the fee schedule.
The written attribution rules matter as much as the split. If produced revenue can't be defined cleanly for your business, that usually means the tracking doesn't exist yet. Building it becomes the first piece of operator work, because nobody can be held accountable to a number nobody can see.
When not to buy this
If you need one specific thing done well, buy that thing. A new website, with someone in-house to run it afterward: hire a good web shop and shake hands. A rebrand, a single campaign, a photo library. Well scoped, clearly needed, one vendor. Done.
The operator seat solves a different problem. It exists for the company with several vendors or channels, real money going out monthly, and nobody on the owner's side of the table who can say which parts are working. If reading your five reports leaves you knowing less than before you opened them, that's the tell. Paying for judgment you don't need is the same mistake as skipping it when you do.
Ask them what to cut
Try this at your next reporting cycle. Ask each vendor one question: what would you cut? Not what deserves more budget. What goes.
A vendor who names something real, even inside their own scope, is worth keeping. Hedging is an answer too. And if nobody in the entire stack can tell you what to cut, you've found the vacancy. Not another channel. Not a sixth vendor. One person, on your side of the table, accountable for the whole number.