The First 30 Days Are for Taking Control, Not Taking Inventory

In my first week at Cubii I let the agency go and cut the search budget from about $20,000 a month to $2,000, with $1,200 of what was left going to branded terms only.

That sounds reckless for a week-one decision. It wasn't a close call. We were selling an under-desk elliptical, and nobody was searching for an under-desk elliptical, because the category didn't have a name yet. We were buying awareness for a product category that did not exist, on a channel that only captures demand somebody already has. Seeing that doesn't take a quarter of context. It takes the keyword report and one honest afternoon.

I had been brought in to turn a delivered Kickstarter campaign into a business. The honeymoon wore off in February, and nobody had quite named it yet, but we were putting something like 40% of revenue into customer acquisition. The company was losing about $40,000 a month.

We moved the money to discovery channels, where you introduce people to a thing they can't yet search for. Organic searches for the brand climbed to roughly 10,000 a month. We were profitable in about sixty days, and Cubii went on to a $100M exit.

The profitability is the part people ask about. The part that mattered is that the money we stopped burning in week one is what paid for everything we learned in weeks two through eight.

Taking control is not the same as changing everything. It means stopping what is demonstrably broken while the slower work of understanding the business runs.

The advice every new growth leader gets is to do nothing

The standard counsel is to spend your first ninety days listening. Change nothing. Build trust. Bring your plan at the end of the quarter.

I understand where it comes from. Arriving and breaking things you don't understand yet is a real failure mode and plenty of people have done it.

The problem is that it prices every month the same, and they aren't the same. If a company is spending into something that returns nothing, ninety days of listening has a cost, and the company pays it while you build rapport. That number rarely appears in anyone's onboarding plan.

Some things need the month. Some things don't.

Real diligence is slow and should be. Which cohorts behave nothing like the blended average. Where the funnel actually leaks rather than where everyone assumes it does. What a churned customer says when you stop asking whether they like the idea and start asking what they already spent trying to solve the problem. None of that resolves in a week, and rushing it produces confident nonsense.

Spend that is buying nothing sits in a different category. At Cubii it was a structural mismatch between the product and the channel, not a bidding problem, and no further study was going to change the answer.

So I separate the two. The slow work gets the month it deserves. The obvious waste gets cut while the slow work runs, because the one funds the other.

What agency "best practices" tend to leave behind

When we took over growth for CamDo Solutions, a B2B manufacturer selling photography equipment into construction, they had already hired a marketing manager out of a large national agency and were seeing no digital growth. Good SEO rankings. Decent organic traffic. A site that wasn't built to sell.

The strategy was education and nurturing across the whole catalog. Free guides, calculators, content downloads. It produced tens of thousands of visits a month and fewer than ten transactions.

That is what you get when marketing isn't accountable to revenue. It's an activity system rather than a revenue system, and it photographs well in a monthly report. We cut the fluff, concentrated spend on the handful of products that actually carried revenue instead of spreading it across the catalog, and pushed hard on paid social and AdWords. That month's digital revenue came in 62.7% above the same month a year earlier. I wrote up the full teardown of that funnel separately, in curing merry-go-round sickness.

Nothing there required another quarter of analysis. It required somebody willing to act on what the numbers already showed.

The decision that can't wait for the full picture

The RealReal's paid social operation was spending heavily without much efficiency when I picked it up. I rebuilt the paid acquisition team and brought the functions in-house.

Not because agencies are inherently the problem. Because you cannot optimize what somebody else keeps in a black box. When an outside team owns the account structure and the reporting, every question about your own business gets answered by the party being evaluated.

This is the part that gets read as a contradiction, so let me be plain. The decision was made in month one. The migration took a quarter. Collapsing those two is how people talk themselves into believing they have to choose between moving fast and moving carefully.

We hired a senior paid acquisition lead first, had them shadow the agency long enough to absorb the account structure and the institutional knowledge, then moved campaigns one channel at a time starting with the highest spend. The transition ran about ninety days. By the end we had the accounts, the data and the accountability.

What came after is the argument for doing it: paid scaled from $300K to $2M+ per month at 7x ROAS, with CAC down 40%, and the channel drove $100M+ a year in attributed revenue. That was possible because we could finally see what was true.

I don't always take it in-house

The rule I use, and I've been on both sides of it.

Keep the agency when you're early and genuinely can't staff a team yet, or when they hold specialized expertise your people don't have, or when they work like an extension of the team rather than a vendor. Embedded, in your standups, operating on your first-party data. That relationship is worth protecting.

Replace them when they're managing your spend but can't explain your unit economics. When they optimize to the number inside the ad platform instead of revenue in the business. When the monthly report is something you nod along to rather than understand. When they resist giving you full access to accounts you pay for. Any one of those is a flag. All of them at once is more common than you'd expect.

If you're carrying line items nobody has been able to explain to you in plain language, bring them to a 20-minute conversation. We can usually tell whether you need more analysis or just need to stop paying for the wrong thing.

Why this order is uncomfortable

Cutting in week one costs you something socially. You're the new person removing a line item somebody chose, sometimes in front of the person who chose it. Standing up in month two to say the most valuable thing you did was stop paying for something doesn't feel like a win, and it doesn't present well next to a new channel launch.

There is a version of that objection I take seriously. The political capital you spend in week one is part of what decides whether you are still there in week eight. The sequencing is my answer to it. What I cut immediately is spend nobody can defend in plain language, because that argument is short and the numbers make it for me. What I move slowly is anything with a person or a working system attached, which is why the RealReal migration got a quarter and a parallel run rather than a memo.

It buys the only thing worth having in month one, which is room. Room to run the slow analysis, talk to the customers who left, and get to a point where the numbers in front of you are honest.

I should be honest about what makes the fast read possible, because it isn't nerve. It's having watched the same few structures break the same way for most of a career. Spend chasing demand that doesn't exist yet. Budget spread so thin across a catalog that nothing ever clears learning. An outside party holding the data you're being asked to decide from. That pattern recognition is the asset, not the decisiveness, and if you haven't built it yet the honest move is to verify before you cut and say plainly that you're verifying. Taking two weeks instead of two days is not the failure. Taking ninety days because nobody wants to be the person who says stop is the failure.

People read this as patience, and it isn't. Killing most of a search budget in week one is not a cautious act. I am aggressive on day one. The aggression just points at the spend before it points at the market, because there is nothing to be aggressive with while the money is going out the door for nothing.

What changes by month three is the target, not the temperament. You own the machine, the waste is gone, the measurement holds, and now the same appetite goes into buying more of what works. Spending hard is a lot easier when you can see what it returns.