The 90-Day Turn: How to Fix a Broken Business in One Quarter

You can turn around a broken business in 90 days. Not heal it, not scale it, but turn the direction of travel. The mechanics are simple enough to fit on a napkin, and the discipline required is harder than anything else you'll do that quarter.

Every broken business I've worked on has been broken in roughly the same two or three places, and the fix has almost never been the one the founder pointed at on day one. Founders tend to diagnose their business by the skill they're worst at. "I'm bad at marketing, so the problem must be marketing." Sometimes they're right. Usually the problem is one level underneath the one they're pointing at, and the real fix looks nothing like "hire another marketer."

The 90-day turn is three moves, in order, on a tight schedule, with a refusal to let anything else onto the list until those three are finished. This is the whole SOP. Scrappy, in this context, isn't cheap execution. It's the refusal to add a single new initiative until the three moves run cleanly, because cheap diagnosis beats expensive action every time.

Month 1: diagnose, don't execute

Most founders try to fix a broken business by doing more of what isn't working. A new ad campaign. A new salesperson. A new website. Another content hire. It feels productive. It rarely moves the number, because acceleration against the wrong diagnosis just gets you to the wrong place faster.

Month 1 is not for new execution. Month 1 is for writing down what is actually happening, in honest numbers, without flinching.

Specifically, you answer five questions.

1. What are revenues by channel, by month, for the last 12 months?
2. Where are leads actually coming from? Not where you think. Where they're coming from, by source, for every closed deal in the last year.
3. What's the conversion rate at each stage of your funnel, actually measured, not assumed?
4. What are the top three customer segments by revenue? By margin? Are they the same three?
5. Which of the founder's weekly activities produce money, which produce noise, and which produce comfort?

That last question is the uncomfortable one, and the most important one. Founders do a surprising amount of work every week that produces neither revenue nor growth, only the feeling of having worked.

Month 1 is the single hardest month of the whole turn, because the founder is used to feeling productive by shipping, and is now being asked to sit still and look at the numbers. Sit still anyway.

Month 2: cut first, then concentrate

By the end of month 1, you'll know, in hard numbers, what's working and what isn't. The shape surprises almost everyone. The top two or three revenue drivers are clearer than expected. The bottom half of activities are producing close to nothing, usually including things you were proud of.

Month 2 is the cut. Before you build anything new, you stop doing the things that aren't working. The pattern is usually obvious once the numbers are on paper:

• The channel producing 5% of revenue and 40% of the effort gets killed.
• The customer segment you chase out of habit but that doesn't make money gets dropped.
• The weekly meeting nobody remembers starting gets cancelled.
• The tool you pay for that nobody uses gets cancelled.

Cutting feels like losing. It isn't. It's reclaiming the capacity you'll need for the second half of the month, which is where the growth actually starts.

In the second half of month 2, with the cuts made, you pick one thing. One channel. One offer. One segment. Whatever the month-1 diagnosis says is producing real results. You double the resources pointed at it. Not 20% more. Double.

You are not growing the business in month 2. You are concentrating it. The growth shows up in month 3.

Month 3: systemize the thing that worked

By month 3, the concentrated thing from month 2 starts producing real results. Revenue ticks up, or the pipeline fills, or the retention curve bends. That's the dangerous moment. The temptation now is to diversify back into other channels "now that we have momentum." Don't.

Month 3 is for systemizing the one thing that worked. Turn the ad hoc version into a repeatable one. Write down the process. Build the reporting. Hire or reassign a junior to run it. Document the playbook. Automate whatever can be automated.

The test at the end of month 3 is binary. If the founder takes a full week off at the end of the quarter, does the thing still run? Not "could run." Does run, under its own momentum, without the founder touching it for seven days.

If the answer is yes, you've completed a 90-day turn. You haven't transformed the business. You've taken the single engine that actually produces revenue, cleared the dead weight around it, and turned it into a machine that operates without you. That's a real turnaround.

What usually happens next

The founders who stop after 90 days tend to drift back into the old pattern. Diversifying too fast. Re-hiring for lost capacity. Re-layering activities that produce comfort rather than money. The gains unravel over the next two or three quarters.

The founders who repeat the cycle, quarter after quarter, each with a tight diagnosis, cut, and systemize, end up with four engines running without them after a year. Each one compounding. At that point, you don't have a job anymore. You have a business. Which, not incidentally, is also the exact version that would sell for a meaningful multiple if you ever decided to engineer for exit.

What I'm not promising

I'm not promising you a 10x in 90 days. I'm not promising you a category-killer brand, or a viral moment, or a funding round. I'm not promising that any of this is easy.

I'm promising that the difference between a broken business and a healthy one is almost never a new initiative. It's a diagnosis, a cut, and a system. In that order. On a schedule. With the discipline not to let anything else on the list until those three are done.

Ninety days is enough time to change the direction of travel of a business. Most founders have never actually tried the full sequence. If you think yours might be one of them, write down the answers to the five month-1 questions this week. That alone will tell you more than any agency proposal ever will. And if you want a second set of eyes on the diagnosis, that's most of what I do at Naughton Ventures.

FAQ

Does the 90-day turn work for any type of business?

The framework works across most service, product, and small business categories. The specifics of the five month-1 questions stay the same, though the channels, segments, and funnel stages you're measuring will vary. Capital-heavy or inventory-heavy businesses sometimes need a slightly longer cycle, closer to 120 days, to see the results of month-2 cuts fully land.

What if I can't answer the month-1 questions because I don't have clean data?

Then month 1 becomes about building the measurement you're missing, not about using it. Don't skip the diagnosis. Build whatever tracking you need, even if it's a messy spreadsheet, and make next quarter the real diagnosis month. Flying blind for another year is more expensive than spending this one instrumenting.

Can I skip the cut in month 2 if everything feels like it's working?

Almost nobody gets to skip the cut. If everything feels like it's working, the month-1 numbers will still show you two or three activities that are generating less than you assumed. The cut is the mechanism that creates the capacity for the month-2 concentration. Without it, you're running the next play on top of a full plate.

How do I know if the month-3 systemization actually worked?

Take an unplanned week off, with no check-ins, and watch what happens to the one thing you concentrated on in month 2. If the revenue, pipeline, or outcome it was producing keeps producing at the same rate without you, the system works. If not, spend the next month hardening the playbook and dashboards until it does.