Chapter 40: Would I Start This Today?

In month eleven Maya spent a Tuesday building a webinar funnel for a service line that should have been dead.

The line was group coaching. She had launched it in the spring, a monthly cohort where six founders fixed their own messaging in a room with her, priced at $600 a head. It had felt like the obvious next leverage play: package the thing she already did one-on-one, sell it to many, watch the margin open up. She had built a curriculum, a Slack space, a slide deck she was genuinely proud of. And it had limped. Three cohorts in, retention was forty percent, fills were soft, and she was spending more hours rescuing it than she spent on the one-to-one work that paid triple.

So that Tuesday she did what she always did when something wasn't working. She tried to fix it from the angle that required no honesty. She built a better funnel. A webinar to feed the cohort, a new landing page, a sharper hook on the registration form. She told herself the offer was fine and the problem was traffic, because traffic was a marketing problem and she was a marketer, and a marketing problem felt solvable in a way the other thing did not. The funnel got prettier. The cohort stayed broken. She poured another afternoon and ninety dollars of ad credit into the top of a bucket that was leaking out the bottom, and she knew it, and she kept pouring.

The substitute had a name this time, and she could see it. Sunk cost. She had built that curriculum. She was proud of those slides. Killing the line meant admitting that the proudest-looking thing she had made all year was the wrong thing, and she would rather optimize a funnel forever than say that out loud.

She closed the ad manager and opened the scoreboard instead.

The council, when she finally consulted them, did not argue about her feelings. They argued about her numbers, and the first voice cut straight through the slides she loved. Ignore what you spent. Ask the only question that matters:

If I had nothing invested in this thing, would I still start this today? Would I still pursue this path today? If the answer is no, it's time to pivot or walk away.

Maya sat with it honestly. If a stranger handed her the messaging business today, clean, no curriculum built, no slides made, would she choose to launch a $600 group program with forty-percent retention against a one-to-one offer that closed at eight times the margin? No. Not for a second. The only reason the cohort still existed was that she had already built it. That was the whole case for keeping it, and it was not a case at all.

Then the council did the thing that saved her from over-correcting. Because the other failure mode was just as real. She could overreact, torch everything, and go chase the shiny webinar-funnel-mastermind future that had nothing to do with what she was good at. The TEDx voice that runs through this whole question kept pulling her back to the difference between the mission and the method:

Pivoting, my friends, very simply, is maintaining your original strategy while changing courses in your tactics.

Her mission was not "run group cohorts." Her mission was the one line she had written on a clean page eleven months earlier, on day thirteen, about helping founders fix their messaging so people actually buy. The cohort was a tactic. A tactic could die without the mission dying with it. She had been treating her own spring decision like a vow when it was only ever a hypothesis, and the hypothesis had returns now, and the returns said no. The council went further and made the kill feel ordinary, almost expected, instead of like a confession of failure:

Research shows that 93% of all entrepreneurial successful companies have pivoted at least one time. They've been willing to do something different.

Ninety-three percent. The pivot was not the anomaly. The cohort surviving on pride would have been the anomaly.

She hit the fork the council always hits here, and for once all three positions were defensible.

The first camp said pivot, decisively, the moment the evidence is in. Survivors of failed models share one trait: they let the numbers, not the ego, make the call. The second camp said the opposite, and it was the camp she had been hiding inside all month. Don't pivot. Push harder. People quit exactly where the work turns boring or one milestone slips, and they mistake a hard stretch for a broken model:

Far too many people think they failed when actually all they did was quit before they figured it out. Don't take feedback for failure.

That voice was right about plenty of businesses. The danger was that it sounded exactly like sunk cost wearing a motivational jacket. So Maya needed the test that told the two apart, and Ali Abdaal handed it to her:

Whatever problem you're struggling with, ask yourself, is it a drama problem fundamentally or is it a maths problem? A drama problem is pretty much everything else.

She ran her cohort through it. The blocker was not drama. She had given it real reps, three full cohorts, a documented curriculum, honest feedback from leavers. The exit surveys said the same thing every time: founders wanted her on their specific business, not a room. The retention number was not a mood. It was maths. The model was wrong, and no amount of consistency fixes a wrong model. This was the one case where "push harder" was the avoidance, not the answer.

The third camp, the protect-what-works voice, told her not to touch a thing that was producing:

All you have to do is look at your balance sheet, your progress, the team that you have, and the amount of joy you have in your life. Now, if everything is going great, you don't need to change.

That one decided itself by pointing at her scoreboard. The cohort produced none of the three. Money, thin. Joy, gone, she dreaded those calls. Growth, flat. The protect-what-works rule did apply, just not to the cohort. It applied to the one-to-one line she had been starving to keep a dying product alive. So the fork resolved on the one variable that cut through all three camps: tracked evidence over time, not a bad week, not a brave week. The numbers had said no for three cohorts running. She believed them now.

Here is the line that does not appear anywhere in this chapter, because it could not. Across town, Marcus had finally launched his course in month eight and was drowning in it, doing every task himself, and when Maya asked him later how the cohorts were converting he could not tell her. He did not track it. He had feelings about it. He had a sense that it was "building." He had no scoreboard to consult, so he could not run the sunk-cost test, because the test needs a number to answer it and he had only pride. He would keep that course alive for another year on vibes alone. Maya killed hers in an afternoon because she had spent month seven building the five numbers that would let her do exactly this.

She paid the toll. She emailed the eleven people still in the active cohort, looped them in early instead of letting it vanish on them, refunded the two who had barely started, and migrated the four who wanted to continue into a higher one-to-one slot at a fair conversion price. She archived the Slack space. She did not delete the curriculum. She stripped it for parts, the three best modules became a lead magnet that fed the offer that actually worked. She kept the mission. She buried the tactic. It took less courage to let go than she had braced for, and that was its own quiet lesson: she had been guarding the door of a room she did not even want to be in.

By Friday the redeployed hours had gone back into the one-to-one pipeline, two new full-price projects closed that she would not have had time for, and the revenue line resumed its climb. The dead weight was gone. The business felt honest again, the way a desk feels after you throw out the thing you kept meaning to fix.

And then, on the Saturday, with the external finally sorted, the team running, the numbers clean, the dead line buried, Maya noticed something that no funnel and no scoreboard could fix. She had killed the cohort the moment the data turned. She still could not kill the small daily voice that told her any of it could vanish, that she was one bad month from the Slack message again, that she had to keep performing to stay safe. The market was solved. The buyers were solved. The team was solved. The last underperforming thing in the whole operation was the operator's own head, and there was no number on any dashboard that would tell her when to pivot that.

My verdict. The hardest thing to kill is the thing you are proud of, and pride is just sunk cost with better lighting. You already know which part of your work is underperforming. You have known for weeks. You are building it a prettier funnel right now so you never have to ask the one honest question, which is whether you would start this today if you were starting clean. If the answer is no, the time you already spent is not a reason to continue. It is the bill for the lesson, already paid, and pouring more in does not get it back. Keep the mission. Bury the tactic. The four-word version of this whole chapter: would you start today? Marcus could not ask it, because asking it requires a number, and he never built one. You did. So look at it.