Stop Funding the Cohort Hole: When Pouring Money Into Meta Is the Wrong Move

Picture a bucket with a slow leak in the bottom. You can keep the water level up by running the tap harder, and from across the room it looks full. Turn the tap off for a day and you realise how much was quietly draining out the whole time.

That bucket is most ecommerce brands that ask me to scale their Meta spend. The tap is paid acquisition. The leak is churn. And turning the tap up doesn't fix the leak, it just hides it and runs your water bill up.

This is the part of my job that surprises founders the most. They come to us wanting more spend, and now and then my honest answer is "not yet". We turn down spend increases when the unit economics don't support them, because pouring money into Meta on top of a leaky bucket isn't growth. It's just a more expensive way to stand still.

Here's how I decide.

Why more spend is sometimes the wrong move

The instinct when revenue plateaus is to buy more of it. Spend's been flat at, say, ~$40k a month, so push it to ~$55k and the top line should follow.

Sometimes it does. But if the reason you plateaued is that customers aren't coming back, all that extra spend is doing is acquiring people to replace the ones quietly walking out the back door. You're not growing the base. You're refilling it.

I think about a business the way you'd think about a bath. New customer revenue is the tap. Returning customer revenue is the water already in there. If the plug's out, you can run the tap full bore and the level barely moves, because the more you put in, the more drains away. The fix is never a bigger tap. It's the plug.

And paid ads are an expensive tap. Every new customer costs you upfront, and if a big chunk of them never buy again, you're paying full acquisition price for a single order, forever. That maths only works if the second and third orders show up. When they don't, scaling spend just scales the loss.

The signals you're funding a cohort hole

A cohort hole is what happens when a run of weak acquisition leaves you with a customer base that isn't compounding. Each new cohort lands, fails to repeat at a healthy rate, and fades out, so next year's returning revenue is smaller than this year's even though you "grew". You spend more to stand in the same spot.

These are the signals I look for before I'll agree to scale. If a few of them are flashing, the budget conversation stops and the retention conversation starts.

• Your active customer file is shrinking, not growing. Forget your total email list, that number only ever goes up and it's lying to you. Count the customers who've actually bought inside the window where repeat purchases normally happen. If more are churning out of that window than you're adding, your base is contracting. That's the canary in the coal mine, and no amount of new spend fixes a base that's leaking.
• Returning revenue is flat while new-customer revenue climbs. Returning revenue is a lagging signal of how well you acquired six and twelve months ago. If you're spending more to bring people in and the repeat line isn't moving, the people you're buying aren't sticking. You've got a hole, and you're shovelling money into it.
• Your first order barely breaks even and there's no second. Plenty of brands run acquisition at break-even on the first purchase, banking on profit from orders two and three. Completely fine, if those orders exist. If your repeat rate is thin, you've built a business that pays to acquire and never gets paid back.
• Contribution margin is sliding as you spend more. When the next dollar of spend brings in revenue that doesn't cover the product, the shipping and the ad cost to get it, you've passed the point where scaling helps. You're buying volume at a loss and calling it growth.
• You're "diversified" but every channel shares one risk. A brand running Meta, Google and TikTok feels spread out. But if all three are cold-traffic acquisition exposed to the same rising CPMs, that's not diversification, it's the same bet three times. None of it protects you when acquisition gets dearer, and scaling into it just enlarges the exposure.

The hard truth about climbing out

Here's the bit nobody wants to hear. A cohort hole can take years to climb out of, not weeks.

Your profit expansion comes from the share of revenue that's returning customers, because those orders carry no acquisition cost. If you've been through a stretch of deflated new-customer acquisition, next year's returning revenue is already set to be lower, and there's no campaign that retroactively fixes a cohort that didn't repeat. The damage is in the past. You can only stop adding to it.

So if you're genuinely in the hole, the way out is usually the uncomfortable one. You may need to spend more to acquire, at slim margin or break-even, to rebuild the base, while also fixing why people didn't come back the first time. That drags your blended efficiency down for a while before it gets better. It's a two-year story, not a two-week one, and the brands that pretend otherwise just keep missing their own forecasts and wondering why.

I'd rather tell a founder that upfront than take their money to scale a hole and watch the numbers get worse on my watch.

What I'd fix before touching the budget

When the signals are flashing, here's the order I'd work in. None of it is a spend increase.

Find the one number you actually win on. Most brands are great at exactly one thing: gross margin, retention, or acquisition. Be honest about which is yours, because it's very rarely acquisition, and stop trying to scale on the leg that's weakest. If you win on margin, the answer might be a higher average order value, not more orders. If you win on retention, protect it before you pour more in the top.

Plug the retention leak first. Outcomes drive repeat purchases, repeat purchases drive margin, and margin is what gives you room to spend on acquisition at all. Get the post-purchase experience, the flows and the product itself good enough that the second order shows up. A brand holding onto customers for one more cycle is worth more than the same brand acquiring 20% harder.

Earn the margin to spend. I keep coming back to a homewares brand I think about, sitting at a ~$55 average order value. Lift that toward ~$70 with a sharper cross-sell and bundle and you've created roughly ~$15 of headroom per order with no new spend at all. That headroom is what lets you absorb a higher cost of acquisition and scale safely. Margin first, then the tap.

Only then, open the tap. Once the base is compounding and the margin's there to support a higher cost per customer, scaling spend does what it's supposed to. You're adding water to a bucket that holds it. That's when more Meta budget is the right move, and not a day before.

Where to from here

The reason this is easy to get wrong is that a leaky bucket and a healthy one look identical from across the room. Both show a full level while the tap's running. You only see the difference when you look at whether the base is actually compounding underneath, and that's exactly the bit most founders never check before they ask to scale.

If you're staring at a plateau and you're not sure whether more spend is the answer or the trap, that's the question a Signal/Noise Audit is built to settle. We'd pull apart your unit economics and your cohorts and tell you plainly whether your bucket holds water, or whether the budget you want to add is just headed for the leak. Better to know which one you're dealing with before you turn the tap up.