The Meta Coefficient: A Smarter Way to Budget Every New Ad Channel

Stop budgeting new channels as a percentage of your total spend. That number is lying to you, and it's quietly capping your growth.

I know that sounds like a strange thing to say. "Share of budget" is how most brands I talk to think about their media mix. Meta gets 70%, Google gets 15%, the new shiny channel gets whatever's left over. It feels tidy. It feels controlled.

In reality it's the wrong unit of measurement, and once you switch units the whole decision gets clearer.

Here's the idea I want to walk you through. Instead of asking "what share of my budget should this channel get", ask "how many cents am I spending here for every dollar I spend on Meta". I've started calling it the Meta Coefficient. It's the single most useful reframe I've picked up on media mix in a long while, and it came straight from watching how some of the sharpest operators actually plan their spend.

Let me show you why the percentage view breaks, then build the worksheet with you.

Why share-of-budget quietly misleads you

Picture a brand spending ~$80k a month on Meta and ~$16k a month on a newer channel like AppLovin. On a share-of-budget chart, that newer channel might show up as a tiny sliver, maybe 4% of total spend, especially if you're benchmarking against an aggregate report that blends in every other brand.

But that 4% hides the thing that actually matters. A big chunk of those other brands can't even access the channel yet. It's gated, invite-only, still rolling out. So the aggregate "4% of budget" isn't telling you whether you're under-invested. It's telling you the channel is hard to get into. Those are completely different facts.

Here's the thing. The relationship that matters isn't the channel against your whole budget. It's the channel against Meta specifically. Because Meta is the engine almost every DTC brand still runs on. For most accounts I see, Meta is still around 60% of total spend. It's the anchor. Everything else should be priced relative to it.

So the question stops being "is AppLovin 4% or 8% of my budget". It becomes "for every dollar on Meta, am I spending 10 cents here, or 20, or 30? And is that the right number for the return I'm getting?"

That's the coefficient. And it changes how you scale.

Step 1: Anchor on Meta and set your true coefficients

Open whatever reporting you trust, your platform numbers or an attribution tool, and pull last month's spend by channel. Now divide each channel's spend by your Meta spend. Don't divide by total. Divide by Meta.

On that ~$80k/month Meta example, here's what a healthy 6-7 figure brand might look like:

• Google: ~$12k. That's a 15-cent coefficient. 15 cents on Google for every Meta dollar.
• AppLovin: ~$16k. A 20-cent coefficient. Roughly a fifth of what you're putting into Meta.
• Everything else (TikTok, Snapchat, Pinterest, connected TV): ~$8k combined. A 10-cent coefficient across the long tail.

Write those down. That's your starting ratio map. Three numbers: 15, 20, 10. Suddenly the whole mix fits on a sticky note, and you can reason about it.

The reason this works better than percentages is that the coefficient holds its meaning as you scale. If you double Meta to ~$160k and you believe in the relationship, your Google line should drift toward ~$24k and AppLovin toward ~$32k to hold those coefficients. Percentages would have you constantly re-slicing a moving pie. Coefficients give you a fixed gear ratio.

Step 2: Pressure-test each coefficient against incrementality

A ratio on its own is just a habit. The number only earns the right to exist if the channel is actually driving incremental orders, not quietly taking credit for sales you'd have got anyway.

This is where I'd slow down. Before you trust any coefficient, the channel needs to have passed a holdout or geo-lift test at least once. Not a dashboard ROAS. An actual test where you withhold the channel from a region or audience and measure whether first-time orders drop.

The operators I respect most run this test repeatedly on any new channel before they let it grow. They'll run three separate holdouts and only ramp once they've seen incremental first-time orders show up every time. That discipline is the whole game. A channel can post a flattering 1-day-click ROAS and still be almost entirely re-attributing existing demand.

So for each coefficient on your map, give it a status:

• Proven incremental. It's passed a holdout. You can scale this coefficient with confidence.
• Unproven but promising. Decent dashboard numbers, no holdout yet. Keep the coefficient small, run the test before you raise it.
• Probably not incremental. Branded search you'd have won anyway, or a channel that quietly fed off Meta's demand. Be honest. This coefficient should come down, not up.

That last one matters. A lot of brands are overspending on Google branded search and calling it performance, when in reality those buyers were coming from a Meta ad they saw two days earlier. If that's you, your Google coefficient is inflated and your Meta coefficient is understated. The fix isn't to scale Google. It's to recognise the dependency.

Step 3: Watch the relationship, not the channel in isolation

This is the part that protected blended ROAS for the brands that got it right, and it's the least intuitive bit.

Meta and your supporting channels aren't independent. When you scale Meta, it tends to lift the channels downstream of it. More Meta impressions means more branded searches, more retargeting pools, more warm audiences for the newer platforms to convert. So if you scale Meta hard and freeze everything else, you'll often watch your Google and AppLovin efficiency improve on paper, because they're feeding off demand Meta created. Cut Meta, and those same channels suddenly look worse for reasons that have nothing to do with the channels themselves.

That's why I budget the mix as a system, not as separate line items. The coefficient view bakes this in. If Meta moves, the supporting lines move with it, in ratio. You scale together.

Here's what that protects you from. The classic blended-ROAS blow-up is when a brand scales one secondary channel in isolation, chasing its standalone ROAS, while the channel is really just harvesting Meta's overflow. The standalone number looks great right up until Meta plateaus, and then the whole house of cards re-rates at once. Holding coefficients keeps each channel honest relative to the engine that's actually feeding it.

Step 4: Decide which coefficient to raise next

Now you've got a map (the ratios), a status on each (the incrementality), and the rule that you move them in lockstep with Meta. The last question is where the next dollar goes.

My take: raise the coefficient on whichever proven-incremental channel still has room to grow before diminishing returns, and where the lift is fastest to capture. For a lot of brands right now that's been the newer scalable channels rather than the mature ones, simply because there's untapped reach there and the creative ports across with almost no extra work. You take your best vertical video and run it as-is.

A rough way to sequence it:

1. Hold or trim the inflated coefficients first. If branded Google is fat and non-incremental, that's the cheapest gain you'll find. You're not spending more, you're stopping a leak.
2. Raise the proven coefficient with the most headroom. If AppLovin passed its holdouts and you're at 20 cents with the efficiency still holding, test 25, then 30. Watch where diminishing returns kick in.
3. Keep unproven coefficients small until they earn it. Run the holdout. If it passes, it graduates and joins the queue. If it doesn't, it stays a rounding error.
4. Re-derive the whole map monthly. Spend shifts, channels mature, gating opens up. The coefficients drift. Recompute them off your current Meta anchor and repeat.

The beauty of this is it turns a vague, anxious question ("are we spending the right amount on each channel?") into four small numbers you can defend and adjust. You stop arguing about percentages of a pie that's always changing shape, and you start managing a set of fixed relationships against a single anchor you understand.

So if you take one thing from this, swap the unit. Next time someone asks what share of budget a channel should get, answer in cents per Meta dollar instead. Then ask the only two questions that matter: is this coefficient backed by a real incrementality test, and is it moving in lockstep with the engine that feeds it?

What would your three coefficients be if you worked them out for last month? I think the answer surprises most founders the first time they run the maths.