From $500k to $1M a month, you need a second channel and a number you can trust.

What happens to your business if Meta’s costs rise by a third next quarter?

For a brand doing $700,000 a month with nearly all of its paid acquisition on one platform, the honest answer is that the whole growth plan moves with it. One auction, one algorithm, one policy team, one account that can be restricted on a Tuesday afternoon. Below $500,000 a month, that concentration is usually the right trade. Above it, the risk starts to outgrow the convenience.

So the second channel is the obvious next move. The trouble is that it is also the moment most brands lose track of what their marketing is doing, because from the day it launches, two platforms will both report that they caused the same sales.

The short version: add demand, and judge it on the store

The rule for this rung has two halves. Add a channel that reaches demand Meta is not reaching. Then judge it on what changed in Shopify revenue and new customers, never on what the new channel reports about itself. Blended MER, total store revenue divided by total ad spend, is the anchor. Everything else is a witness.

Why Meta alone gets more expensive as spend grows

Dependency is the visible risk. The quieter one is cost. As the arithmetic of doubling revenue shows, each extra customer on one platform costs more than the last, because more budget reaches people further from your best buyers. A second channel has its own curve, and the first customers on a new curve are cheaper again.

That is the real case for the second channel at this band. It is not diversification for its own sake. It is buying the next thousand customers somewhere they have not yet become expensive.

Which second channel: three options for a Shopify brand

For most Shopify brands, the second channel is Google, but Google is really three different jobs, and they should be judged separately.

• →Shopping and Performance Max: capturing people who already search for what you sell. Google describes Performance Max as a goal-based campaign type that reaches all Google Ads inventory, including YouTube, Display, Search, and Gmail, from a single campaign. For a catalog brand with searched products, this is usually the first place to go.
• →Brand search defense: showing up when people search your name. It is cheap and often sensible, because competitors can bid on your name. It is also mostly demand you already created, so it must never be pooled with non-brand when you judge the channel. The reasoning is in brand vs non-brand, the measurement error.
• →YouTube and Demand Gen: creating demand among people who are not searching yet. Google describes Demand Gen campaigns as serving ads on YouTube, Discover, Gmail, and the Display Network. This is the right second channel when your category is not searched for, and the hardest one to measure.

A simple way to choose: look at how people find you today. If a real share of your Shopify orders already arrives from organic Google search, demand exists in search and Shopping is the natural second channel. If almost nobody searches for your category, a search channel has little to capture, and the second channel has to create demand, which points to YouTube and Demand Gen. Brand defense sits alongside either one. It is never the whole plan.

The readiness test: five conditions before you launch

A second channel launched too early splits a small team and muddies the numbers the brand relies on. I would want all five of these to be true first.

1. Meta is supplied with creative. If you are not shipping enough distinct concepts, the first channel has cheaper growth left in it. That is the previous rung, creative supply from $250k to $500k a month.
2. Meta’s marginal cost is rising. A spend step test shows the next customers on Meta costing noticeably more than your average, and close to what you can afford.
3. Tracking is clean. Purchases reach both platforms reliably and are not double-counted within each one. For the Shopify side of Meta, see Shopify and the Meta Conversions API.
4. You have margin headroom. Blended MER sits well above breakeven MER, which is one divided by your contribution margin rate. At 45% contribution margin, breakeven is 2.22.
5. You can fund 60 days of learning without cutting Meta to pay for it. If you are reading this in October, either launch now and judge it against the same weeks last year, or wait and launch in January, because a 60-day window that starts now runs straight through Black Friday and December. Cutting the first channel to fund the second makes it impossible to tell what either did.

Why two platforms will claim more than you sold

Each platform counts conversions under its own rules. Google’s click-through conversion window defaults to 30 days for a new conversion action, so a purchase up to a month after an ad click can be counted. Meta runs its own attribution system with its own settings. Neither platform can see the other’s touchpoints, and each is built to count the sale when its ad played a part.

Shopify has its own view too. Its marketing reports support several attribution models, and the default, last non-direct click, gives full credit to the last non-direct channel before purchase. The full list of reasons these ledgers disagree is in why Meta, Google, GA4, and Shopify report different revenue. For this rung, one consequence matters: once you run two channels, the platform numbers will add up to more than the store took in.

One month, three ledgers: $740,000 claimed, $700,000 sold

Here is an illustrative month for a brand at $700,000. This is arithmetic, not a client result. It spends $150,000 on Meta and has just added $50,000 on Google. Average order value is $100.

The two platforms claim $740,000 of revenue in a month when the store recorded $700,000 from every source combined, including organic search, email, and people who typed the address. Neither platform is lying. They are counting overlapping journeys under different rules. But it means the platform ROAS figures cannot be added, averaged, or compared directly. Only the MER column is a fact about the business.

Before and after: what the second channel actually added

The better question is what changed when Google was switched on. Compare the same brand’s month before the launch with the month after, holding Meta spend constant. This is still illustrative arithmetic, and a single month against a single month is a noisy comparison, which we will come back to.

Google reports $260,000 of revenue at a 5.2 ROAS. The store grew by $80,000. On the store’s evidence, each Google dollar added $1.60 of revenue, below the 2.22 breakeven for a 45% margin, and the month’s contribution fell by $14,000.

That does not mean Google failed. A new channel spends its first weeks learning, part of the $260,000 claim is brand search that would have converted anyway, and one month is not a verdict. It does mean the dashboard figure is not the decision figure. If you had judged the channel on its 5.2 ROAS, you would have scaled it. The store suggests you should first find out which part of it is adding revenue.

How noisy is one month against another? Noisy enough that I would never scale or cut a channel on it alone. A promotion, a shipping delay, or one strong email can move store revenue by more than a new channel adds in its first month. That is why the decision rule below compares against a baseline written in advance, using the same weeks last year where you have them, and why a gray result goes to a holdout test rather than to a debate.

The post-purchase survey as a third witness

Store revenue tells you whether the total moved. It cannot tell you why. A one-question survey on the order confirmation page, “how did you first hear about us?”, adds a perspective no platform has: the customer’s own memory.

Memory is imperfect, and people over-report the last thing they saw. Treat the survey as direction, not precision. Its value is in the comparison. If Google carries a quarter of your spend but almost no new customers name Google or YouTube as where they first heard of you, the channel is probably capturing demand Meta created. That is useful work, but it is not the independent demand a second channel is supposed to add.

The decision rule for the first 60 days

Write the scorecard before the channel launches, so the result cannot be argued into a win afterward.

1. Project a baseline: what Shopify revenue and new customers would have been over the next 60 days without the channel, from the recent trend and the same weeks last year.
2. Set the bar in new customers: channel spend over the period divided by your marginal CAC ceiling. At $100,000 of spend and a $60 ceiling, the channel needs to add about 1,667 new customers above baseline.
3. Judge brand search and non-brand separately. Brand defense is kept or cut on its own logic. It never rescues or sinks the rest of the channel.
4. Read the survey and the platform numbers as supporting evidence. They can explain a result. They cannot overrule the store.
5. At day 60: above the bar, scale in steps. In the gray zone, run a proper holdout test before scaling, as described in incrementality testing without a data team. Clearly below, cut back to brand defense and your highest-intent Shopping and rethink.

The ceiling in step two comes from the same margin arithmetic as the rest of this ladder. Your MER floor sits underneath it. The breakeven ROAS calculator works out that floor from your margin, and MER vs ROAS explains why MER against store revenue is the number to set it on.

What common advice gets wrong: the second channel as a hedge

The usual advice treats the second channel as insurance: get off Meta dependency by spreading the budget. It sounds prudent. As stated, it is wrong, and the reason matters.

A second channel that mostly harvests Meta’s demand does not reduce your dependency. It hides it. Brand search and a good share of Shopping capture people Meta made curious. If Meta’s costs rise and you pull back, those channels shrink with it, because the demand they were catching stops being created. You thought you had two engines. You had one engine and a second dashboard.

Real diversification means a channel that brings in people Meta never reached. That is why the survey question is about where customers first heard of you, and why the 60-day bar is set in new customers rather than in reported revenue.

When this advice is wrong

• →You are a search-first category. If Google was your first channel, the logic runs in reverse. Meta becomes the second channel, and the question is whether it creates demand Google cannot.
• →Your tracking is not fixed yet. A second channel launched on a broken measurement base produces two sets of numbers you cannot trust. Repair first, then launch.
• →Much of your revenue sells off-store. If a marketplace or retail partners take a meaningful share, Shopify understates what ads produce, and MER against the store alone reads too harshly.
• →Your season is lumpy. If demand swings sharply month to month, before-and-after comparisons need last year’s same weeks, or a geographic holdout, to mean anything.
• →The first channel still has cheap growth. If Meta’s marginal cost is flat, the second channel is a distraction. Go back to creative supply.

What to do this month

1. Reconcile the last three months: platform-claimed revenue for each channel next to Shopify revenue. Note how far the sum exceeds the store.
2. Compute blended MER and breakeven MER. If the gap between them is thin, margin is your constraint, not channels.
3. Run the five-point readiness test honestly. Any condition that fails is this month’s work instead of the launch.
4. Add a one-question post-purchase survey now, so you have a baseline before the second channel starts.
5. Write the 60-day scorecard, including the new-customer bar, and share it with whoever will run the channel.

This is the second rung of the ecommerce growth ladder. If Google is your likely second channel, our Google Ads for Shopify team builds it with brand and non-brand separated from day one and judges it against the store. And if you want an outside read of where your current account is spending without return, start with the free audit.