How to scale an ecommerce business: the revenue ladder.

What would have to be true for your store to be twice its size a year from now?

Most founders answer with a tactic. More spend, another channel, a new product, a better agency. Each of those is sometimes right. None of them is an answer, because the question is not what you could do. It is what is stopping you.

A growing store is almost never blocked by five things at once. At any moment one constraint binds, and effort spent anywhere else produces very little. Which constraint binds changes as the business grows, and it changes in a fairly predictable order. A brand at $300k a month and a brand at $3M a month can run the same playbook and get opposite results, because they are standing on different rungs of the same ladder.

This guide is that ladder: five revenue bands, the constraint that tends to bind on each, the arithmetic underneath all of them, and a way to check which rung you are actually on. Every number in it is either sourced or labeled as arithmetic. None of it is a client result.

The short version: one constraint per rung

Read the table as a map, not a verdict. The band tells you where to start looking. The symptom tells you whether you have found it.

Each rung also gets its own deep guide through October and November, starting with creative supply at $250k to $500k a month. This page is the map they all hang from. Underneath all five sits a floor: if you cannot say whether an extra order makes money, you are not on any rung yet, whatever your revenue. That comes first, and it is short.

The floor under the ladder: know whether an order makes money

Platform ROAS is computed on revenue. Your business runs on contribution margin: revenue minus product cost, shipping and fulfillment, payment fees, returns, and any sales tax or VAT sitting inside the revenue figure. Those deductions are not small, and the ad platforms see none of them.

A worked example, as arithmetic. An $80 order with $30 of product cost, $9 of shipping, $3 of fees, and $4 of expected returns leaves $34 of contribution margin, or 42.5%. That order breaks even at a ROAS of about 2.35, which is 80 divided by 34. A 2x ROAS on this product loses money on every sale while the dashboard calls it healthy.

The fix is unglamorous. Enter cost per item on every product in Shopify so the store knows its own margins, compute contribution per order for your real product mix, and derive the breakeven with the breakeven ROAS calculator. Then govern the account on MER against Shopify revenue rather than on platform-reported ROAS, because a blended ratio cannot be inflated by attribution. MER vs ROAS covers why, and ecommerce unit economics walks the full waterfall.

This sounds like accounting. It is often the growth move that pays for itself fastest, because once margin is visible by product, the spend to cut and the products with unused headroom tend to become obvious. Reallocating from the first group to the second grows revenue and profit at once, with no new channel and no new budget.

One more check belongs on the floor. When every channel underperforms at the same time, the common factor is rarely three separate media problems. It is the destination. Conversion rate multiplies everything upstream: moving from 1.5% to 2.2% means about 47% more orders from the same traffic, which cuts the cost of each order by roughly a third without touching a bid. Check mobile speed on the pages that receive paid traffic, the checkout step by step, and whether the page a visitor lands on answers the promise the ad made. Landing pages covers the detail.

The arithmetic of doubling

Before the rungs, the one equation every one of them is working on. Monthly revenue has three terms.

That is all the growth there is. You can acquire more new customers, make each first order larger, or earn more from the people who already bought. Each rung of the ladder is, in the end, the point where one of those terms stops moving and the others cannot carry the load.

Here is the arithmetic for an illustrative brand at $500k a month that wants to reach $1M. Today it wins 5,000 new customers a month at a $70 first order, and returning customers add $150,000. The table assumes that the cost of a new customer rises as you buy more of them in a month, because each extra customer is reached further from the people already looking for you. The rate of that rise is an assumption chosen to show the shape of the problem, not a benchmark. Put your own numbers in the ecommerce growth milestone planner.

Read path A first, because it is the plan most founders default to. Doubling on new customers alone means buying 2.4 times as many of them each month. If that pushes the average cost of each one from $25 to $36, the spend needed rises 3.5 times and advertising takes 44% of revenue. On the 42.5% margin from the floor example, that is a store twice the size that loses money every month.

Path D reaches the same revenue with 1.5 times today’s new customers, and spends 21% of revenue doing it. Nothing in D is cleverer media buying. It is a larger first order and a customer base that comes back. That is why this is not a media ladder. Two of its five rungs are about the other two terms.

Two caveats keep the table honest. It is a single month, and in real life returning revenue grows on its own as each month’s new customers join the base, which helps every path. And the rising cost per new customer is an assumption: yours may rise faster or slower, and the only way to know is to step spend up and watch what the store reports. The shape holds either way. The term you can move cheapest is rarely the one you are currently pushing. To see how long each new customer takes to pay back, use the CAC payback calculator.

How to tell which rung you are really on

Revenue band is where you start looking, not the answer. A $3M brand can be stuck on creative supply, and a $400k brand can already have a margin problem. The rule is simple: you are on the lowest rung whose test you fail. Constraints lower down do not go away because the business got bigger. They just get more expensive.

The order matters because solving a later constraint while an earlier one holds produces work that looks productive and changes nothing. A forecast built on creative that is already tired. A margin project in a store that cannot tell a new customer from a returning one. The ladder is a sequence, not a menu.

Rung one: $250k to $500k a month, creative supply

The symptom is a heartbeat. One or two ads carry most of the Meta spend, and the month rises and falls with them. When they tire, cost per new customer climbs, the team launches variations of the same idea, and none of them beats the incumbent. More budget mostly buys more frequency on the same people.

It binds here because most brands reach this band on one channel and a handful of winning concepts. On Meta in particular, the creative now does much of the targeting: the ad decides who stops, and the system learns who to show it to from the people who stop. A small library of ideas reaches a small set of people, however much you spend behind it. The constraint is not budget. It is how many genuinely different ideas you can put into market each month, which is the argument in creative volume is the new targeting.

The first three moves:

1. Set a production cadence you can actually sustain, a fixed number of new concepts every week, and count only concepts that differ in angle, customer, or format. Ten crops of one video are one concept.
2. Feed the cadence from customers rather than from the team’s taste. Reviews, support tickets, return reasons, and a one-question post-purchase survey are where the next hooks come from.
3. Agree promote and kill rules before launch, and name ads so results read by concept. The system is laid out in the creative testing system that scales and scaling Meta creative production.

Ignore for now: a third ad platform, attribution modeling, and conversion rate test programs. At this volume most site A/B tests will not reach a result in a useful time, so fix obvious breakage instead of running experiments.

The number that proves progress is the share of spend going to concepts launched in the last 60 days, read next to cost per new customer. When new work regularly earns budget and new-customer cost holds as spend rises, this rung has stopped binding.

Rung two: $500k to $1M a month, a second channel and measurement you can trust

The symptom is two arguments at once. The first channel gets more expensive at every step up in budget. And when you add up what each platform says it drove last month, the total is larger than what Shopify booked, so every allocation meeting becomes a debate about whose number to believe.

It binds because the two problems arrive together. One channel has a ceiling, so growth needs a second, usually Google search and Shopping catching demand your Meta work created. But the moment there are two channels, both claim the same sale, and without a number anchored to the store you cannot tell whether the second channel is adding customers or relabeling them. Growth needs the channel. Allocation needs the measurement. Why Meta, Google, GA4, and Shopify report different revenue walks through the gap.

The first three moves:

1. Build one scoreboard anchored on the store: weekly MER from Shopify net sales and total ad spend, and cost per new customer from Shopify’s new customers count. Report platform numbers beside it, never instead of it.
2. Add the second channel with a written test plan and a fixed window, starting with non-brand search and a clean Shopping feed. Keep brand search separate everywhere. People typing your name were already coming, so brand search is harvesting demand, not capturing it, and brand vs non-brand explains what mixing the two does to your numbers.
3. Fix signal quality before you scale on it. Server-side conversion tracking keeps purchases visible to the platforms when browsers drop them, covered in server-side tracking and the Conversions API.

Ignore for now: a third and fourth channel, marketing mix modeling, and the search for a perfect multi-touch model. You need a number the whole team trusts more than you need a sophisticated one.

The number that proves progress is cost per new customer across all channels combined as total spend rises. If it holds while spend grows, the second channel is adding customers. If it climbs, the channels are trading the same ones.

Rung three: $1M to $2M a month, a new-customer engine

The symptom is growth that flatters. Revenue is up, blended ROAS looks the best it has ever looked, and the number of new customers each month has barely moved. Retargeting and brand campaigns take a rising share of spend and report the strongest returns.

It binds because at this size the existing customer base is large enough to carry the optics. Returning customers are the cheapest conversions in any account, and bidding systems chase cheap conversions unless told otherwise, so without a deliberate split the ads drift toward selling to people who already bought. The doubling table shows why that cannot last. Returning revenue helps, but the new-customer term is the one that compounds, because this month’s new customers are next year’s returning revenue.

The first three moves:

1. Separate new from returning revenue, by channel, using your own customer list rather than each platform’s guess, and make cost per new customer the headline number. What a new customer really costs makes the case.
2. Tell the platforms what you want. Google Ads customer acquisition settings can bid higher for new customers, or bid only for them, on Search, Shopping, Performance Max, and Demand Gen campaigns, and they work best when you define existing customers with your own first-party data, such as a customer list.
3. Set a ceiling on what a new customer may cost, from first-order contribution plus the payback window your cash can fund, and hold every channel to it.

Ignore for now: loyalty program polish and new channels. Retention matters, but at this rung its job is to be measured separately, not to be the growth story.

The number that proves progress is new customers per month, and what each one cost, trended over a quarter. Shopify’s customer cohort analysis groups customers by the month of their first order, which is the simplest honest view of whether the engine is running.

Rung four: $2M to $3M a month, margin, offers, and AOV

The symptom is a store that grows while the money does not. Revenue sets records, MER looks acceptable, and contribution after marketing is flat or falling. Discounts have crept deeper, a promotion runs most weeks, and nobody can say which products the ad spend is really paying for.

It binds because at this size a point of margin is a lot of money. As arithmetic, one percentage point of contribution on $2.5M a month is $25,000 a month, or $300,000 a year, before anyone touches a bid. It also binds because margin sets the ceiling on everything else. Breakeven MER equals your contribution margin, so a brand whose margin slides from 45% to 38% has quietly lowered the spend it can afford on every channel at once. And first-order AOV, the middle term of the doubling equation, is set by offers and bundles, not by media.

The first three moves:

1. Compute contribution margin by product group and give each group its own target, rather than one account-wide ROAS that overbuys low-margin items and starves high-margin ones. The mechanics are in profit on ad spend and the Google Shopping operator’s guide.
2. Rebuild offers around order value instead of sitewide percentage off: a free-shipping threshold set above today’s AOV, bundles, and gift-with-purchase. Read every offer against contribution per order, not revenue.
3. Move spend toward the high-margin products that are under-bought. At this band that is often the cheapest growth left, because it needs no new customers at all.

Ignore for now: revenue records set during promotions, and any new channel. A bigger promotion that lowers contribution can simply be paying customers to buy what they would have bought anyway.

The number that proves progress is contribution after marketing in dollars each month, with first-order AOV beside it. If both are rising, the rung is moving.

Rung five: $3M to $5M a month and beyond, forecasting, incrementality, and the team

The symptom is surprise. The plan says one thing and the month says another. Finance and marketing hold different forecasts. Every channel report looks fine and the total does not add up. And most decisions still wait on the founder, who has become the only place the numbers meet.

It binds because the cost of being wrong has grown. At this spend, the gap between the average return and the return on the last dollar is large money, and average numbers hide it. Inventory, cash, and media plans have to agree months ahead, especially around peak. And no single person can hold that much together by attention alone.

The first three moves:

1. Run one incrementality read per quarter on your largest line of spend. Google and Meta both offer conversion lift studies that compare people who could see your ads with a held-out group who could not, and geo holdouts do the same by region. The options are in incrementality testing without a data team.
2. Forecast from your own spend curve. Record what each step up in spend bought in new customers and contribution, plan the next quarter from that curve with finance in the room, and consider a marketing mix model once you have enough history. Google’s Meridian and Meta’s Robyn are both open source.
3. Give every number an owner, and hold a weekly operating review where the store-anchored scoreboard is the agenda. Decide deliberately which work sits in-house and which with partners, a trade-off laid out in in-house vs agency.

Ignore for now: debates about which attribution model is right. At this rung, attribution tells you where to look and lift tests tell you what is true.

The number that proves progress is forecast error: last quarter’s plan against what happened, for new customers and contribution. When that gap narrows, you are running a business that can decide where the next $100k goes before it spends it.

What common scaling advice gets wrong

The most common advice is that scaling is a spending problem: find the channel, raise the budget, hold the ROAS. The arithmetic says otherwise. In the doubling table, the path that relies only on buying more new customers is the most expensive way to reach $1M a month, and on ordinary margins it loses money. Spend is the lever that feels most like growth, which is why it gets pulled first. It is often the one that should be pulled last.

The second error is advice without a band. A tactic that is right at $2M a month, such as margin-band bidding or a mix model, is noise at $300k. A tactic that is right at $300k, such as concentrating on one channel, is a risk at $2M. When advice reads as irrelevant, it is usually aimed at a rung you have not reached or have already left.

The third is treating retention as a separate department. Returning revenue is a term in the same equation as acquisition. A brand whose customers come back can afford to pay more for each new one than a competitor bidding on first orders alone, which is how some brands outbid rivals for the same customer without losing money. The discipline is to license that deliberately rather than hopefully. "Our LTV is high, so we can afford a bad CAC" is the sentence that precedes a cash crisis when the payback window is longer than the business can fund.

When this ladder is wrong for your business

The bands describe a store that sells physical products direct to consumers, mostly through its own site, with paid social and search as the main acquisition channels. Outside that shape the order changes.

• →Subscription brands. Returning revenue is the business, so retention and payback can bind before creative supply does.
• →High-ticket, low-frequency products. If customers buy once every few years, the returning-revenue term is small and the new-customer engine binds from the start.
• →Brands with a large marketplace or wholesale channel. The store is only part of revenue, so store-anchored MER understates what advertising drives, and the measurement rung arrives earlier.
• →Heavily seasonal or gifting brands. When one quarter carries most of the year, monthly bands mislead. Read the ladder on trailing twelve-month revenue divided by twelve instead.
• →Cash or inventory constraints. If you cannot fund the stock to sell twice as much, no rung is the constraint. The balance sheet is.

The bands are also approximate. A brand at $950k a month with a working second channel and trusted measurement is already on rung three. Use the diagnostic, not the revenue line.

What to do this month

It is the end of September, so peak is weeks away. That changes the order of this month’s work, not the method.

1. Fill in cost per item for every product in Shopify and write down contribution per order. You will need the breakeven for every promotion decision in November.
2. Run the diagnostic table above and write your rung on one line. Share it with whoever owns the budget.
3. Put last month into the doubling equation: new customers, first-order AOV, and returning revenue. Decide which term you can move cheapest.
4. If you are on rung one, lock the creative cadence for November now, because concepts briefed in November arrive after the peak they were for. The BFCM 2026 paid media playbook covers the rest of the peak plan.
5. Set a stop rule before peak spend begins: the cost per new customer above which you stop adding budget. Decided in advance, it holds. Decided on Black Friday, it rarely does.

If you want the diagnostic run on your own data, the free audit reads your Google Ads or Meta account and works from what is there rather than from benchmarks. If you are past $50k a month in total paid media and want a team to work the ladder with you, that is what our Shopify marketing agency service is built around, and the ecommerce overview shows how we approach it. For what changes once ad spend itself passes six figures a month, scaling ecommerce ads past $100k picks up the media side.

The method is the sequence, not the list. Find the constraint that binds now, fix that one thing, and let the next one announce itself. The stores that scale are rarely the ones doing the most. They are the ones doing the right thing in the right order.