Ecommerce Growth Patterns: Compound or Stall

Why some ecommerce brands compound year over year while others burn budget chasing revenue that never turns into profit.

Updated on: 2026-09-11

The same conversation happens over and over. A founder tells me revenue is up, the team is busy, the agency sends confident weekly reports, and yet the bank balance feels tight and every attempt to spend more makes efficiency worse. Nobody can say exactly which lever is broken. The dashboards look green. The business does not feel like it is winning.

After years of running an agency and advising brands mostly in the $2M to $20M range, the difference between the ones that compound and the ones that stall is rarely a traffic problem. It is whether acquisition, conversion, retention, margin, measurement and execution reinforce each other, or quietly cancel each other out. A compounding brand is one where each new customer improves the economics of the next customer. A stalling brand is one where each new customer costs slightly more and returns slightly less, and the growth chart hides it until it doesn't.

What "compounding" actually means in ecommerce

Compounding is not a motivational idea. It is arithmetic that punishes weak systems and rewards durable ones.

A brand growing 20% a year becomes about 2.49 times larger over five years. A brand growing 10% a year becomes about 1.61 times larger. That 10 percentage point gap in annual growth produces a business roughly 54% bigger after five years. Small, persistent improvements in retention, conversion or margin, reinvested rather than spent, are what create that gap. They rarely show up in a single monthly report.

This is why I push clients away from revenue as the north star. Revenue can rise while profit and cash fall. The market itself will not save you here. Global ecommerce is still growing, with EMARKETER's forecast summarized by Shopify putting worldwide sales at $6.419 trillion in 2025, up 6.8%. But category growth does not flow into your account automatically. FTI Consulting's 2025 U.S. report makes the harder point: pandemic ecommerce was mostly a pull-forward of trends already in motion, acquisition costs are rising, and brands now have to win share and frequency inside a maturing category. The easy-growth period is over.

Pattern one: compounders manage contribution profit, not ROAS

The single most common mistake I see is treating platform ROAS as if it were profit. It is not. It is a claim a platform makes about revenue after someone clicked an ad.

The number that decides whether you compound is contribution profit: what remains after the variable costs of selling and fulfilling an order. Depending on your accounting, that includes product cost, fulfilment, shipping, payment processing, returns and variable marketing. A brand that quietly excludes shipping, returns or discounts from its margin math will always look healthier than one that includes them, and will be surprised when the cash doesn't appear.

The 2025 DTC Mega Report, which analyzed over $10.1 billion in revenue across 107 million orders from Shopify DTC brands, shows why this matters. Median CAC was $54. Median first-order profit was $31. That leaves the median brand roughly $23 underwater on the first transaction, recovering only about 57% of acquisition cost from the first order. The report also found 37% of brands did not recover CAC on the first order at all, and 31% ran contribution margins below 10%.

Treat these as diagnostic ranges, not targets. The report does not publicly disclose its date range, brand count, geographic mix or weighting, so I would not hold any brand to a specific median. What the numbers establish is directional and important: strong first-order ROAS can coexist with an unprofitable acquisition model. The question worth asking in your weekly review is not "what's our ROAS." It is "after variable costs, how much contribution profit does a new customer generate, and how long until we recover the acquisition investment?"

Pattern two: retention is what turns acquisition into an asset

Acquisition without retention is a recurring expense. Acquisition with retention is an asset, because the second order changes the value of the first.

The same DTC dataset shows returning customers repurchased within 90 days at 33.0%, versus 14.7% for new customers. That's about 2.24 times higher. Repurchase likelihood also collapses with time since the last order: 89% within 0 to 30 days, 41% at 61 to 90 days, and 12% past 120 days.

I want to be careful here. Higher retention is not proof that email flows or loyalty tactics caused it. Returning customers self-select, and some products replenish naturally. But the economic asymmetry is real, and it is where operating leverage lives. On a $10M base, the report estimates a 10% improvement in retention is worth about $1.4M in impact, compared with roughly $1.0M for conversion, $1.0M for AOV, and $0.6M for CAC. Again, source-specific math, not a law of physics. Still, it matches what I see: the highest-leverage improvement is usually not the one everyone is fighting over in the ad account.

Retention works as a system, not a tactic. Klaviyo's case study of the eyewear brand L. Eyes Eyewear reported email driving 44% of revenue and repeat-customer revenue climbing from 16% to 50%. It is a vendor case with no independent control group, so I read it for mechanism rather than promised results. The mechanism is what matters: capture identifiable visitors, follow up while intent is still warm, trigger on behavior like browse and cart abandonment, tell the brand story, use post-purchase messaging, and give existing customers a real reason to come back.

Pattern three: product economics cap how much retention can compound

Retention is not equally available to every brand, and pretending otherwise leads to wasted effort.

A consumable with a predictable replenishment cycle creates repeat demand on its own. A mattress or an occasional luxury piece may have excellent customer value but a repurchase interval measured in years. Apparel can drive frequency through assortment, but returns and sizing erode margin. The DTC report's category contribution margins range from around 60.5% in fashion and apparel to 41.0% in beauty and cosmetics, with health, home and food in between. I would not treat those as universal standards, because the methodology isn't disclosed, but the spread makes the point: there is no single "good margin" or "good repeat rate" that applies to everyone.

So before chasing a higher repeat rate, I ask:

• Is the product naturally replenishable, or are you forcing frequency?
• Can you build bundles or subscriptions without destroying margin?
• Is the second purchase a genuine new order, a refill, or just a discount you trained the customer to wait for?
• Does higher frequency create more returns, support load and fulfilment cost?

Retention bought through permanent discounting will grow order counts and shrink contribution profit at the same time.

Pattern four: scale conversion and creative before scaling spend

Here is the pattern I get called about most: "ROAS drops every time we increase the Meta budget." The reflex is to blame the channel. The reflex is usually wrong.

More spend does one reliable thing. It increases your exposure to weaknesses that were already there. If your creative volume is thin, your conversion is soft, or your new-customer economics are marginal, more budget makes all three more visible, not better. Before touching the budget, I look at creative testing capacity, campaign structure, conversion, new-customer economics and measurement, in roughly that order.

Channel data supports treating platforms as different tools rather than one interchangeable "paid" line. Triple Whale's 2025 benchmark report, covering $18.4 billion in spend across more than 33,000 brands, found Meta made up 68.3% of ad spend, Amazon posted an 11.02% conversion rate with the lowest CPA, and TikTok ROAS varied by 234% between industries. AI-driven orders grew 1,481%, though the channel is still early. My read: Amazon's conversion rate likely reflects high purchase intent rather than demand creation, while Meta tends to create new demand but needs real incrementality testing to trust its reported conversions. Don't crown a channel good or bad without accounting for intent, category and creative fit.

Pattern five: measurement quality separates real scaling from attribution theater

Attributed revenue and incremental revenue are different things, and the gap between them is where a lot of budget goes to die.

Attributed revenue is what a platform claims credit for after an interaction. Incremental revenue is what would not have happened without the ad. Incremental ROAS divides the second by spend. A channel can show excellent attributed ROAS while capturing customers who were going to buy anyway.

The platforms know this and argue their own case. Meta's 2025 white paper, based on 307 studies across 54 advertisers, claimed rules-based attribution undervalues Meta by a median 31% versus incrementality measurement. That is commercially interested research, and I would not present it as neutral truth. Independent practitioners often argue the opposite, that platform reporting over-credits retargeting, branded search and view-through conversions. The resolution is not picking one platform's number. It is triangulating with experiments, blended CAC and contribution profit.

What I have teams track:

• New-customer CAC, not only blended CAC
• Blended CAC across paid, organic, referral and owned
• Contribution profit by cohort
• 30, 90 and 365-day repurchase rates
• Payback period in orders and days
• Net revenue after returns and refunds
• Incremental lift from controlled tests
• Cash conversion and inventory needs

This is the technical work that makes leadership confident in the numbers, and it is where first-party attribution and server-side tracking earn their keep.

Pattern six: returns and fulfilment are growth economics

Returns are not a back-office issue. They sit directly inside your unit economics, and most brands underweight them until they scale.

The National Retail Federation's 2025 returns research estimated $849.9 billion of U.S. merchandise returned in 2025, with 19.3% of online sales returned and 9% of all returns fraudulent. Meanwhile DHL's 2025 ecommerce trends research found 71% of U.S. consumers would abandon a purchase if the returns process failed expectations, and 73% would not buy from a retailer they didn't trust with delivery and returns.

That is a genuine trade-off, not a slogan. Frictionless returns lift conversion and confidence. They also cut net revenue, add processing cost and create fraud exposure. So the metric that matters is not gross conversion rate. It is net contribution per visitor: revenue after discounts, refunds and returns, minus variable fulfilment, payment, product and marketing costs. A brand that lifts conversion by attracting low-quality, high-return orders can stall with a better headline number.

Compounder vs staller

This table is a synthesis of the patterns above, not a measured taxonomy. Read it as a diagnostic, not a scorecard.

What I would do first

If growth has stalled and the numbers feel murky, I don't start with the ad account. I start with the economics and the ownership.

1. Rebuild contribution margin with every variable cost included, especially returns, shipping and discounts. Most surprises live here.
2. Split CAC and repurchase into new versus returning cohorts. You cannot fix retention you can't see.
3. Pick one commercial target and one owner. Fragmented ownership, where the agency optimizes ROAS and the email team optimizes opens and nobody optimizes contribution profit, is the most common cause of expensive stalling I encounter.
4. Establish a weekly review rhythm with clear priorities, decisions and escalation rules, so the plan actually moves.

This is the core of how I work with brands, whether as a three-month growth advisory or as a fractional head of growth. The point is not to replace a capable agency or team. It is to sit above execution, name the real constraint, and hold the whole plan to one commercial standard. You can see more of the background and track record here.

FAQ

Why does my ecommerce growth stall even though we keep spending?

Usually because spend is amplifying a constraint that lives somewhere other than the ad account. If contribution margin is thin, new-customer repurchase is weak, or measurement over-credits channels that capture existing demand, more budget makes the problem bigger and more visible. The fix is diagnosing which part of the system is limiting the others, not buying more traffic.

Is ROAS a bad metric?

It is not useless, but it is a poor primary target. ROAS is attributed revenue over spend, and attributed revenue is a platform's claim, not your profit. I'd rather anchor on new-customer contribution profit and payback period, then use ROAS as one input among several. A channel can post strong ROAS and weak incremental economics at the same time.

Should I fire my agency if growth has stalled?

Often no. The more common issue is unclear ownership and no shared commercial target, so the agency optimizes what it can see while margin and retention drift. Before switching anyone out, I'd clarify what good performance means, put one owner over the full plan, and set a review rhythm. Sometimes the same agency performs far better inside a system that holds it accountable.

How important is retention compared to acquisition?

For most brands in the $2M to $20M range, retention carries more leverage than founders expect, because the second order changes the economics of the first. But retention is capped by product category and repurchase cycle, so it is not a universal cure. The right move is matching your retention effort to what your product can actually support, without buying repeat orders through permanent discounting.

How do I know if a channel is really working?

Look past attributed revenue to incrementality. Run controlled tests, watch blended CAC, and check contribution profit by cohort. If pausing or scaling a channel barely moves total new-customer volume or profit, its reported numbers are probably crediting demand you already had.

Further reading

• 2025 DTC Mega Report for benchmark ranges on CAC, contribution margin and repurchase
• FTI Consulting 2025 U.S. Online Retail Report for market context and the shift toward omnichannel
• Insights from Miguel Casteleiro for more on operating rhythm, measurement and channel strategy