Training One Metric Does Not Win the Race
A 400% ROAS is a real number and it still leaves out most of what decides whether the business makes money. Product cost, inbound freight, payment fees, pick and pack, outbound shipping, returns and discounts all come out of the same sale the ad account is taking credit for.
Formula 1 teams hit the same trap in a different form. Braking takes enormous leg strength, which makes leg press numbers easy to track and easy to improve. A team that trains only that one thing loses on everything it never measured.
Ecommerce brands do this with ROAS. It’s fast, it’s visible inside the ad platform and it moves when you push on it, which is exactly what makes it a weak summary of the business.
Two questions are worth asking instead. What is one order worth after every cost has come out? Is that number where a business in your category should be?
This guide answers both, then covers the five habits that stop one metric from running the company.
Why Metrics Like ROAS Don’t Tell the Full Story of Profitability
This isn’t a problem that only hits weak operators. We see it most often in brands whose ad account is genuinely well run, because a well-run ad account produces a number worth celebrating.
We’re using ROAS (return on ad spend) as the example because it’s the shiny metric that distracts the most businesses and agencies. For many teams it’s the go-to number for judging ad performance.
Blended MER is the marketing efficiency ratio: total revenue divided by total ad spend. It catches what ROAS hides and it’s also the number that should set your ecommerce marketing budget in the first place.
Spend $1 on ads. Make $4 in revenue. That’s a 400% ROAS.
It reads like profit and it usually isn’t. ROAS doesn’t account for COGS (cost of goods sold), operating expenses, or anything that happens after the sale.
Three questions test it. What’s left after shipping, returns and fulfillment? Is retention holding as the customer base grows?
If you or your agency only watch ROAS, then the factors that decide long-term profitability go unwatched. The metrics worth watching instead are slower and less flattering. They’re the ecommerce KPIs that track fundamentals:
- Product quality
- Operational efficiency
- Profit margins
- Long-term customer value
A team that pushes hard on one performance metric is repeating the F1 team’s mistake: training the leg press and leaving the car alone. The KPI gets better and the result doesn’t.
A Strong ROAS Can Sit on Top of a Losing Order
The sale that produced your 400% ROAS is the same sale six other costs come out of. We usually get asked how to increase ecommerce profit before anyone can say what one order keeps. Walk one order down the page and you can see whether it made money.
- Product cost, inbound freight and duty: what you paid the factory plus what it cost to land the goods. Freight and duty belong here rather than in overhead, because they move with volume.
- Payment and platform fees: Shopify’s published online card rate runs 2.9% plus 30 cents on Basic and 2.5% plus 30 cents on Advanced. On a $60 order that’s roughly $2 gone before anything is picked.
- Pick, pack and outbound shipping: the labor, the box and the carrier. If you offer free shipping, then this line is a discount you already gave and never called one.
- The returns provision: NRF and Happy Returns estimate that 19.3% of online sales will be returned in 2025. So about one order in five carries a second shipping leg, a restock and often a unit you can’t sell at full price again.
- Discount give-back: our own 2023 discounting research found 62% of shoppers delay clothing purchases until they can buy at a discount. If most of your revenue arrives at 20% off, then your real selling price isn’t your list price.
- Acquisition cost: the ad spend the ROAS calculation already counted, now sitting underneath every line above it.
What survives all six is contribution margin. That is the revenue from an order minus every cost that moved because the order happened. It’s the number the ROAS conversation should have been about.
Ecommerce profitability comes down to four levers: what a product costs to make and land, what it costs to get to the customer and back, what it costs to acquire that customer, and what the price holds after discounts. Pricing work, demand forecasting and acquisition-cost reduction each move one of those levers and none of them moves the others.
Which lever matters most depends on where a brand’s margin is actually leaking. That is why the audit comes before the strategy. A brand that cuts acquisition cost while its returns rate quietly runs at twice its category’s norm has improved a number and not its profit.
Once you have the per-order figure, it belongs in your business budgeting: a budget built on revenue targets will keep funding the orders that lose money. A minority of your SKUs usually carries most of the contribution. The walk above is what tells you which ones.
Some of the profitability advice written for ecommerce sellers starts with a profit first allocation: set a profit share aside before any expense gets paid. That is a cash habit and the per-order walk is what tells you whether the orders can carry it.
If your team can’t say what one order is worth after all six lines, then an outside read on the numbers tends to be faster than another month of dashboards.
Your Category Sets the Profit Margin Range and Four Choices Decide Where You Land
We once built a full labor-allocation model with a client. His two highest-margin customer segments were losing money once we counted the service time those customers consumed. The lowest-margin segment turned out to be funding the whole business.
No category average could have seen that, which is why a benchmark is a place to start reading rather than a target to hit. It’s still worth having, because most teams argue about ecommerce profit margins without agreeing on which margin they mean.
| Margin | What it subtracts | The question it answers |
|---|---|---|
| Gross margin | Product cost, freight and duty | Is the product priced above what it costs to make and land? |
| Contribution margin | All of the above, plus fees, fulfillment, shipping, returns, discounts and acquisition cost | Does one more order make us money? |
| Net margin | All of the above, plus payroll, rent, software and the rest of fixed overhead | Did the whole company make money this year? |
Public-company data gives you the shape of the ranges a profitable ecommerce business works inside. Aswath Damodaran’s January 2026 industry margin dataset at NYU Stern puts apparel’s average gross margin near 57% and its net margin at 3.85%. Household products run about 51% gross and 11.7% net, while furniture and home furnishings run about 30% gross and 1.1% net.
So a strong gross margin says almost nothing about whether a category keeps money. Apparel and furniture both give nearly all of it back, in different places and for different reasons.
That spread also answers the question people usually ask first, which is whether ecommerce is actually profitable. It’s profitable the way retail is profitable, unevenly and by category. Most direct-to-consumer brands land below the middle of their category’s range because they carry an acquisition cost the wholesale model never had to.
Where a brand sits inside its own category’s range is mostly the result of four choices it already made. Average order value sets how far a fixed pick-and-pack cost gets spread. A free-shipping policy moves a real cost onto your side of the ledger.
The other two are the share of revenue going to paid acquisition, which decides how much of every order is bought rather than earned, and repeat rate, which decides how many orders one acquisition cost has to cover. A subscription brand and a one-time-purchase brand in the same category can sit at opposite ends of the range on that last factor alone.
Reading your own position takes a quarter of data and an afternoon. Pull your last full quarter, run the cost walk on your ten highest-volume SKUs and work out contribution as a share of revenue.
If you land below where your category sits, find the line that did it. Many brands find it in one or two places, often returns depth or discount depth. That gives them a repair with a name, which is more than a general push to be more profitable ever gives anyone.
The Risk of KPI Tunnel Vision
Push hard enough on a single KPI and the view of performance skews. It also pushes teams toward short-term thinking.
Under pressure to make one number look good, teams reach for tactics that lift short-term revenue. Whether those tactics help the business is a separate question. It’s rarely the one on the agenda.
The KPI quietly becomes the business goal.
Say ROAS is down. The first instinct is usually to meet the ad agency and push for improvement. To hit the target they might cut traffic volume to lift conversion, rework the ad messaging, or narrow the audience.
None of those moves is wrong on its own. ROAS may well improve for a quarter. Whether it was a win is a different question.
Three questions rarely get asked in that meeting:
- How will these changes affect the long-term goals? Cutting traffic volume to lift ROAS shrinks the top of the funnel, reduces brand awareness and shrinks the pool of future customers.
- How will these changes affect other key metrics, now and in a year?
- Is the ROAS problem a symptom of something deeper: friction in checkout, a weak post-purchase experience, or a product that doesn’t quite fit the market?
Keep goals as goals and KPIs as the tools you reach for. A temporary lift in one metric can undo a year of foundation work. The lift is the part that shows up in the meeting. Ecommerce growth hacking tactics fall into the same trap: a team reaches for whatever moves the number fastest without checking whether the tactic fits the business.
Improve Ecommerce Profitability By Shifting Focus from Short-Term Metrics to Long-Term Success
Five steps keep the cost work above from being overwritten by next month’s dashboard.
1 – Start With the Plan the KPIs Are Supposed to Serve
Why it’s needed: a strategic plan ties targets, actions, metrics and KPIs to the business goals. Without one, you end up picking metrics by which of them move. An ecommerce growth consultant can pressure-test that plan before your team commits to next quarter’s targets.
How to build it (a quick overview; our guide to ecommerce planning covers the work properly):
- Identify your three to five year vision, which is the part your ecommerce business plan should already state.
- Set long-term goals that support the vision.
- Break the long-term goals into short-term actions.
- Identify combinations of KPIs that track progress toward the goals.
The next step decides which of those KPIs your team should be looking at daily.
2 – A KPI You Chase Every Month Stops Telling You Anything
Why it’s needed: a KPI is a tool for measuring progress. Once it becomes the endgame it stops measuring anything. Confusing the two is the most common version of this mistake.
Cyclical work and fundamental change do different jobs. Short-term pushes on a metric like ROAS are the espresso: useful when you need to get through the day and ruinous as a diet. If you do push on specific metrics, then rotate which one you push on each month.
Most of the attention belongs on sustained work: product quality, operational efficiency and customer experience. Those changes hold when nobody is pushing on them.
Shiny metrics are easy to measure and give immediate feedback, which is most of their appeal. Slower measures such as customer lifetime value and retention rates track the fundamentals instead. Customer lifetime value is the total contribution one customer produces across every order they place.
Use a KPI to ask a question rather than to declare a result:
- Why is our customer acquisition cost rising?
- Are we seeing diminishing returns on ad spend?
- Do customers stay past the first purchase?
TRY THIS: take five minutes and a napkin. Write down the last KPI you had a meeting about, then decide whether it’s still a tool or has quietly become a goal. Ask a colleague the same question and compare answers.
3 – Ask Your Team What They Are Optimizing For, Then Check It Against the Plan
Why it’s needed: agencies and internal teams drift toward short-term KPIs, usually because that’s what leaders ask about and pay for. The assumption underneath is that short-term wins add up to long-term performance. They often don’t.
Open communication is what keeps the daily work pointed at the strategy.
1. Ask them the right questions.
- What are you trying to achieve? Pay attention if the answer is a list of KPIs.
- How are those metrics moving us toward the long-term goals?
2. Compare the answers to your strategic plan. If the daily work is pointed somewhere your plan isn’t, then something has to change, and it’s usually the incentive rather than the person.
4 – Teams Optimize What You Measure Them On, So Change That First
Why it’s needed: the doers need to see how their daily work supports the long-term goals. They also need tools and incentives pointed the same way.
1. Adjust the KPIs and targets to reflect long-term success. Make sure each team is measured on numbers that move sustainable growth rather than this month’s total.
2. Rebuild the dashboards. Platform dashboards ship with defaults that are rarely the metrics your strategy needs. Once you’ve defined the KPIs that support the strategy, rebuild each team’s view around them.
3. Tie incentives to long-term outcomes. The most reliable way to hold a team’s focus is to pay for the outcome you want or for the actions that directly produce it.
TRY THIS: take two minutes and answer two questions.
- Have I actually incentivized my team and my agencies?
- What are those incentives tied to, short-term KPIs or long-term goals?
If those two answers don’t match your plan, then a working call against your actual numbers is usually the fastest way to see what your team is really being paid to move.
Those four steps won’t hold without the fifth.
5 – A Leadership Team That Panics at a Dropped Number Undoes the Other Four
Why it’s needed: a team that doesn’t know what matters most costs you profit. That’s what happens when you tell people to focus on long-term strategy and then press them for short-term wins.
1. Get leadership genuinely on board. The most reliable way is to involve them in building the strategy, then to check that each of them can say how their own department contributes.
2. Train leaders to advocate for the long view. Show them how to make the case in meetings and strategy discussions. Show them clear examples of what not to do.
3. Agree the process for a bad number before you get one. Decide now what happens when a key metric drops so the default response isn’t a panic push. Three questions make a workable process:
- Is this single metric really a sign of a long-term problem?
- What fundamental issues could have contributed to it?
- If we work to lift this KPI, how will that affect the rest of our key metrics?
Leadership has to promote the long-term strategy and then act like it when a number moves. Teams read how leadership reacts far more closely than they read the plan.
Take the Broader Perspective Into Next Week’s Meeting
Don’t be the F1 team that only trains the leg press. Three things carry most of the value in this guide.
- Contribution margin per order is the number that tells you whether a sale was worth making: revenue minus every cost that moved because the order happened.
- A category benchmark tells you what businesses like yours tend to keep. Your own four choices on order value, shipping, paid share and repeat rate decide where inside that range you land.
- Keep goals as goals and KPIs as tools. That holds only if the doers and the executives share the same plan and are paid against it.
Action item. Work through this checklist before your next leadership meeting.
- Can three people on the team state our strategic plan the same way?
- What was the last KPI we held a meeting about? Is it still a tool?
- What is one order worth after product cost, fees, fulfillment, returns, discounts and acquisition cost?
- Which of our four position choices is holding our margin below the category?
If that checklist turns up more work than a quarter can absorb, start with the 60-day quick lift. It picks the two or three repairs that pay first.
