In the summer of 2025, Valentte, a UK maker of candles and reed diffusers, put a new bundle offer live. Its founder, Luke Bream, has told the story twice on camera since, once on his own show and once at an industry event. Sales jumped by roughly $270,000 a week overnight (we have converted his figures to US dollars), and they stayed there. It looked like the week every store owner hopes for.
Underneath, the store was losing roughly $40,000 to $67,000 every week. Ad spend had climbed past 30 percent of revenue. By his own account, the company went from very profitable to a loss of about $3.3 million for the year. And for the first few weeks, he did nothing about it, because every number he was watching said the growth was working.
That is the trap in scaling an online store. The dashboard reports averages, and averages are made mostly of the customers you won before you pushed. The question that decides whether to spend more is narrower: what did the last increase in spend actually buy, and did those customers make you money on their first order? Here is how to answer it with numbers your store already has, and what to watch in the rest of the business while you do.
Your Average Cost per Customer Counts People Who Were Coming Back Anyway
Start with the number most owners use to judge their ads: total ad spend divided by total orders, or by total customers. It feels like a cost per customer. It is really a blend of two very different groups.
Some of the people your ads reach have bought from you before. A returning customer sees your ad, remembers you, and buys again. They are cheap to win, because they already trust you. In Valentte's store, Bream puts the ad cost of bringing an existing customer back at about one-tenth the cost of winning a new one. Blend those two together and the new customers look far cheaper than they are.
Here is a simple example store. These numbers are made up to show the math:
- Ad spend this month: $6,000
- Orders from returning customers: 250
- Orders from first-time customers: 150
- Blended cost per order: $6,000 / 400 = $15
- Cost per new customer: $6,000 / 150 = $40
Fifteen dollars and forty dollars describe the same month. If the owner judges growth on the $15, every budget increase looks safe. The $40 is the one that matters, because growth means new customers, and new customers are what each extra dollar has to buy.
Treat every dollar as spent on new customers when you do this math. Some of it reached returning buyers, so the true new-customer cost is a bit lower. But counting the whole budget gives you a number you can trust not to flatter you, and it is the number that rises when you push.
You do not need special software to get it. In Shopify, the New vs returning customers report (Analytics, then Reports, Customers category) gives you the count of first-time customers for any period. Divide your total ad spend across every platform for the same period by that count. Write it down every week.
The First Customers Your Ads Find Are the Cheapest Ones You Will Ever Buy
Even the new-customer number is an average, and it hides the part that decides whether to scale.
Ad platforms spend your first dollars on the people most likely to buy: the ones already searching for what you sell, or the ones who look most like your past buyers. Those people convert quickly and cheaply. There are only so many of them. Every extra dollar has to reach someone a little further away, someone less ready, less familiar, harder to convince. They cost more to win.
So when you raise your budget, the new money does not buy more customers at your old average price. It buys the next customers in line, and they are always pricier than the ones before. Bream describes his own new-customer cost as an average of people won for very little and people won for two or three times as much. "As you spend more, the cheap customers run out first," he says. Every extra bit of budget goes on people who are harder to reach.
The averages move slowly, because the cheap early customers are still in them. That is why a store can be losing money on its newest customers while its average cost per customer still looks fine.
We have covered how each platform behaves when you raise a budget: Meta spends your easy money first, and a bigger Google Ads budget buys pricier clicks before it buys more of them. This article is about what that looks like in your store's own numbers, across every channel at once.
Your Last Budget Increase Paid More per Customer Than Your Average Shows. Divide to Find Out How Much.
There is one calculation that tells you what your most recent increase bought. It takes five minutes and two periods of numbers.
Take the example store. Last month it spent $6,000 and won 150 new customers. This month it raised spend to $9,000 and won 190 new customers.
- Extra spend: $9,000 - $6,000 = $3,000
- Extra new customers: 190 - 150 = 40
- What each extra customer cost: $3,000 / 40 = $75
The new average is $9,000 / 190, about $47. That looks like a modest rise from $40. But the 40 customers the extra money bought cost $75 each, nearly twice the old average. The average has spread the damage across everyone, including the cheap customers you were already getting.
Now compare that $75 to what a first order actually earns. Not the order total. What is left after you pay for the product, packaging, shipping, card fees, and the share of orders that come back. If you have not worked that number out, our pieces on the return your ads need to break even and which of your products actually makes money walk through it.
Say a typical first order at this store leaves $55 after all of that. At the old average of $40, each new customer made $15 on the first order. Each of the 40 extra customers lost $20. The extra $3,000 bought $2,200 of first-order profit. The store spent more, sold more, and made $800 less from those new customers than if it had never raised the budget. Every report it was watching said the month went well.
A few rules keep the comparison honest:
- Compare like with like. Two weeks in November will beat two weeks in August no matter what you spent. Compare against the same stretch before the change, or the same weeks last year, and be suspicious of a jump that lines up with a holiday, a sale, or a product launch.
- Give the sales time to arrive. Some people buy days after they first see an ad. Judge a budget change after at least a week, and two is better for a small store with few orders.
- Change one thing at a time. If you raised the budget, launched a new offer and redesigned the product page in the same week, you cannot tell which one did what.
- Use a range, not a single number. With 40 extra customers, one good or bad day swings the answer. If the number comes out at $75, read it as somewhere between $60 and $90, and decide from there.
Without a Top Price for a New Customer, a Bigger Ad Budget Just Buys Losses Faster
Once you can see what the next customer costs, you need a line you will not cross. Without one, the next increase always looks reasonable, because the averages never sound the alarm.
The line comes from the first-order number. If a first order earns $55, a new customer you win for more than $55 loses money on the way in. You can choose to accept that, but only on purpose.
Valentte's version of the rule is simple: scale spend until new-customer cost hits a number, then stop. Bream now goes further. His paid-media lead has no spending limit at all, only a target for profit from new customers on their first order, every day. Spend can go as high as it likes, as long as that profit target still clears.
Two practical notes on setting your own line:
Profit on the first order is the safe default for a small store. You can pay more than a first order earns if customers reliably come back, but then you are counting on money that arrives months later to cover a bill that arrives this month. That is a timing problem, and we covered it in Lifetime Value Does Not Pay This Month's Invoice. Bream's store got its payback window from about 150 days down to zero, meaning it now makes money on every first order. His view is that a long payback window is dangerous for anyone who is not very experienced or has no outside money behind them.
If you are already over the line, the fix is to spend less, not more. This is the part owners resist. When your newest customers cost more than they earn, turning the budget down drops the costliest ones first and brings your cost per new customer back under the line. Then you work on the things that make every customer cheaper to win: better ads, a stronger offer, a product page that converts more of the people who arrive. When those improve, you raise the budget again, and the line holds further out.
There is also a rough check you can run in your head: ad spend as a share of revenue. During its bad stretch, Valentte's ad spend ran above 30 percent of revenue. Today it runs between 11 and 15 percent, and the business is profitable again. Bream's own view is that most stores struggle above 20 or 25 percent, because few small businesses have their other costs tight enough to carry it. That is one founder's opinion, not a rule, and the right number depends on your margins. But if your ad spend is a third of your revenue and you do not know what your newest customers cost, find out before your next increase.
The ad platforms will let you set a cost limit inside the account too. Meta and Google both have bid strategies that take a cost or return target, and they will try to stay near it. Treat them as a speed limit, not a guarantee. They can run over it, and they are only as good as the target you give them. The target still has to come from your store's own first-order number.
A Sales Jump Breaks Your Shipping and Support Before It Breaks Your Ads
The money is only half of what goes wrong when sales jump. The other half happens in the building.
Valentte ships its own orders, and Bream laid out exactly what broke when sales jumped, in order:
- Shipping. On a normal day there were about 500 orders still unshipped at 5 p.m. That number went past 10,000.
- Customer service. About a week later, open support tickets went from around 50 to around 1,200. Late orders produce "where is my order" emails, and those arrive after the delay, not during it.
- People. He brought in 90 temporary staff overnight. That overwhelmed the managers and supervisors, who were now training new people while fighting the backlog.
- Cost per order. The labor cost to make, pick, pack and ship each unit went from about 47 cents to more than $1.70. Nearly four times as much, on tens of thousands of units a day.
- Stock. The company used up its own materials, then the stock its suppliers held for it, and then suppliers started sending substitute parts that did not fit its filling machines. Work went back to hand-filling.
None of that shows up in an ad account. Your ads can be bringing in customers at a fine cost while late shipping, slow replies and refund requests quietly turn those customers into one-time buyers who leave bad reviews. That costs you the repeat orders your numbers were counting on.
The lesson Bream draws is the useful one: growth comes in steps, and each step up breaks something. It might be shipping, support, cash, stock, or the owner's own time. After each step, look for what broke and fix it before you push again.
For a small store, that means watching a few numbers every day while you scale, not once a month:
- Orders not yet shipped at the end of the day. Know your normal number. If it climbs two days running, slow the ads before it becomes a backlog.
- Open customer emails or tickets. Again, know your normal. This one lags shipping by about a week, so a rise means a problem that started earlier.
- Days of stock on your best sellers. Ads that sell out a product in a week are buying customers you cannot serve. Scaling spend and running out of stock is the worst of both, and it ties up cash in the slow movers you still have, the problem we covered in You Did Not Lose the Money. You Turned It Into Inventory.
Raising Ad Spend in Small Steps Shows You What Broke. Doubling It Overnight Buries the Damage.
Put the two halves together and scaling becomes a routine you can repeat, not a leap.
- Know your line. Work out what a first order earns after every cost, and decide the most you will pay for a new customer.
- Raise spend in a modest step. Something like 20 percent at a time keeps the platforms stable and keeps the math readable.
- Wait. At least a week, two if you can. Keep everything else the same.
- Divide the extra spend by the extra new customers. Compare the answer to your line, as a range, not a single number.
- Check the building. Unshipped orders, open tickets, stock on your best sellers. Is anything climbing?
- Decide. If the extra customers cost less than they earn and nothing broke, take another step. If they cost more, step back down and work on the ads, offer or page. If something broke, fix it before you grow again.
Doubling the budget overnight skips all of that. The extra customers arrive mixed in with your normal ones, the averages absorb the damage, and the shipping backlog shows up a week after you stopped watching. Small steps cost you a little speed. They also show you, step by step, the price of the next customer and the first thing that breaks.
One more thing changes as you scale, and it is worth knowing before you start: your store's conversion rate will usually drop. The new visitors your bigger budget reaches are less ready to buy than the ones you were getting, so a smaller share of them buy. That is not a broken site. We explained why in Your Conversion Rate Went Up. Your Store Made Less Money. Judge the step by what the extra spend bought, not by whether the store-wide rate held.
Bream summed up the cost of learning this the hard way: "We really damaged the business getting here." Most stores never face a jump as big as Valentte's. The same thing still happens one budget increase at a time, slowly enough that the averages never catch it.
This is how we look at growth when we work with online stores: what the next customer costs, what the first order earns, and what in the business has to hold before the budget goes up. If you want a second set of eyes on your numbers before your next increase, our ecommerce marketing team does this every week. If you would rather do it yourself, start with one number this week: your extra spend divided by your extra new customers.
