Why ecommerce sales plateau, and what to investigate next

Separate a real plateau from normal variation

Before you diagnose anything, check that there is something to diagnose. A month that is lower than the one before is not a plateau. Seasonality, a promotion last year, a stock-out, a price change or one large wholesale order can make a flat year look like a fall, or a fall look flat.

The question I ask first is: flat compared with what? Look at the same store four ways before you decide.

Read month on month, this store is falling. Read against last year, it is growing. The trailing twelve months tell the real story: revenue is flat, orders are down, and a higher average order (up from about $95 to $100) is holding the total up. That is a plateau, and a more specific one than ‘sales stopped growing’.

• Use rolling twelve-month totals, or compare each month with the same month last year, so seasonality cancels out.
• Split revenue into orders and average order value. Flat revenue with falling orders is a different problem from flat revenue with flat orders.
• Mark the months with promotions, stock-outs, price changes, site changes and large one-off orders. Explain those before you explain the trend.
• Take out wholesale, marketplace or trade revenue if it is lumpy, and read the direct store on its own.
• Check that tracking did not change underneath you. A new checkout, theme or analytics setup can move numbers the business did not.

If, after all that, revenue and orders have been flat on a like-for-like basis for several quarters and no single event explains it, treat the plateau as real.

Check acquisition, conversion and repeat purchase

Revenue is the product of a few numbers: how many people arrive, how many of them buy, how much they spend and how many come back. A plateau is what you see when one of those falls while another rises, or when they all stop moving at once. The total hides which.

Here is a different hypothetical store, averaging about $60,000 a month in both periods.

The owner of this store sees a flat line. Underneath it, the business is spending 37.5% more on ads to buy more first orders, converting a smaller share of a larger audience, and losing almost a third of its repeat orders. New customers are refilling a bucket that has started to leak.

Once you have your own version of this table, the numbers usually point to one or two of the scenarios below. They are distinct, and each calls for a different next test.

• New customers are getting dearer.

  Spend rises, first orders rise more slowly, and each additional customer costs more than the last. You may have reached most of the people who are easy to find with your current offer and creative. More budget in the same place buys less each time, so the test is new audiences, new angles or a new channel. When to increase your ad budget covers how to read the next spend increment.

• More visitors, the same orders.

  Traffic grows but conversion falls. Either the new traffic is less qualified, from a broader audience, a new channel or a cheap placement, or something on the site got worse. Split conversion by traffic source and device before you blame the website.

• The repeat base is shrinking.

  Returning orders fall while new orders hold. Customers are not coming back at the rate they used to, because the first experience disappointed, the range has nothing for them next, or the follow-up stopped. How to get more customers to buy again works through this one.

• Customers are worth less than they cost.

  The offer puts a hard limit on what you can pay for a customer. At Apothecary by Justin in January 2026, a $69 average order left about $55 of gross profit per customer, and buying one cost about $400. No amount of optimisation crosses a gap like that. The fix was the offer: raising the average order changed what a customer was worth. The full case is on our results page.

• Nothing is broken, and the business is full.

  The numbers are healthy, but the founder, the warehouse or the cash cannot take more. That is a capacity ceiling, covered in the next section.

Ads amplify an offer. They don't fix one that can't pay for itself.

Identify capacity, stock and offer constraints

Some plateaus have little to do with marketing performance. The demand is there, or could be, and the business cannot serve more of it without something else giving way.

• Founder time.

  If you pack orders, answer customers, brief the freelancer and write the emails, growth stops when your week is full. The symptom is ideas that never get done, not campaigns that fail. When to outsource ecommerce marketing covers what to hand over and what to keep.

• Stock and cash.

  Selling more means buying more stock, earlier. If a larger order would tie up cash you do not have, or a supplier's minimum order is beyond you, the ceiling is working capital. Growing into it without a plan can run you out of cash while the sales report looks good.

• Fulfilment and service.

  Late dispatch, rising returns or slower replies are signs that operations have reached their limit. More orders make those worse, and they feed straight back into repeat purchase.

• The range.

  If most revenue comes from one or two products, you may have sold them to most of the people who want them at that price. The next step is a new product, a bundle or a new audience, not more of the same.

• Margin headroom.

  A thin margin caps what you can pay to win the next customer. As you reach colder audiences, the cost of a customer rises until it meets that cap, and growth stops there. Improving profit without chasing more orders looks at contribution and product mix.

There is one more kind of ceiling: the one you chose. If you are later in your career and have decided not to scale, a flat top line is not a problem to fix. The useful question becomes how much profit the business produces at the size you want. Say that early, so nobody plans for growth you do not want.

A capacity plateau looks exactly like a marketing plateau on a revenue chart. The difference shows when demand rises: orders slip, stock runs out, or you do.

Estimate which next test is worth running

Once you have a likely scenario, resist fixing everything at once. Put a rough value on each lever, then weigh it against what the test costs and how confident you are in the assumption behind it.

Using Store B, and assuming 60% of revenue is left after product, shipping and payment costs (contribution before media), three candidate tests look like this.

The ad test has a large revenue figure and the smallest return, and its assumption, $50 for each extra customer when the current blended figure is about $35 and rising, is the one you know least about. The conversion test is worth most on paper, but only if the cause of the drop is something you can fix. The repeat test is smaller and often cheaper to run.

None of these numbers is a forecast. They tell you what each test has to prove to be worth running, and which assumption to check first.

1. 1

   Name the scenario.

   Write down which of the scenarios your numbers point to, and what result would prove you wrong.

2. 2

   Size the lever.

   Estimate the monthly contribution if the test works, using your own margin rather than revenue.

3. 3

   Price the test.

   Include media, production, your own time and any outside help. A test that costs more than a few months of its upside needs a stronger reason.

4. 4

   Set a review point.

   Decide in advance what result means continue, change or stop, and when you will look. At low order volumes, allow longer and treat small differences as noise.

5. 5

   Run one change at a time where you can.

   Two changes at once leave you unsure which one moved the number.

If you would rather investigate this with someone, our ecommerce strategy work starts with the same question: what the business has to earn, and which lever moves first. For working through the order once you know, what to fix first, in the order the money flows is a useful companion.