When should an ecommerce business increase its ad budget?

Distinguish average performance from the next spend increment

Your average return is total ad revenue divided by total ad spend. Your marginal return is what the extra spend brings in: the additional revenue divided by the additional spend.

They differ because audiences are not uniform. The first dollars tend to reach the people most likely to buy: returning visitors, people already searching for what you sell, people who resemble your best customers. As spend rises, the platform has to reach further, into people who are less ready or less suited. Each extra dollar usually does less work than the one before. How quickly that happens depends on the product, the audience size and the creative, which is why nobody can tell you in advance.

So an account can show a healthy average while the latest increase loses money. Here is how.

Take a hypothetical store spending $10,000 a month on ads, with $30,000 of revenue credited to them: an average ROAS of 3.0. After product costs, freight, payment fees and returns, it keeps 50% of revenue as contribution before media. Its break-even ROAS is therefore 2.0. Below that, ads cost more than the orders contribute. The founder is considering adding $5,000 a month.

Look at the last row. Blended ROAS is 2.33, comfortably above the 2.0 break-even, and the dashboard looks fine. But the extra $5,000 brought in only $5,000 of revenue, and the business is now $2,500 a month worse off than before the increase. The average hid the loss.

The question is not whether your ROAS is above break-even. It is whether the increment is. If you have not worked out your break-even point yet, start with what a good ROAS is for your business, or use the break-even ROAS calculator.

How to estimate the marginal return

You cannot read marginal return directly off the ad platform. You estimate it by comparing comparable periods before and after a change, using store numbers as well as platform-reported revenue, because platforms can credit themselves with sales that would have happened anyway.

• Compare total store revenue and new customers, not just platform-reported revenue, before and after the increase.
• Change one thing at a time. If you raise spend, launch new creative and run a sale in the same fortnight, you will not know which did what.
• Allow for seasonality. An increase in the run-up to a gifting peak will look better than the same increase in a quiet month.
• Accept uncertainty at low volumes. With a few dozen orders a week, normal variation can be larger than the effect you are trying to measure.

Check stock, cash and operational capacity

An increase that works inside the ad account can still hurt the business. Before you spend more, check three things outside it.

• Stock.

  More orders need more product. If the increase in the example works as hoped, an extra $15,000 of monthly revenue at a $100 average order is about 150 extra orders a month. If the best seller runs out halfway through, you are paying to send shoppers to a sold-out page, and pausing mid-test muddies what you learn.

• Cash.

  Ad spend leaves your account quickly. Cash from orders follows, but replacing the stock those orders used may need paying for weeks or months before it sells again. A profitable increase can still leave you short in between. Map when money goes out and comes in over the next three months before you commit.

• Capacity.

  Someone has to pick, pack, answer questions and handle returns. If dispatch is already stretched, 150 more orders a month means slower delivery, and reviews follow. The same applies to marketing: higher spend shows the same few ads more often, so you usually need more creative to keep them working.

If any of these is tight, the answer may be 'not yet', or 'increase, but only to the level stock and cash can carry'. Our guide to setting an ecommerce marketing budget covers how to plan cash timing alongside media.

Plan a bounded increase and observation period

A bounded increase is a test with limits set before it starts: how much more you will spend, for how long, what you will measure, and what result means continue or stop. It turns 'let's see what happens' into a decision you can make on evidence.

1. 1

   Set the size of the step

   Choose an increase your stock, cash and team could carry if it works, and that you could afford to lose if it does not. There is no universally correct percentage. A step that is too small can disappear into normal weekly variation. One that is too large puts a lot of money on a guess.

2. 2

   Decide where the money goes

   Adding budget to an existing campaign, opening a new audience and testing new creative are different tests. Be clear which one you are running.

3. 3

   Write down the threshold

   Using your break-even ROAS, state the marginal return that would justify keeping the increase, and the store-level numbers you will check: total revenue, new customers and contribution after media.

4. 4

   Fix the observation period

   Pick a period long enough to cover your normal weekly pattern and produce enough orders to read, and leave everything else alone during it. Each platform describes its own learning or adjustment behaviour in its documentation, and those rules change, so check current guidance rather than relying on a rule of thumb.

5. 5

   Record the baseline

   Capture the same numbers for a comparable period before the increase, so the comparison is clear when the test ends.

Decide what counts as success before you spend the money, not after you have seen the numbers.

The same logic applies across channels, with a difference in what extra budget buys. On Meta, more budget mainly means reaching further into audiences. On Google search and Shopping, much of it competes for searches that already exist, and once you appear for most of them, extra budget has less to buy or moves into broader placements. Our pages on Meta ads for ecommerce and Google Ads for ecommerce describe how we approach each.

Define reasons to continue, pause or change direction

When the observation period ends, most of the decision should already be made. You wrote the threshold down. Now compare.

Sometimes the answer is neither more nor less, but different. If every increase lands below break-even, the constraint is probably not budget. It may be the offer: at the current average order, a customer may simply cost more to win than they are worth. It may be conversion on the site, or creative that only works on the warmest audience. That is when a plateau stops being an ad question, and why ecommerce sales plateau covers how to work out which it is.

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

And if the increments keep clearing break-even with room to spare, and the business can carry the orders, keep stepping up. One bounded test at a time.