Should my next $10k a month go to more Meta spend or to the website?

Why doesn't more spend bring more sales?

You raised Meta spend by a third over the summer. The dashboard still shows a respectable ROAS. Revenue went up, but nowhere near a third, and when you look at net sales after discounts and refunds, the quarter feels flat. Now there is another $10k a month in the budget and two people in the room with opinions.

Your media buyer wants it in ads: the account can take more, give it room. Your site person wants it on the website: why pour more traffic into a leaky bucket? Both can be right. Which one is right for you this quarter is a question about numbers you already have.

Here is the short version. Spend and the site are not competing for the same job. Spend decides how many people arrive. The site and the offer decide what each arrival is worth. The question is which of those two is currently the weaker link, and whether the next dollar of spend still earns more than it costs.

What is revenue per ad dollar actually made of?

Every dollar you put into Meta turns into revenue through a chain of three numbers. Each one has a different owner, and each one can move on its own.

Now split last quarter against the quarter before. The example below is a brand that raised spend and got a fraction of the growth it expected.

Read it from the top. Spend rose 38% and revenue rose 11%. Almost all of the gap sits in sessions per dollar: each dollar bought 18% fewer visits. The site lost a little, the offer gained a little. In this account, the website is not what broke. The extra spend simply bought more expensive, lower-quality visits. Redesigning the product page here would have spent the $10k on the wrong number.

Flip the pattern and the answer flips. If sessions per dollar held and conversion rate slid from 3.0% to 2.4%, the media was doing its job and the site or the offer was dropping the ball.

Where to get the numbers: spend from Ads Manager, sessions, orders and revenue from Shopify for Meta-tagged traffic, by week, US only. Use Shopify rather than Meta's reported purchases for the bottom of the chain. Our guide on Meta ROAS vs Shopify revenue explains why the two never match.

What did your last spend increase actually return?

This is the number almost nobody looks at, and it decides the question. Your ROAS in Ads Manager is an average: total revenue divided by total spend. It blends your best dollars, the ones that reach your warmest buyers, with your last dollars, the ones Meta had to work hardest to place. The next $10k will not earn the average. It will earn something closer to what the last $10k earned.

Marketing mix modelers have a name for this. Google's Meridian documentation defines marginal ROI as the return on a small increase in spend, and says that "if the mROI is much lower compared to the ROI, then the channel is beginning to saturate at historical spend level" (Google Meridian). Meta's own open-source model, Robyn, calls the same idea the "next dollar response" (Robyn, Meta Marketing Science).

You do not need a model to get a first read. Find the last time spend went up and stayed up, and compare a stretch of weeks before with a stretch after.

Say your contribution margin before ad spend (after product cost, shipping, payment fees and discounts) is 60%. Your break-even ROAS is then 1 ÷ 0.60, about 1.67. The average of 2.28 looks comfortably profitable. The marginal 1.40 is not: the extra $10,000 a month brought $14,000 of revenue, $8,400 of contribution, and a loss of $1,600 a month. Nobody sees that loss on the dashboard, because the average hides it. Our guide on what a good ROAS is for a supplement brand walks through the break-even math.

Keep the read honest

• Match the volume. Compare weeks with similar spend on each side, so one big launch week does not pass for the new normal.
• Match the season. Compare the same weeks last year as well. A spend increase in October meets a pricier auction no matter what you do.
• Check total net sales too. Some of Meta's effect shows up in direct, branded search and email. If store-wide net sales rose more than Meta-tagged revenue, the marginal return is better than the Meta-only line says. If they rose less, it is worse.
• Subscriptions change the bar. If first orders roll into subscriptions, the right break-even is on payback, not first-order revenue. Our guide on LTV to CAC for subscription brands covers that version.

Why does each extra dollar on Meta buy less than the last?

Diminishing returns on ad spend are not a theory agencies use as an excuse. They are built into how the delivery system works, and Meta says so in its own documentation.

When Meta bids on your behalf for the lowest-cost results, its Marketing API documentation spells out the catch: "Costs can rise as you exhaust least expensive opportunities or as you increase budget" (Meta for Developers). In plain terms, the system spends your first dollars on the cheapest conversions it can find. When you hand it more, it has to go further out: people less likely to buy, placements less likely to produce a real visit, or the same buyers at a higher price.

The auction adds to it. Meta's auction does not simply pick the highest bid; it weighs several factors, and "estimated action rates and ad quality measure ad relevance" (Meta). The people most likely to act on your ad are a limited pool. Once the likeliest have seen it, more impressions tend to go to people less likely to act.

And the market itself keeps getting more expensive. In Q2 2026, ad impressions across Meta's apps rose 14% year over year, and the average price per ad still rose 12% (Meta). At those prices, a flat budget buys less than it did a year ago, and a growing one meets that on top of its own diminishing returns.

Robyn puts the general rule in one line: "each additional unit of advertising investment increases the response at a declining rate" (Robyn, Meta Marketing Science). The curve bends for every brand. The only question is where you are on yours.

What does the same $10k do on the site instead?

Take the brand from the last table and give it two choices. It spends $40,000 a month on Meta, Meta-tagged revenue is $100,000, and the whole store does $250,000 a month in net sales across all channels. Contribution margin before ad spend is 60%.

Four things fall out of that table.

1. Spend above break-even wins. At a marginal return of 2.0, more spend makes money every month and beats a site fix over the first quarter. If your last increase earned well above break-even, keep spending.
2. Spend below break-even loses every month. At 1.4, the loss repeats for as long as the budget stays up, and the average ROAS keeps it hidden.
3. Where the fix lands matters. A checkout or offer fix touches every session from every channel: email, search, returning customers. A landing page fix only touches the traffic sent to that page. Same lift, very different value.
4. A fix costs once, spend costs every month. That is why the gap widens with time. It is also why site work has to be chosen by what it is worth, not by what looks most broken.

One more line worth noticing: a 5% conversion lift turns a 1.40 marginal return into about 1.47. Better, and still under the 1.67 break-even. Site work raises what every dollar earns. It does not repeal diminishing returns. If the curve is bent hard, the answer may be new creative angles and new audiences, not either option above.

If the site budget is a monthly program rather than a one-off fix, compare the lift it finds each month with the marginal return of the same money in ads.

When is more Meta spend the right answer?

More spend is the right call when the curve has not bent yet. You will usually see all of these at once:

• The last increase paid. The marginal return on your most recent step up sits clearly above break-even ROAS, on volume-matched weeks.
• Frequency is low and stable. Spend went up and frequency barely moved, so Meta is finding new people, not showing the same ads to the same people more often.
• Reach grows with spend. Weekly reach rose roughly in line with the budget.
• Sessions per dollar held. CPM and the share of impressions that become visits stayed steady as budget rose.
• Creative is not saturated. New angles still find winners, and your best ads are not sliding in click-through rate week after week. Our guide on how many creatives to test covers what a healthy pipeline looks like.
• The site already converts. Conversion rate on paid traffic is at or near what similar stores see, so there is no obvious leak eating the extra visits.

When those line up, the site is not the bottleneck. Raise spend in steps, read the marginal return after each one, and stop stepping when it drifts toward break-even.

When is the website the right answer?

The site is the better home for the $10k when the visits are fine but too few of them turn into orders, or the orders are worth too little.

Conversion rate is below your category

For context, Shogun's H1 2026 data across 745 active Shopify stores puts Health and Wellness at a 3.31% median conversion rate (orders divided by sessions), against a 1.74% median for all stores (Shogun). That is a blended store number, including email and returning customers, so cold Meta traffic usually sits lower. Our guide on conversion rate benchmarks for supplement stores shows how to compare like with like.

The checkout leaks

Baymard Institute puts the average documented cart abandonment rate at 70.22%, across 50 studies. In its survey of US online shoppers who abandoned during checkout (leaving out those who were just browsing), the top reason was extra costs too high, such as shipping, tax and fees, at 40% (Baymard Institute). Baymard also estimates that the average large ecommerce site can gain a 35.26% increase in conversion rate through better checkout design. Treat that as a sign of how much room checkouts usually leave, not a forecast for yours. If add to carts are healthy and orders are not, start with our guide on add to cart but no purchases.

The offer is the problem

Price, bundle structure, the subscribe and save discount and the free-shipping threshold set both conversion rate and AOV. If AOV is flat while costs rise, or the subscription offer confuses first-time buyers, no page design fixes it. See our guide on subscribe and save discounts.

The page does not match the ad

Cold traffic from an ad about one problem landing on a generic product page is a common leak. Our guide on landing page vs product page for Meta traffic covers when a dedicated page earns its keep.

What are the signs that more spend is hitting a wall?

Diminishing returns rarely announce themselves. They show up as small drifts once you put spend next to the rates, day by day.

The first two are the clearest. If CPM held steady but each dollar buys fewer sessions as budget grows, you are not paying more for attention; you are getting worse attention for the same price. And if the days you spend the most are the days you convert the worst, the marginal buyer is colder than the average one. Both mean the next $10k will land on the weakest part of the curve. If CPA has climbed alongside, our guide on a rising Meta CPA splits that part.

Who should make the call?

Part of why this question is so hard to answer is that nobody owns the whole chain. The media buyer is measured on ROAS and spends what they are given. The site team is measured on conversion rate and wants a redesign. Each owns one number, and revenue per ad dollar falls between them.

That is a structure problem as much as a data problem, and it is one of the reasons brands stall around the same revenue band. Our guide on why DTC brands plateau covers the pattern. If you are deciding whether that ownership should sit with an agency, a freelancer or a hire, see agency vs freelancer vs in-house.

How we decide between spend and the site at Succession

When a founder asks us where the next $10k should go, we do not start with an opinion about media or design. We start with a split, and the split usually makes the decision for us.

1. One market first. We isolate the US before reading any rate. Mixed international traffic drags every number down and hides real movement.
2. Clean sessions. We filter Meta's crawlers and link checkers out of session data. They arrive in bursts from Meta's data centers when ads are created or edited, mostly showing as Sweden and Ireland, and they never buy.
3. The three-way split. Sessions per dollar, conversion rate and average order value by week, with Shopify orders as the scoreboard and revenue per session next to ROAS, so we know which number moved.
4. Volume-matched days. To read what the last increase returned, we compare days with similar spend, and the same weeks last year, so scale and season do not pass for performance.
5. Per-page funnel steps. Session to add to cart to checkout to order for each landing page, not sitewide, so a leak on one page is not averaged away.
6. A leak map ranked by orders per day. Each leak gets a ceiling in orders per day, so site work and a spend increase compete on the same unit.
7. Checkout and offer before redesign. Those fixes touch every session. A new page design comes after, if the map still points there.

Our decomposition workbook stays on our side. The split itself is something any founder can ask their team or agency to produce this week.

What should you do this week?

1. Pull the last two quarters by week, US only: Meta spend and CPM from Ads Manager, and sessions, orders and revenue from Meta-tagged traffic in Shopify. Add total store net sales.
2. Compute sessions per dollar, conversion rate and AOV for each quarter. Write down which one moved most.
3. Find your last spend increase. Compare volume-matched weeks before and after, divide the extra revenue by the extra spend, and put that marginal return next to your break-even ROAS.
4. Sort the last 60 to 90 days by spend. Compare sessions per dollar and conversion rate on your top-spend days with your middle days.
5. Map session to add to cart to checkout to order for each landing page, and list what a buyer pays at checkout that they did not see on the page.
6. Decide: if the marginal return is above break-even and frequency is low, raise spend in steps. If not, put the $10k on the checkout and the offer first.

If you would rather run the split with someone, our free growth audit goes through your last quarter live and shows which number is holding the next dollar back.

FAQ