Jewelry advertising metrics: what to measure instead of ROAS

Somewhere right now, an agency is sending a jeweler a report with “6.2x ROAS” in a big green box. The jeweler’s bank account has not been informed.

That is the trouble with most jewelry advertising metrics. They describe the ad account beautifully and the business not at all. ROAS is a ratio, and you cannot deposit a ratio. Your bench jeweler has never once asked to be paid in multiples.

We are not against the number existing. Leave the column open if you like it. We are against letting it pick your ads, set your budget and tell you how the store is doing, because it is bad at all three. Most jewelers don’t have an advertising problem. They have an analytical problem.

TL;DR

ROAS tells you the ratio of sales to ad spend and nothing about how much money is left. Grade ads on gross profit per sale: average order value minus cost per purchase. Watch total dollars left, because a lower ROAS on more spend often leaves far more money. Check the store’s own numbers, not only the platform’s. Test whether your margin survives costs rising 10 to 20%. And find out which first purchase brings customers back.

IROAS cannot tell you which ad made money.

Here are two ads.

Ad A: sells a $10 item at a $2 cost per purchase. ROAS is 5x. It leaves $8.
Ad B: sells a $100 item at a $40 cost per purchase. ROAS is 2.5x. It leaves $60.

A ROAS report tells you to kill Ad B. Ad B leaves seven and a half times as much money on every sale.

Now the same thing with jewelry prices (our example, not anyone’s real account). A $400 pair of studs at $80 per purchase is a 5x. A $2,400 ring at $800 per purchase is a 3x. The report crowns the studs. The studs left $320. The ring left $1,600.

Ad B also has somewhere to go. It is easier to lower a cost per purchase than it is to raise an average order value. A $40 cost has room in it: a better ad, a better page, a faster reply. A $2 cost does not. Nobody is getting that one to zero, and you are still selling a $10 item.

There is a second thing the ratio hides. Two ads from the same store can bring in very different buyers, and the order value is how you see it. When you chase the best ROAS, you are often choosing the cheapest customer on purpose and then wondering why nobody comes back.

IIROAS falls when you grow, and that is usually fine.

As you spend more, you reach people who are colder than the ones you started with. They convert less often, and the cost of reaching them tends to rise. So ROAS tends to fall as spend rises. That is not a malfunction. It is what growing looks like.

The reverse is also true. ROAS is at its most beautiful on the smallest, warmest audience you can find, the people who were about to buy anyway. That is why the prettiest number in many accounts belongs to retargeting. It is the marketing equivalent of standing inside your own front door, handing flyers to customers already walking in, and sending the owner an invoice for the sale.

Look at it in dollars, before cost of goods:

$1,000 in ad spend at 5x: $5,000 in sales, $4,000 left after the ads.
$10,000 in ad spend at 3x: $30,000 in sales, $20,000 left after the ads.

ROAS fell by 40%. The money left after ads went up five times. Anyone paid to protect the multiple will keep your budget small forever and call it discipline. A 6x on $300 a month is a very impressive way to make jack shit.

This is also why your starting budget should come from your order value, and why increases should be small and only ever into margin you actually have.

IIIThe column to add: gross profit per sale.

Average order value minus cost per purchase. That is what is left after you bought the customer, and it is the first number we look at.

To build it in Meta Ads Manager:

Go to Columns, then Customize Columns, then Create Custom Metric.
Name it “Gross profit per sale”.
Set the formula to Purchase Conversion Value ÷ Purchases − Cost per Purchase.
Set the format to currency, save it and add it to your column view.

Then sort the ads by amount spent, highest first. The ads spending the most are the ones that matter most.

One thing the label hides. Ads Manager has no idea what the piece cost you. That number is before cost of goods, before the setter and before the rent. It is excellent for comparing ads inside one account and useless for deciding whether you are rich. Take cost of goods off it yourself, by product line, before you make a budget decision.

How to read it:

High order value, plenty left: make more ads like this one.
High order value, not much left: good customers on an expensive path. Look at the page and the follow-up before you blame the ad.
Low order value, plenty left: leave it running. Do not bet the year on it.
Low order value, not much left: turn it off.

And give it time. Early numbers lie, ROAS included. A two-day read on any ad is a horoscope. Where the money is going tells you what Meta has decided, so read spend, CPM and frequency first, and cost per purchase last. We go through that in how to read jewelry ad creative. Advertisers who skip it tend to kill the ads that made the sale possible, then complain about ad fatigue.

IVCheck it against the till.

Every ad platform grades its own homework. Each one’s number is useful, and none of them is the truth.

Jewelry makes this worse than most categories. Someone sees your ad on their phone on a Tuesday, walks in on Saturday and pays at the counter. Meta does not know about that sale unless you tell it. On the other side, a platform can claim a sale that was going to happen anyway.

So once a month, from your own store system and your own bank statements, write down three numbers:

Total sales.
New customers.
Total ad spend across every platform.

Total ad spend divided by new customers is your blended cost per new customer. It is the number no platform can flatter, and when the dashboard says “record month” and the till does not, believe the till.

Read the two together. If Meta claims 10 sales and the store saw 15, you can afford a higher cost per purchase on Meta than you thought. If Meta’s own numbers look a little worse but email and search picked up more orders, that is a win, and it will never appear in a ROAS column. To close the gap as far as it can be closed, send your store’s sales back to Meta through the Conversions API.

VFind out if your margin can take a punch.

“Am I profitable?” is the wrong question. Growing raises your cost per purchase and adds some bad weeks, so the real question is whether you have room for that.

Take your numbers and raise the cost per purchase by 10 to 20%. Here is 20%, with cost of goods included this time (examples, not real accounts):

A $2,400 ring, $1,200 cost of goods, $800 per purchase: leaves $400. With the cost at $960, it leaves $240. You can grow this.
A $150 silver pendant, $60 cost of goods, $70 per purchase: leaves $20. With the cost at $84, it leaves $6. That pendant now pays you less than the sandwich you ate while you shipped it.

The first one survives the punch, so scale it in small steps. The second one needs a lower cost or a higher order value before it gets another dollar. Scaling a sale that barely works does not make it work. It just makes it louder.

VIThe metric nobody puts in the monthly report.

The first sale carries the entire cost of finding the customer. The second one, sent by email to someone who already trusts you, carries almost none of it. That is where a jeweler’s real margin lives, and it does not appear anywhere in Ads Manager.

The baseline is low. Bluecore’s benchmarks, as reported by MobiLoud, put the repeat rate for first-time luxury and jewelry customers at about 9.9% within a year. So there is plenty of room, and small moves count.

The math is simple. If a first sale costs $500 in ads and that customer later buys again from an email, you paid $500 for two sales: $250 each. Every repeat lowers what you really paid, which is what lets you pay more up front than a competitor can.

Pull two things from your store data:

Your second-purchase rate. How many first-time buyers from a year ago have bought again?
The piece they started with. Your bestseller and the piece that brings people back are not always the same piece. The ads should lead with the second one.

If you cannot measure lifetime value yet, order value is a decent stand-in: customers who spend more are more likely to come back. And if nothing brings them back, start with email. Ads acquire. Retention compounds.

VIIWhat to ask whoever runs your ads.

No trick questions. Whoever is spending your money should be able to answer these from memory or within a day.

What is the gross profit per sale on our top five ads by spend?
What did the platforms report in sales last month, and what did the store actually record?
What is our blended cost per new customer, this month against last?
Can our margin take a 20% rise in cost per purchase?
What is our second-purchase rate, and what are we doing about it?

If the answer to all five is a bigger font on the ROAS, you have learned something useful. We do not care about vanity metrics. We care about business growth. For the rest of what to look for, see whether you should hire someone to run your ads at all.

VIIISelf-diagnostic: is the report telling you anything?

The metrics diagnostic.

Do you know the gross profit per sale on your top five ads?
Have you taken cost of goods off that number, at least by product line?
Do you sort your ads by amount spent, not by ROAS?
Do you wait for real spend before judging an ad?
Do you compare what the platforms report with what the store recorded, every month?
Do you know your blended cost per new customer?
Could your margin survive a 20% rise in cost per purchase?
Do you know your second-purchase rate?
Do you know which first piece brings customers back?

Three or more no’s and the problem is not your ads. It is that nobody can tell you whether they work.

Straight answers

What metrics should a jewelry store track for advertising?

Gross profit per sale (average order value minus cost per purchase), total dollars left after ad spend, blended cost per new customer from your own store records, and your second-purchase rate. Read spend, CPM and frequency first when judging an individual ad, and cost per purchase last. ROAS can stay in the report, but it should not make the decisions.

What is a good ROAS for jewelry ads?

There is no useful answer without a price and a margin attached. A 3x on a $2,400 ring and a 3x on a $150 pendant are completely different businesses, and a benchmark is just someone else’s margin. Work out how much is left per sale after the ad and the cost of goods, then decide whether that is enough for you.

Is ROAS useless?

No. It is a quick read on the relationship between sales and spend, and a sudden change can be worth looking into. It is just a poor way to choose between ads, set a budget or judge the business, because it says nothing about order value, cost of goods or how much money is actually left.

How do you calculate gross profit per sale in Meta Ads Manager?

Go to Columns, Customize Columns, Create Custom Metric. Use the formula Purchase Conversion Value ÷ Purchases − Cost per Purchase, format it as currency and add it to your column view. Ads Manager does not know your cost of goods, so take that off separately before making a budget decision.

Why does ROAS drop when you increase your ad budget?

More spend reaches people who are colder than your first buyers. They convert less often and tend to cost more to reach, so the ratio falls. That can still be a good trade: 3x on $10,000 leaves far more money than 5x on $1,000. Keep increases small, and only spend into margin you actually have.

Why don’t Meta’s sales numbers match my store?

Meta only counts what it can connect to an ad. A customer who saw an ad and later paid at the counter may not be counted, and a platform can also claim a sale that would have happened anyway. Compare both every month, trust your store records for the business decision, and send sales back through the Conversions API to narrow the gap.

How do you measure lifetime value for a jewelry store?

Start with your store records: how many first-time buyers bought again within a year, and how much they spent. Because jewelry purchases are far apart, this can take time to show up. Until you have enough history, order value is a reasonable stand-in, because customers who spend more are more likely to come back.

This is part of our fine-jewelry operating series. For the budget side of this, see how much a jewelry store should spend on Meta ads. For measurement plumbing, see the Conversions API for jewelry stores.

Add the column. Sort by spend. Check the till. Punch the margin. Pull the repeat rate. If you want an operator who reports what is left instead of what is pretty, see how we work or email info@endicodatastrategic.com directly. The founder reads every inquiry within one business day.

Work with us

Want an operator running this instead of reading about it?

The founder is on every account, and we do not work with two jewelers in the same trade area. No binding contracts, and you own everything: the ad account, the pixel, the creative, the audiences.

Book a strategy call →