A jewellery advert can generate clicks, orders and an impressive-looking return on ad spend while the business still loses money. Advertising platforms usually compare attributed sales with media spend. They do not automatically protect the product margin, delivery subsidy, packaging, payment fees, discounts, returns or the cost of replacing the stock that sold. This guide shows how to calculate a jewellery brand's real break-even ROAS, set a maximum acquisition cost and diagnose what to fix before increasing the budget.
ROAS measures attributed revenue against advertising spend. It does not tell you whether enough money remains after every other cost.
Start with the problem: jewellery ad revenue is not profit
Return on ad spend is normally calculated as attributed conversion value divided by advertising cost. If a platform reports £4,000 of sales from £1,000 of spend, the reported ROAS is 4.0, or 400%. That sounds strong because the revenue is four times the media cost. It says nothing about what it cost to supply those orders.
The £4,000 may include VAT, discounts, delivery charges, orders later refunded and sales the customer might have made after another marketing touchpoint. The business still needs to pay for the jewellery, packaging, fulfilment, transaction fees, delivery support, returns, customer service and overhead. A high-revenue, low-margin product can tolerate far less acquisition cost than a product with the same selling price and stronger contribution.
Do not begin with the question, ‘What is a good ROAS for jewellery?’ Begin with, ‘How much can this business afford to pay for this order, product, customer type and market?’ The answer is calculated from your numbers, not borrowed from an agency screenshot or a generic ecommerce benchmark.
| Metric | Simple calculation | What it tells you | What it can hide |
|---|---|---|---|
| CTR | Clicks ÷ impressions | Whether the advert earns a click | Click quality and purchase intent |
| CPC | Ad spend ÷ clicks | Average media cost per visit | Whether the visit converts |
| CPA | Ad spend ÷ attributed actions | Cost per chosen action or order | Whether the action is valuable |
| CAC | Acquisition cost ÷ new customers | Cost to win a new customer | Repeat value and contribution |
| ROAS | Attributed revenue ÷ ad spend | Revenue reported per £1 of ads | Product margin and non-ad costs |
| Contribution after ads | Net sales − variable costs − ads | Money left for overhead and profit | Future refunds not yet known |
Calculate the maximum acquisition cost for a jewellery order
Start with net sales from the order, using a consistent revenue basis. Deduct every variable cost that exists without advertising: product or production cost, preparation labour, packaging, payment fees, marketplace fees where relevant, delivery absorbed by the business, gifts, commissions and a realistic allowance for returns or replacements. The remainder is contribution before advertising.
If the business is prepared to use all of that contribution to acquire the order, the result is the absolute break-even acquisition cost. That leaves nothing from the first order for overhead, owner pay or profit. A commercially safer target keeps a required contribution after ads. Maximum acquisition cost is therefore contribution before ads minus the contribution the business needs to retain.
Calculate by product family and market, not only across the whole store. A £160 ring, a £55 pair of earrings and a £30 add-on can have different product costs, return risk, postage bands and repeat behaviour. One account-wide target may cause the platform to chase the easiest revenue rather than the orders that strengthen the business.
- Use net sales after discounts and expected refunds.
- Remove VAT from the internal revenue basis where the business accounts for VAT.
- Include the true landed or production cost of the jewellery.
- Include packaging, fulfilment time, fees and delivery support.
- Set a required contribution after advertising.
- Recalculate when price, costs, market or offer changes.
Work through a jewellery break-even ROAS example
Consider a simplified £100 order from a business using £100 as its consistent internal net-sales basis. Product cost is £28, packaging and fulfilment cost £6, payment fees £3, delivery support £4 and the returns and replacement allowance £4. The order contributes £55 before advertising.
Spending the full £55 to acquire the order produces a break-even ROAS of £100 divided by £55: 1.82, or 182%. But the order would then contribute nothing towards overhead or profit. If the business requires £20 after advertising, the maximum ad cost becomes £35 and the target ROAS becomes £100 divided by £35: 2.86, or 286%.
This is why break-even and target ROAS are not interchangeable. Break-even is the edge of loss under the assumptions. The operating target needs room for forecast error, refunds, creative production, agency fees and the fact that not every platform-attributed order is purely incremental.
| Line | Amount | Running contribution |
|---|---|---|
| Net sales | £100 | £100 |
| Jewellery cost | −£28 | £72 |
| Packaging and fulfilment | −£6 | £66 |
| Payment fees | −£3 | £63 |
| Delivery support | −£4 | £59 |
| Returns and replacement allowance | −£4 | £55 before ads |
| Required contribution after ads | −£20 | £35 maximum ad cost |
| Target ROAS | £100 ÷ £35 | 2.86 / 286% |
Recalculate jewellery ROAS whenever the offer changes
A discount lowers revenue while most fulfilment costs remain. If the illustrative £100 order is reduced to £80, the £28 jewellery cost, £6 fulfilment, £4 delivery and much of the returns exposure still exist. Payment fees may fall slightly, but the money available for advertising contracts much faster than the customer price.
Suppose the £80 order contributes £36 before advertising and the business still needs £20 after ads. Maximum ad cost is now £16, so the target ROAS is £80 divided by £16: 5.0, or 500%. The campaign may convert more often while becoming much harder to acquire profitably.
Update the target for percentage discounts, free gifts, bundles, free delivery and market-specific pricing. Do not let the advertising platform optimise against an old revenue target when the current campaign economics are materially different.
| Scenario | Net sales | Contribution before ads | Required contribution | Maximum ad cost | Target ROAS |
|---|---|---|---|---|---|
| Full-price illustration | £100 | £55 | £20 | £35 | 2.86 |
| Discount illustration | £80 | £36 | £20 | £16 | 5.00 |
| Higher-AOV illustration | £140 | £75 | £25 | £50 | 2.80 |
Check jewellery purchase tracking before judging the ads
Advertising optimisation depends on the conversion events and values the website sends. Google states that target ROAS bidding requires conversion values; its system then tries to maximise conversion value while reaching the chosen average target. If purchase value, currency, transaction ID or the purchase event is wrong, the bidding system is learning from the wrong commercial outcome.
Place a real test order and trace it through the website, analytics, advertising account and Shopify order. Confirm that one order creates one purchase, uses the paid amount rather than an incorrect catalogue price, preserves currency and records the same transaction ID where the integration supports it. Check that add-to-cart, begin-checkout or page-view events are not mistakenly marked as the primary sales goal.
Test refunds and cancellations too. Some setups report the original order value indefinitely unless adjustments are imported or the business evaluates profit elsewhere. A campaign that looks stable before returns can become unprofitable after a product-specific fit or quality problem emerges.
- One completed order produces one purchase event.
- Purchase value matches the amount the business intends to optimise.
- Currency is correct for every market.
- Transaction IDs prevent avoidable duplicate purchases.
- Primary goals represent real commercial outcomes.
- Test and staff orders are excluded where appropriate.
- Refunds are included in the final profitability review.
Expect Shopify, analytics and ad platforms to disagree
A customer might discover a necklace through an advert, return through organic search, join the email list and buy after an email. Different systems can credit that order differently. Shopify explicitly notes that platforms use their own attribution rules and that more than one platform can count the same sale. Its own reports can apply last non-direct click, last click, first click, any click or linear models.
The any-click model gives full credit to every clicked channel in a journey, so credited sales can exceed the number of actual orders. Last click gives all credit to the final channel. First click favours discovery. None creates a perfect causal answer. Use a consistent primary view for decisions, then compare models to understand how sensitive the conclusion is.
Reconcile spend from the advertising account, orders from Shopify and cash or contribution from the finance view. Platform ROAS is useful for optimisation inside that platform. Blended marketing efficiency and contribution after marketing are better checks on whether the whole business improved.
| View | Likely credit | Useful for | Risk |
|---|---|---|---|
| Ad platform | Interactions within its window | Optimising campaigns on that platform | Can claim a sale another channel also influenced |
| Shopify last non-direct click | Final non-direct channel | Consistent channel comparison | Undervalues earlier discovery |
| Shopify first click | First known channel | Understanding initial acquisition | Ignores later persuasion |
| Shopify linear | Credit shared across clicks | Seeing multi-touch journeys | Still based only on observed touchpoints |
| GA4 data-driven | Modelled contribution from account data | Cross-channel analysis | Will not match ad-centric reporting |
| Finance view | Actual revenue, cost and contribution | Profitability and cash decisions | Does not by itself explain the journey |
Account for Shopify's 2026 marketing-report changes
Shopify changed the Performance by marketing channel report in September 2026 so it includes all online-store traffic grouped by channel, not only traffic attributable to a marketing activity. Shopify warns that sessions, orders and sales can therefore look higher than they did in earlier versions of the report.
It also updated referring-channel, referring-medium and traffic-type values on 18 October 2026. A date range crossing that date can show similar traffic under different labels. Those classification changes do not alter total orders or sales, but they can distort a channel trend if the business assumes the rows remained identical.
Annotate reporting changes, save the exact date range and attribution model, and avoid comparing a newly defined channel row with an old export as though the measurement were unchanged. A marketing decision needs a stable denominator as well as a current number.
Let jewellery conversion lag settle before changing the campaign
Jewellery is not always an immediate purchase. A customer may compare dimensions, delivery dates, reviews, materials, sizing, gifting suitability and returns before buying. Google Ads conversion windows can be set from one to 90 days depending on the source, and Google recommends reviewing the time-lag report when selecting the window.
Recent data is incomplete. Google advises allowing for processing delay and notes that conversion reporting can take 24 to 48 hours, with some conversions arriving later. Shopify says advertising spend from connected marketing apps can take up to 24 hours to sync. A morning comparison of yesterday's spend and sales can therefore make the newest day look artificially weak.
Choose a review rhythm that matches spend and purchase delay. Use early indicators such as qualified sessions, product views and basket additions for monitoring, but do not declare profitability from them. Finalise the commercial result only after the conversion window, cancellations and a meaningful portion of returns have matured.
Diagnose where the jewellery advertising funnel is failing
Do not respond to weak ROAS by changing every part of the campaign. Read the sequence. Impressions show delivery, CTR shows whether the creative and proposition earn attention, CPC shows what the click costs, landing-page engagement shows whether the promise continues, add-to-cart shows product confidence, checkout behaviour shows purchase friction and contribution shows whether the order is worth buying.
A high CTR with weak product engagement often means the advert is interesting but mismatched. A reasonable product-view-to-basket rate with an expensive CPC points back to audience, auction or creative efficiency. Strong baskets with weak checkout completion suggest delivery, payment or trust friction. Strong reported ROAS with weak contribution is an economics problem rather than an advert problem.
Compare by product, audience, placement, market, device and new-versus-returning customer where volume supports it. Avoid slicing a small campaign into dozens of cells; apparent winners based on one order are not reliable enough to scale.
| Symptom | Likely question | First investigation |
|---|---|---|
| Low CTR | Does the creative earn relevant attention? | Product visibility, hook, audience and placement |
| High CTR, fast exits | Does the page continue the promise? | Landing-page match, speed and mobile layout |
| Product views, few basket additions | Is the piece clear and compelling? | Imagery, scale, price, variants and delivery |
| Baskets, few purchases | What changes at checkout? | Total cost, payment, dispatch and trust |
| Orders, weak ROAS | Is click cost or order value the constraint? | CPC, conversion rate, AOV and product mix |
| Good ROAS, weak profit | Are margins and costs represented? | Contribution, discounts, returns and new-customer mix |
Keep the jewellery advert, product and landing page in one story
The page should fulfil the specific promise that earned the click. If the advert shows a gold-colour birthstone necklace under £70 with delivery before a gifting date, the landing page should show that product, selected option, current price, availability and dispatch information without making the customer reconstruct the offer.
Do not send every paid visit to the homepage. Use the closest relevant product, collection or considered campaign page. Keep the advertised item in stock, preserve tracking parameters through redirects and remove old ads when a variant, price or dispatch promise changes.
Jewellery needs recognition at a glance. The main image, title, price and variant language on the page should resemble the advert closely enough for the customer to know they arrived correctly. Then answer the next questions: scale, material, fit, fastening, personalisation, packaging, delivery and returns.
Separate cold jewellery acquisition from warm remarketing
A previous visitor, email subscriber or existing customer already knows something about the business. A cold prospect may be meeting the brand, product and price for the first time. Mixing those audiences can make the campaign look more efficient than its new-customer acquisition really is.
Report prospecting, remarketing and existing-customer activity separately. Exclude recent purchasers where the goal is new-customer acquisition, and check whether branded search campaigns are capturing demand created elsewhere. A strong blended ROAS can be dominated by people who were already likely to buy.
Remarketing still has value when it restores a considered decision with useful evidence, not endless pursuit. Use product views, basket stage and purchase history to keep the message relevant. Measure incrementality cautiously: an attributed order is not proof that the advert caused the order.
Measure jewellery CAC by new customer—not simply by order
Cost per order and customer acquisition cost are different when existing customers reorder through ads or one customer places multiple orders. For a new-customer campaign, divide the relevant acquisition cost by genuinely new customers acquired. Include direct media spend and decide whether creative, agency, app and campaign-production costs belong in the management view.
Compare CAC with first-order contribution, not first-order revenue. Then review customer cohorts by acquisition month, first product, discount and channel. Shopify's customer cohort report can show acquisition and retention patterns, including the marketing channels responsible for directing a cohort's customers.
Do not defend an unprofitable first order with a guessed lifetime value. Use observed repeat behaviour after refunds, service costs and future marketing. Jewellery purchase frequency can be slow and occasion-led. A forecast that assumes every customer returns can justify almost any acquisition cost and protect none of the cash.
Send advertising systems better jewellery value signals
Google supports conversion values and value rules so advertisers can express that some conversions are more valuable than others. It can optimise target ROAS towards reported sales revenue or profit-margin values. That does not mean every small jewellery account should immediately change bidding strategy; it means the value signal deserves the same care as the creative.
If all purchases send gross revenue, the system is rewarded equally for £100 of high-contribution sales and £100 of heavily discounted, low-contribution sales. Where the implementation and volume support it, use values that better reflect the business goal, or keep a separate profit view that governs budgets and product exclusions.
Google currently states that Search and Shopping campaigns generally need at least 15 conversions in the past 30 days at the conversion-tracking level before target ROAS can be applied. Low-volume campaigns need restraint. Stable tracking, a realistic budget and enough time can be more valuable than repeatedly changing targets before the system has evidence.
Set a jewellery advertising test budget that can answer a question
A test budget must be affordable to lose and large enough to create useful evidence. Start with the maximum target CAC, expected website conversion rate and the clicks likely to be needed for an order. If target CAC is £30 and conversion rate is 1%, the implied affordable CPC is about £0.30 before allowing for uncertainty. If the auction routinely costs far more, the economics need a different product, conversion rate, order value or acquisition route.
Give each test one primary question: creative angle, hero product, audience, landing page or offer. Hold the other elements as stable as practical. Set the spend limit, success metric, minimum observation window and stop condition before launch. Do not keep weak ads alive because the next order might repair the average.
Scale in controlled steps after performance matures. Watch marginal results, not only the historical average; the next £500 can be less efficient than the first £500. Confirm that stock, fulfilment, customer service and cash flow can support the extra volume before increasing budget.
| Decision | Set before launch | Do not substitute |
|---|---|---|
| Commercial goal | New customers, profitable orders or stock objective | More traffic |
| Target | Maximum CAC and required contribution | A generic ROAS benchmark |
| Product | Known margin, stock and delivery capacity | Whichever image is available |
| Measurement | Purchase value, spend source and attribution view | One platform screenshot |
| Window | Time for conversion lag and early returns | Yesterday's unfinished result |
| Stop condition | Spend or evidence threshold | Hope that one order changes everything |
| Scale condition | Mature profit plus operational capacity | Revenue alone |
Use a jewellery advertising scorecard that reaches profit
Build one weekly view that joins media, website and commercial outcomes. Record spend, impressions, clicks, CTR, CPC, sessions, product views, basket additions, orders, new customers, net sales, AOV, refunds, contribution before ads, CAC and contribution after ads. Keep the revenue basis and attribution view consistent.
Review products and markets separately where the data is adequate. A campaign can hit its account target by selling one strong product while another drains budget. It can also appear profitable before international delivery, currency conversion or return costs are assigned.
Add notes for creative changes, stockouts, price changes, promotions, tracking repairs and reporting changes. Numbers become more useful when the business can explain why they moved. The objective is not a perfect dashboard; it is a reliable decision about what to stop, repair, continue or scale.
- Use advertising-platform spend as the cost source.
- Use Shopify orders and refunds as the sales reality check.
- Separate new and returning customers.
- Compare platform ROAS with adjusted ROAS and contribution.
- Mark incomplete conversion and return periods.
- Record product, market, device and audience differences only when volume is credible.
A 30-day plan to make jewellery advertising profit-aware
In week one, calculate contribution before ads and maximum CAC for the main product families, offers and markets. Choose a required contribution after ads. Reconcile the last 90 days of ad spend, Shopify orders, refunds and new-customer status using one stated attribution model.
In week two, test the complete purchase tracking journey. Confirm purchase value, currency, transaction ID and primary goals. Review Shopify's current marketing-report definitions and annotate the September and 18 October 2026 changes where historical comparisons cross them.
In week three, diagnose the funnel by campaign and hero product. Repair the largest constraint: creative relevance, audience, CPC, landing-page match, product confidence, checkout or economics. Separate prospecting, remarketing and existing-customer orders.
In week four, run one bounded test with a target CAC, spend limit, observation window and stop condition. Wait for reporting delay and early refunds, then calculate adjusted ROAS and contribution after ads. Scale only if the mature result and operational capacity both support it.
- Days 1–4: calculate product-level contribution and maximum CAC.
- Days 5–7: reconcile spend, sales, refunds and customer type.
- Days 8–12: test purchase tracking and conversion values.
- Days 13–17: compare attribution models and reporting definitions.
- Days 18–22: repair the weakest funnel stage.
- Days 23–30: run and mature one controlled campaign test.
Buy jewellery customers at a cost the business can keep
Paid advertising can be a useful growth channel when the product, margin, measurement and customer journey are ready. It becomes dangerous when reported revenue is treated as retained value and a platform target replaces the business's own economics.
Know the maximum acquisition cost before the campaign starts. Send accurate purchase values, allow the result to mature, compare attribution views and bring the final decision back to contribution and cash. The aim is not the highest ROAS screenshot. It is a repeatable way to acquire the right jewellery customer while leaving enough money to serve them and grow.
Jewellery business consultant · Devon, UK · Clients worldwide