The Ads Strategy Behind Growing Jewellery Brands: What We Learned in Our Webinar with Francesc Alsina

On 29 July, I hosted a live webinar with Francesc Alsina, a digital marketer with nearly 15 years in e-commerce and more than 400 brands behind him. He launched the e-commerce for PdPaola, the Barcelona jewellery brand, and today runs his own consultancy while collaborating with Common Thread Collective, a US growth agency.

A brand doing £4 million a year that couldn't tell new revenue from returning revenue

One story from the session stuck with everyone who attended. A jewellery brand came to Francesc making around £4 million a year, but their growth had started to slow. When he looked under the bonnet, he found they had no proper e-commerce report at all.

They were reading Meta's dashboard and Shopify's revenue total, and treating both as one number. On paper, the return on ad spend and cost of acquisition looked healthy. In reality, acquiring a new customer had become three to five times more expensive than the year before. The blended figures looked fine only because the brand happened to be good at retention, without knowing it, or measuring it.

That is the whole problem in one example. Most jewellery brands judge their advertising by the wrong numbers, and the wrong numbers tell a flattering story right up until growth stalls.

This article pulls the substantive parts of that 90-minute session into something you can act on.

What the webinar actually covered

The session ran through eight connected ideas about how paid advertising, business performance, and brand growth fit together for a jewellery brand. Not eight separate tactics. One system. The parts that matter most for a brand owner are below.

How Meta decides who sees your ad

Every time someone opens Instagram or Facebook, Meta runs a real-time auction for that ad slot. It is not simply whoever pays the most.

Meta combines three inputs:

• Your bid — what you're willing to pay.

• Estimated action rate — how likely that specific person is to complete the action you're optimising for, such as a purchase.

• Ad quality — Meta's relevance and experience score for your creative.

The ad with the highest combined value wins the impression. This is why two jewellery brands can spend the same budget and get completely different results. The one with stronger creative and cleaner data signals wins more useful impressions for less money.

One practical detail worth knowing: Meta reserves a large share of ad inventory for video. If you're running only static images, you're competing for a smaller pool of placements.

Why patience is mechanical, not optional

When you launch a new campaign, or make a significant edit to a live one, that ad set re-enters what Meta calls the learning phase. During roughly seven days, Meta tests placements, audiences, and times of day to work out where your ad performs best.

As a rough guide, Meta wants around 50 conversion events per ad set per week to exit the learning phase efficiently. Until then, cost per result bounces around. Some days look great, some look terrible. That volatility is the algorithm working, not the campaign failing.

The most common mistake Francesc sees is founders panicking after 48 hours, pausing or editing the campaign, and resetting the learning phase over and over. His rule of thumb:

• Wait at least three days before small tweaks.

• Wait at least seven days before bigger changes.

• If you can wait a full month, better still.

And when you doubled a working campaign's budget from £100 to £200 overnight and it stopped performing? You reset the learning phase. To scale a live campaign, increase the budget by around 19% at a time, or duplicate the campaign and give the copy a larger budget.

Why targeting is no longer your main lever

For years, interest targeting and lookalike audiences were where founders spent their attention. Francesc's position is that this changed, with Meta's fuller rollout of its AI advertising systems (known internally as Andromeda and GEM) around October 2024.

Meta's models now predict buying behaviour across the whole platform well enough that, in most jewellery accounts, broad targeting paired with strong data signals performs as well as, or better than, hand-picked audiences. The algorithm has more data on who is likely to buy a gold necklace this week than you ever will.

So your leverage has moved. The things you actually control now are:

• The quality and variety of your creative, so the algorithm has different angles to find different buyers.

• The cleanliness of your data signals — a well-implemented pixel and Conversions API.

• The stability of your budget and structure, so the algorithm isn't constantly relearning.

  
  

Brands still trying to out-target the algorithm tend to be the ones underperforming.

The metrics jewellery founders actually need to read

Most founders live inside Ads Manager, watching advertising metrics: click-through rate, cost per click, cost per purchase, and return on ad spend (ROAS). Those tell you how an ad performs inside the platform. They don't tell you whether the business is healthy.

For that, you need business metrics: average order value, gross margin, contribution margin, lifetime value, and repeat purchase rate.

Here is why reading only ROAS is a trap. A 3x ROAS on a high-margin fine jewellery piece can be very profitable. The same 3x ROAS on a heavily discounted, lower-margin line might barely break even, or lose money, once you count cost of goods, packaging, shipping, returns, and payment fees. Jewellery carries a real-world cost that ROAS never sees: rings, in particular, generate returns and resizing because of sizing issues.

Francesc's habit is to read blended metrics as a trend line over several weeks, not to react to a single day's platform number.

Metric	What it measures
ROAS (in-platform)	Revenue Meta attributes to its own ads
Blended ROAS / MER	Total revenue ÷ total ad spend across all channels
aMER	Marketing efficiency ratio for acquisition (new customers only)
CAC vs LTV	Acquisition cost judged against lifetime value, not in isolation

A CAC of £80 sounds high on its own. If your average customer's lifetime value is £400, it's an excellent number. Judging the first purchase in isolation is the error.

The one thing to take away: separate new revenue from returning revenue

If you remember a single idea from the session, this is it. Your revenue comes from two engines:

1. New customer revenue (growth by acquisition). This is what Meta ads are excellent at. It is also the most expensive way to grow, because you pay for attention every single time.

2. Returning customer revenue (growth by retention). This comes from product experience and channels you already own — email, WhatsApp, SMS, loyalty. Getting an existing customer to buy again is widely estimated to cost five to seven times less than acquiring a new one, and it compounds over the life of the brand.

Brands that scale sustainably run both engines and track them separately. Every pound of lifetime value you build through retention raises the price you can afford to pay for a new customer, which lets you bid more aggressively than competitors who are only judging the first sale.

The gap Francesc sees most often: jewellery brands pour the whole budget into acquisition and have a thin or non-existent email and WhatsApp flow structure. The natural repeat opportunity simply evaporates. His advice is to build those automations — welcome series, post-purchase, care flows, gifting reminders, win-back — from day zero, even before returning revenue arrives. Klaviyo is his recommended platform for this on Shopify.

Your strategy has to change with your stage

There is no single correct approach. It depends on where the brand is.

• Early stage — validate and learn. The goal isn't scale, it's proof of concept: what price point, what story, what photography style actually converts. Concentrate spend on a small number of hero products rather than spreading a thin budget across the whole catalogue. Prioritise testing creative angles over testing audiences, and expect volatility.

• Scaling stage — build efficiency. Once you know what works, the question shifts from "does this work?" to "can I trust this as I add budget?" You layer in retargeting and retention campaigns, and creative becomes a pipeline, not a one-off. Watching ROAS alone is dangerous here, because margin can erode underneath a stable-looking number.

• Rocket stage — compound growth. At maturity you optimise for brand equity and lifetime value. A large share of revenue can and should come from people who already know you. Diversifying beyond Meta — Google, affiliates, organic, partnerships — reduces dependency on one algorithm. Francesc has seen mature brands where 60–80% of revenue comes from returning customers.

The four layers of business growth

Francesc frames the whole thing as four layers stacked on each other:

1. Customer acquisition (Meta and Google ads) — the most visible, easiest to measure, and most unpredictable source of revenue.

2. Owned audiences (email, WhatsApp, SMS, organic) — a more reliable, more forecastable stream that stabilises acquisition.

3. Business performance (margin, cash flow, operations) — this decides whether all that activity turns into a fundable business.

4. Long-term brand growth (retention, brand equity, word of mouth) — this compounds over years and lowers acquisition cost in layer one.

His point: most growth problems in jewellery brands are layer two or layer three problems wearing a layer one disguise. "Fix the ads" rarely fixes anything on its own.

Six mistakes that repeatedly cap growth

1. Judging campaigns too early — editing on one or two days of learning-phase data.

2. Chasing ROAS or CAC instead of profit — scaling on a good-looking number without checking the margin.

3. Ignoring returning revenue — all budget into acquisition, no retention system behind it.

4. Under-investing in creative variety — the same one or two ads for a month, then blaming the algorithm for fatigue.

5. Scaling budget too fast — big jumps that reset the learning phase before the account has stabilised.

6. No system connecting channels — Meta running in isolation with nothing to catch and convert traffic into repeat revenue.

Individually, each is recoverable. Together, they quietly cap how big the brand can grow.

Where Chocianaite fits

Ads move faster and cost less when the creative feeding them is strong and varied. That is the layer where we work. Chocianaite produces jewellery photography and creative — product, model, and content built specifically for a jewellery audience — so brands have the steady pipeline of angles a scaling account needs, rather than one or two ads a month running to fatigue. During the webinar, I mentioned we're already collaborating with Francesc's team on ads, Klaviyo, and Shopify development for shared clients: their performance expertise paired with our content is the combination we keep coming back to.

If you'd rather learn the thinking yourself first, the free Ultimate Jeweller's Playbook is the place to start.

Why act on this now

The single most damaging habit in the whole session was measuring the wrong things and finding out too late. The £4 million brand didn't have a stock problem or a creative problem when they arrived — they had a measurement problem that hid a widening acquisition cost for a full year. Francesc's homework for attendees was blunt: work out your blended MER and margin today, map your new-versus-returning revenue split for the last 90 days, and place your brand honestly at early, scaling, or rocket stage. You cannot fix what you haven't separated.

Get the full picture

We covered far more in the live session than fits here — campaign structure, marketing-calendar moments, loyalty tiers, and the detail behind each layer. The full breakdown, plus recordings of sessions like this one, lives inside our Skool community, where jewellery brand owners can ask questions and go deeper. That is where the working conversation continues.