When a subscription model changes what a customer is worth, steering on ROAS becomes the wrong goal. Here is how we flipped the strategy and made the account stronger than it had ever been in peak season.

The Brand

A Dutch supplements brand that sells across multiple European countries through its Shopify store. We have managed this brand's Google Ads account from Adpact for over three years. The brand has a loyal customer base with a high share of returning buyers, and since late 2025 it runs a subscription model that lets customers sign up for automatic repeat deliveries.

It is a logical step for supplements: if you take a product daily, you do not want to reorder every month.

The Starting Point

For years we steered, like most e-commerce accounts, primarily on ROAS. We knew exactly what the break-even ROAS was and from which ROAS we turned a profit. Logical, but with the arrival of the subscription model the economics of the brand changed fundamentally. A new customer is no longer worth a single order, but a potential subscriber: new customers stay subscribed for an average of 7 months with an average order value of €45. That means an expected customer value of over €300 per acquired customer.

With that LTV, a new customer is simply allowed to cost more. Continuing to steer on ROAS would mean we systematically buy too cautiously and leave growth on the table.

Our Approach

As of 1 January 2026 we flipped the strategy: from ROAS steering to New Customer Acquisition, with CPA and the number of genuinely new customers as the leading KPIs.

1. KPI shift from ROAS to CPA on new customers

We deliberately accept a higher cost per order, because the subscription LTV more than earns it back. The question is no longer “what does this click return today?”, but “what is this customer worth in 7 months?”

2. Identifying the winning products

An important learning: not every product lends itself to new customer acquisition. A select number of items structurally bring in more new customers at lower CPCs. By analysing at product level which items attract new customers, instead of only looking at total performance, we could direct the budget where the acquisition effect is greatest.

3. Dedicated campaigns per winning product, against the consolidation trend

Current best practice prescribes consolidating as much as possible. We deliberately chose the other route: the winning acquisition products got their own dedicated campaigns. The reason? Search intent differs per item and each product has its own seasonal moments. By segmenting we can scale each product at the right moment, instead of strong products drowning in a consolidated structure.

The Results

First an important nuance for anyone who wants to understand the numbers: January is always a peak month for supplements brands. New year's resolutions push search volume and the number of new customers up artificially. This is a seasonal effect that returns every year and says little about the underlying strength of a strategy. February, once that effect has worn off, is therefore the honest starting point to measure the impact of the new approach.

Three things stand out in these numbers:

1. Structural growth of 46% in three months

From the realistic starting point in February, the number of new customers grew every month, to 1,247 in May. An increase of +46%.

2. The growth accelerates

Not only does volume rise month over month, the growth rate itself increases too: from +8.5% in March to +13.7% in April and +18.1% in May. That is exactly the pattern you want to see with this approach: as the dedicated campaigns collect data and we can steer more sharply per product, Google Ads becomes more effective every month.

3. May beats the January peak

Perhaps the strongest signal: in May we brought in 10.5% more new customers than during the January peak, a month without any seasonal advantage. The structural acquisition power of the account now sits above the level that was previously only reached in high season.

Each new customer represents not one order, but an expected subscription value of over €300. The growth in volume therefore translates directly into future, recurring revenue.

Key Takeaways

• A subscription model changes what a customer is worth, and therefore what you are allowed to pay to acquire one.

• Keep steering on ROAS in that situation and you optimise for the wrong goal.

• Not every product is an acquisition product. Analyse at product level which items actually bring in new customers.

• Sometimes the best practice is not the best choice. Segmenting at product level gave us the control that consolidation would have taken away.