Meet the Client Partner: A Fashion Retailer Building Its Customer Base Across Europe

Our Client Partner is the EMEA arm of a global fashion and lifestyle retailer. The business sells apparel, accessories and home products to a young, style-led audience, and it competes in a category where acquisition costs climb every peak season and loyalty determines profitability.

The account already ran a mature Performance Max setup. Search, shopping and discovery inventory worked together, and the feed was in good shape. Returning customers and existing site visitors, however, absorbed a large share of the conversions. That pattern flattered the reported return on ad spend while leaving the customer file thinner than the business wanted.

The Growth Challenge: Reaching First-Time Shoppers Without Losing Control of Efficiency

Performance Max optimizes toward conversion value, and it finds the cheapest conversions available. In a retail account with a strong brand and a large base of past buyers, the cheapest conversions often come from people who already know the brand. The campaign looks efficient, yet it adds little incremental revenue. A retailer that depends on repeat purchase for margin needs a steady inflow of new buyers, and a bidding system that favors existing customers works against that need.

We set out to change the bidding signal. Our Client Partner wanted to shift investment toward first-time shoppers and accept a higher cost to win them, provided the economics held up over the customer lifetime. We agreed on the following success conditions:

1. Increase the share of orders from new customers within Performance Max.

2. Grow qualified traffic from audiences that have not yet bought.

3. Keep the cost per new customer within a range that lifetime value supports.

4. Protect audience quality so that added reach does not dilute spend.

These conditions treat a higher cost per click as an expected outcome and make the cost per new customer the primary guardrail. We built the test around that logic.

Our Strategy: Bidding Higher for New Customers and Tightening the Audience

We compared two periods. The baseline ran from February 8 to June 10, 2025, and the test period ran from June 11 to October 11, 2025. Within the test period, we changed three things.

Activating New Customer Value Mode in Performance Max

We switched on the new customer optimization feature in Performance Max, which instructs the system to bid higher when it identifies a user as new to the brand. Instead of treating every conversion equally, the campaign assigns extra value to a first purchase. That change moves the auction logic toward the outcome the business wants, which is a larger pool of customers who can be retained and monetized later.

We expected this to raise click costs. A system that bids more aggressively for unfamiliar audiences competes in pricier auctions, and it reaches people with less purchase history. We told our Client Partner this before launch, which let us judge the results against the right benchmark.

Excluding Audiences That Weaken New Customer Quality

New customer mode improves reach, but it cannot judge whether a new user fits the brand's core buyer. We reviewed performance by age band and excluded the 55+ segment. This step raised the share of new customers within spend, because budget stopped flowing to a segment that converted poorly for this assortment. The exclusion sharpened audience quality and supported spend efficiency.

Setting a Learning Window and a Cost Guardrail

Smart bidding needs time to recalibrate after a change in value signals. We allowed a learning period of four to six weeks and avoided reacting to early volatility. We then monitored cost per new customer against lifetime value, so that bids stayed aligned with what a first-time buyer is worth over time.

We also considered timing. Activating the feature before peak trading gives the system a head start on building the customer base, so we recommended it as a pre-peak lever. For retailers with strong retention, we see a case for running it year-round as a standing growth setup.

Results: More New Customers at a Higher but Controlled Cost

The test delivered the shift in mix we targeted, and the cost increases landed where we forecast them. Figures compare the test period with the baseline period.

New customer acquisition

- New customer share increased by 8%.

- Traffic increased by 15%.

- New customer orders showed an uplift by 10%.  

Engagement

- Click-through rate increased by 25%.

Cost

- Cost per click rose by 10%.

- Cost per new customer rose by 15%.

Audience quality

- The new customer percentage improved after we applied the 55+ age exclusion.

The cost movements deserve a close read. A 10% rise in click cost and a 15% rise in cost per new customer reflect a campaign that now bids into more expensive auctions for people who have not bought before. Those increases came alongside a larger and better-composed new customer share, which is the trade the strategy intended. Whether that trade pays back depends on lifetime value, so we recommend that our Client Partner track cohort revenue from these new buyers over the coming quarters to confirm the return in EBITDA terms.

One caveat applies to the reading of these results. The two periods differ in length and seasonality, so some of the movement may reflect trading conditions rather than the feature alone. A holdout or geo-based incrementality test would isolate the feature's effect more cleanly, and we see that as the logical next step.

Lessons Learned: What Senior Leaders Can Apply to Their Own Accounts

1. Change the signal before you change the budget

Performance Max follows the value it receives. When a first purchase carries the same value as a repeat purchase, the system favors whichever is cheaper to win. New customer value mode gives marketing leaders a direct way to correct that.

2. Plan for higher unit costs and judge them against lifetime value

A rise in cost per click and cost per new customer is a predictable result of reaching unfamiliar audiences. Teams that set the guardrail at lifetime value avoid cutting a strategy that is working.

3. Allow four to six weeks of learning

Early weeks produce noisy data. Stable evaluation needs a full learning window, and early intervention resets the system's progress.

4. Prune the audience while you expand it

Broader reach raises the chance of wasted spend in weak segments. Age exclusions, such as the 55+ segment here, improved the quality of new customer volume.

5. Time activation around peak

Switching the feature on before peak trading builds the customer base when demand is highest. Retailers with strong retention economics can evaluate a year-round setup.

6. Back the result with incrementality testing

Platform-reported new customer share shows direction. A controlled test shows how much revenue the strategy created that would not have occurred otherwise, and that evidence supports budget decisions at the C-suite level.

---

Would a new customer bid strategy strengthen your next trading peak? Reach out to us.

Relevant insights

· Case study: The Role of Google’s Demand Gen in the Customer Journey

· Report: The State of Search 2026 and Beyond: How AI, Automation, and Commerce Are Reshaping Discovery

· Video: How Can Predictive CLV Improve Bidding Strategies in Paid Search?

About Crealytics

Crealytics is an award-winning full-funnel digital marketing agency fueling the profitable growth of over 100 well-known B2C and B2B businesses, including ASOS, The Hut Group, Staples and Urban Outfitters. A global company with an inclusive team of 100+ international employees, we operate from our hubs in Berlin, New York, Chicago, London, and Mumbai.