Black Friday-Ready Growth Through Tracking, CRO & Paid Media Overhaul – Fitness Apparel Case Study

Midsummer Agency approached the project as a full-funnel recovery and scale initiative, rebuilding the account across media, measurement, feed infrastructure, website experience, and profitability.

1. Rebuilding the tracking and technical foundation

The first priority was regaining control of the data. We audited the setup across Google Ads, Merchant Center, GA4, GTM, Meta Ads, and Shopify, to identify what was compromising attribution and campaign efficiency.

To strengthen measurement quality, we implemented server-side tracking through Stape.io, including a custom loader, Cookie Keeper, and a multi-domain setup to improve data continuity across domains. We also implemented Stape’s bot detection solution to reduce invalid traffic and improve the quality of the signals being sent back to advertising platforms.

At the same time, we supported the brand during a bot attack by helping set up Cloudflare for Shopify, adding another layer of protection during a commercially important period.

2. Turning Google Ads and the feed into stronger growth levers

From a Google Ads and feed management perspective, the work went far beyond campaign maintenance. We reviewed account structure, assessed branded and non-branded activity, identified missed product coverage opportunities, and improved the way the catalog supported Shopping and Performance Max.

One of the biggest unlocks was expanding the catalog beyond its previous limitations so campaigns could show a much broader range of products, sizes, and variants. 

Alongside this, we corrected foundational feed issues: faulty product titles, HTML characters in descriptions, incorrect google_product_category assignments, and broken product_type structures, all of which were limiting how Google could index and surface the catalog.

We then rebuilt the product titles with search behavior in mind for both markets. In Sweden, this meant replacing underperforming English terms with higher-volume Swedish equivalents, swapping terms like “sweatshirt” for “tröja” and “jacket” for “jacka”, in some cases unlocking search volumes more than 5–10x higher than the original terms, with edge cases kept where the data justified it. 

In Norway, color attributes were translated from English and mapped based on actual search volume, choosing “lyseblå” over “light blue”, or keeping simpler terms like “brun” over obscure equivalents. We also fixed the material attribute, which had been comma-separated and unreadable by Google, reformatting it to slash-separated so multiple materials could be properly indexed, handled through custom regexp rules built directly in the feed management tool.

We also laid the groundwork for smarter segmentation through custom labels and product-bucketing logic, helping turn the feed from a basic catalog export into a more strategic performance asset.

3. Building a more scalable Meta Ads system

On Meta Ads, the goal was not simply to run campaigns, but to build a more scalable creative testing system around the brand’s strongest product stories and buying triggers.

The foundation of this approach was an ABO (Ad Set Budget Optimization) structure. We chose ABO deliberately over a standard CBO setup: with a single testing campaign and centralized budget, Meta’s algorithm would have naturally favored whichever product angle performed better, leaving us with no clean data on the underperforming ad sets and no ability to scale them independently. By managing budget at the ad set level instead, we guaranteed equal learning opportunity across every angle, isolated performance signals per product category, and kept the data clean enough to make confident decisions.

We structured the account around clear creative angles for cold traffic, separating messaging by product category and USP rather than relying on generic setups.

For tights, this meant angles around scrunch expertise, fit, shape enhancement, and fabric quality, with specific creative concepts such as “Scrunch Done Right” (focusing on stitching and fabric construction) and “Scrunch Creator” (a single squat demonstrating how the fabric stretches and holds the scrunch).

For jeans, the focus shifted to curve-friendly fit, comfort, stretch, and “real denim” positioning — with angles like “Try-On Proof,” “Built for Curves,” “No Waist Gap,” and “All-Day Wear.” For hero products, we leaned into back-in-stock demand, comfort, styling, and momentum-driven messaging with concepts built around scarcity (“Sold Out Instantly,” “Back in Stock”). Each ad set contained three creatives mapped to a single angle, targeting 20 conversions per creative as the threshold for statistical significance.

We ran fully broad targeting with no age or gender restrictions to minimize CPMs without sacrificing reach into the core audience, excluded past purchasers, and set the attribution window to 7-day click + 1-day view to accelerate the algorithm’s learning phase.

Following the successful launch, we implemented multiple 20% budget increases per ad set to capitalize on high-performing angles. We also ran a URL test (collection page versus product page) for creatives where a single product was prominently featured, adding a layer of landing page optimization to the creative testing framework.

We paired this with a mix of UGC, influencer content, and in-house assets to create a more deliberate prospecting and retargeting framework. We also solved an important execution bottleneck: resizing creatives across multiple placements had become highly time-consuming for the client. Using a proprietary AI solution, we automated creative resizing across formats, making testing and production much faster and easier to scale.

4. Creating a CRO roadmap around the real friction points

Because better traffic alone would not solve the underlying performance ceiling, we also built a CRO roadmap focused on improving the efficiency of the traffic already being generated.

The roadmap took a mobile-first approach, since mobile was driving the vast majority of traffic and revenue. It also reflected two different conversion problems by market: in Norway, the main weakness was deeper in the funnel, especially around checkout progression, while in Sweden the friction appeared earlier, around product discovery, persuasion, and movement into checkout.

Based on that, we prioritized the highest-impact improvements across the shopping journey: sticky navigation, more visible add-to-cart actions, clearer Klarna and payment method visibility, stronger reviews and social proof, better fit guidance and size charts, improved filtering and sorting, quick-buy opportunities, stronger category navigation, and more supportive cart and checkout messaging. The result was a roadmap designed to improve both conversion rate and average order value over time.

5. Restoring profitability with clearer economics

Finally, we addressed the commercial issue behind the media performance. The brand was no longer operating profitably, so we installed and configured Lifetimely to identify breakeven sales levels and give the team a clearer view of actual business economics.

This made it easier to evaluate performance beyond platform ROAS alone, improved decision-making around scale, and helped steer the brand back toward profitable growth.

PPC:

• Google Ads management
• Feed optimization and management

Paid Social:

• Meta Ads management

CRO

Web Analytics

• Tracking and measurement troubleshooting
• Lifetimely setup and profitability analysis
• Shopify support