Case study / Lolla's
Fashion e-commerce / Performance growth
Turning scattered winners into a commercial growth system
Lolla's was already running paid acquisition and testing different formats and creatives. The problem was not a lack of activity. The useful signals were scattered across products, locations, pages, marketing angles and the customer journey. Nour's role was to find those signals, consolidate the winners and turn them into a system that could scale more intelligently.
Brand context
Egyptian women's fashion e-commerce brand.
Verified outcome
EGP 1.6M
Gross sales
525
Orders
8x
ROAS
3 months
September to November 2024
Published with figures. The gross sales, orders, return and timeframe below belong to this engagement only and do not guarantee future outcomes.
The decision in one view
- 01
Signal
Winners existed, but they were fragmented across products, governorates, pages, marketing angles and the customer journey.
- 02
Diagnosis
This was not a lack-of-testing problem. It was a fragmented-learning and consolidation problem.
- 03
Decision
Consolidate the existing winners first, then build acquisition and the on-site experience around them before adding media weight.
The situation
Lolla's was selling around EGP 200K per month, and it was hard work to get there. The account was not idle. Multiple ad types and creatives had already been tried, so the store had history, spend and results on record.
That history is the reason this engagement did not start with a new campaign plan. It started with what had already happened.
The problem was not a lack of advertising activity. It was fragmented learning.
The signal
The account had produced winners, but they were scattered. Some products performed better than others. Some governorates converted better than others. Some pages behaved differently from the rest of the store. Some marketing angles resonated more than others, and different stages of the customer journey responded to different messages.
Each of those was a real signal. None of them was being used together.
My diagnosis
The business did not have a testing problem. It had a consolidation problem. Learning was being generated and then left behind in individual campaigns instead of being carried forward into the way the store sold.
Scaling spend on top of that structure would have scaled the fragmentation with it.
Before testing more, consolidate what the market has already told us.
What I ruled out
I ruled out starting another broad testing cycle, since the account had already paid for the answers. I ruled out treating Ads Manager reported ROAS as the decision metric on its own. I ruled out scaling budget before the winning signals were combined into something the store could actually convert on.
The decision
Filter the existing performance first, isolate the recurring winners, and build the acquisition and on-site experience around them. Only then increase the weight of media behind that system.
The system I built
- 01Step
Winner filtering
Existing performance was filtered across products, governorates, pages, marketing angles and stages of the customer journey, so the recurring winners could be separated from one-off results.
- 02Step
Winner consolidation
The winners were combined rather than run as separate experiments, so a winning product, a winning region and a winning angle could reinforce each other instead of competing for budget.
- 03Step
On-site merchandising
Frequently Bought Together was implemented, which increased average order value, and product merchandising badges such as Best Seller were added so the store itself reflected what the demand data already showed.
- 04Step
Campaign execution
Execution shifted increasingly toward Catalog and Flex formats, which suited a catalog with identified winners better than one-off creative pushes.
- 05Step
Destination logic
Traffic was sent either to a relevant collection page or directly to a product page, decided by context rather than by a single default landing rule.
Measurement and decision infrastructure
Meta Pixel carried the acquisition and conversion signal inside paid media. GA4 covered on-site behavior, traffic quality and journey analysis. Shopify was the commercial source of truth for orders and gross sales.
Decisions were not based on Ads Manager reported ROAS alone.
Acquisition signal
Meta Pixel
Acquisition and conversion signals inside paid media.
Behavior signal
GA4
On-site behavior, traffic quality and journey analysis.
Commercial truth
Shopify
Orders and gross sales used as the commercial source of truth.
Final commercial validation
I do not let one platform define the truth. I look for alignment between acquisition signals, user behavior and commercial outcomes before I scale.
The operating loop
The engagement ran as a loop rather than a launch. Every cycle fed the next one, and the store, the campaigns and the measurement layer moved together.
- 01Filter
- 02Find the signal
- 03Combine winners
- 04Build
- 05Send users to the right destination
- 06Measure behavior and sales
- 07Feed the learning back into the system
The last step feeds the next filtering cycle.
The outcome
Across September, October and November 2024 the store produced EGP 1.6M in gross sales from 525 orders at 8x ROAS, on approximately EGP 200K of total media budget.
Average order value increased after Frequently Bought Together was implemented. The exact uplift is not published here because it is not verified.
Documentary evidence
Shopify performance snapshot
EGP 1.6M
Gross sales
525
Orders
8x
ROAS
3 months
September to November 2024
Approx. EGP 200K total media budget
What this case says about how I work
The growth here did not come from spending into a new idea. It came from reading what the account had already proven, consolidating it, and rebuilding the acquisition and on-site experience around those winners before increasing media weight.
More execution is not always the answer. Sometimes the business already has the answer buried inside its existing signals. I rarely begin with 'What campaign should we launch next?' I prefer to begin with 'What is the system already telling us, and what should that change in the way we grow?'
Evidence summary: EGP 1.6M Gross sales · 525 Orders · 8x ROAS · 3 months September to November 2024