Case Study · Ecommerce
Five years active and scaling. $30K per month to $120K per month. Advertising that was producing net losses turned into $156K per month in sustained profit. 130% profit ROAS means every $1 spent returns $1.30 in profit after product cost and financing commissions.
A national ecommerce retailer offering Buy Now Pay Later financing on automotive products came to LGG Media in 2021 after abandoning paid advertising entirely. Every campaign attempt had produced a net loss and the business could not diagnose why.
The core problem was measurement. They had no way to calculate profit per transaction, so their campaigns had been optimizing on transaction revenue instead. Revenue and profit are not interchangeable signals in this business.
The retailer generates income from two sources on every transaction: product margin from the sale, and a financing commission from the BNPL partner. Combined, the two numbers produce actual profit per transaction. Neither number alone is sufficient.
The core failure: the campaigns ran at a net loss because the algorithm was maximizing revenue, not profit. Revenue looked acceptable. Profit was negative. Optimizing on revenue finds the wrong transactions.
Revenue-based optimization in a dual-profit-source business produces the wrong answer on every transaction where margin and revenue do not move in the same direction.
A mid-tier automotive product with a 40% product margin and a high-commission BNPL partner can generate $350 in combined profit on a $600 transaction.
A premium product at the same revenue level with a 15% margin and a low-commission lender generates $140 in combined profit on the same $600.
The algorithm targeting revenue sees two identical $600 transactions. The business experiences a $210 profit difference between them. The signal was telling the algorithm to find more $600 transactions. The business needed it to find more $350-profit transactions.
Without the combined profit figure flowing back to the algorithm, there was no mechanism to communicate the difference. The algorithm was performing exactly as instructed. The instruction was wrong. Fixing the outcome required fixing the input.
LGG Media built a custom integration connecting the retailer's ERP system to the ad platforms through a proprietary data management layer. The ERP calculates combined profit per transaction in real time: product margin plus financing partner commission.
LGG Media built the API connection that pulls that profit figure from the ERP the moment a transaction completes, passes it through a deduplication and data hygiene layer that removes duplicate events and validates the profit calculation against the source record, and delivers the verified profit amount to the ad platforms as the conversion value.
The deduplication and hygiene layer is a non-trivial component. In a BNPL business, financing confirmations and product fulfillment events can fire at different times from different systems. Without deduplication, the same transaction can send multiple conversion signals with different profit values depending on which system fires first.
The hygiene layer holds the conversion event until both the product margin and the financing commission are confirmed, then sends a single verified profit figure. Attribution accuracy depends on that sequencing being correct on every transaction. For the tracking mechanics behind this, see how attribution infrastructure works →
What changed: the payload sent to Google and Meta now contained the actual combined profit amount, not transaction revenue. A $220-profit transaction sent $220. A $350-profit transaction sent $350. The algorithm stopped pursuing high-revenue traffic patterns and started pursuing high-profit ones.
LGG Media set a target profit ROAS of 130% across both Google and Meta. The target was set at 130% rather than a higher rate because this business is volume-dependent.
Fixed operational costs, warehouse, logistics, customer service staffing, are spread across all transactions regardless of volume. At 710 transactions per month, fixed costs represent a smaller proportion of each transaction's economics than they would at 300 transactions per month.
A 130% profit ROAS at 710 monthly transactions produces more total profit than a 150% profit ROAS at 400 monthly transactions once fixed cost dilution is accounted for. Setting the target too high would restrict volume below the threshold where fixed cost dilution makes each transaction more profitable. Setting it too low would erode the per-transaction margin.
Both Google and Meta hold 130% profit ROAS at the same target simultaneously. Equal performance across both channels validates the profit signal rather than making results channel-dependent. Each platform's algorithm receives the same conversion values and reaches the same efficiency independently.
In Q1 2023, when profit-based bidding went live across both platforms, the account turned profitable for the first time in the retailer's history of running paid advertising. It has held at 130% profit ROAS since, more than two years of consistent performance across spend levels that grew from $30K to $120K per month.
The account currently generates approximately 710 transactions per month at an average combined profit of $220 per transaction and a customer acquisition cost of approximately $170. The $50 difference between the $220 average profit and the $170 CAC is the per-transaction net margin before fixed costs.
At 710 monthly transactions, total advertising-attributed profit reaches $156K per month. Monthly ad spend has grown from $30K at engagement start to $120K at current scale, a 4x increase over five years with profit ROAS held constant. The account has not dropped below 130% profit ROAS in any month since Q1 2023.
The five-year engagement duration reflects the consistency of the attribution infrastructure rather than a dependency on any single channel or campaign structure.
The profit signal has remained accurate as the product catalog, financing partners, and platform algorithms have all changed over that period. Because the ERP calculates combined profit per transaction at the source, the conversion value stays correct even when margins, lenders, or commission rates shift.
That is why the account has never dropped below 130% profit ROAS since Q1 2023. The infrastructure measures the right number on every transaction, and it keeps measuring it correctly as the inputs move.
Explore the approach: Pay-per-click management → · Attribution tracking → · All client results → We run the same profit-based approach for other online stores through our ecommerce PPC management. See how the underlying tracking and bidding work in our guides to Google Ads offline conversion tracking and target ROAS.
Next Step
If revenue looks acceptable but profit is inconsistent across products, the conversion signal reaching the algorithm is optimizing on the wrong number.
Request a Discovery Call →