Performance marketing: measure acquisition beyond the ad platform
Performance marketing manages acquisition against defined, measurable outcomes such as qualified inquiries, purchases or subscriptions. Its value to a product depends on what happens after that event: whether the customer receives the promised value, returns and contributes enough to cover acquisition costs.
Separate the payment model from the optimization goal
Performance marketing is sometimes used narrowly for arrangements where a partner is paid for an agreed action. In everyday paid-acquisition work, it also describes campaigns optimized for outcomes even when the advertiser pays for clicks or impressions. State which meaning you are using.
A target CPA is an optimization target, not a promise that every action will cost that amount or that an unconverted click is free. Google's target CPA documentation, for example, describes setting cost-per-click bids to help achieve a target average action cost. Billing, bidding and the business outcome are related but different.
Paid search, paid social, affiliate partnerships and programmatic placements can support acquisition. Retargeting describes the audience's prior interaction, not a separate guarantee of cheap incremental customers. Choose a channel from the reachable audience, product restrictions and test economics, with a clear spending limit.
Choose an event that predicts useful customers
Define the customer journey before the conversion event. A mobile app install, a completed first task, a paid subscription and a renewal answer different questions. For a sales-led product, a form fill may be far from a qualified opportunity. Check whether a cheaper early event still brings people who reach the important later stages.
Use CPC for click cost, CPA for a precisely named action and CAC for new-customer acquisition with an explicit cost scope. ROAS divides attributed revenue by advertising spend; it does not deduct product delivery costs or prove that the advertising caused the revenue. LTV also needs a definition: revenue and gross-profit value are not interchangeable.
Reconcile platform events with product and payment records. Check duplicates, refunds, time zones, attribution windows and the delay between acquisition and purchase. Keep unobserved journeys visible. More tracking tools cannot reconstruct every missing interaction or turn attribution into a causal experiment.
A subscription example: revenue is not acquisition capacity
Suppose a hypothetical subscription product spends $5,000 on advertising and acquires 100 new paying customers. Media-only CAC is $50. If each customer produces an average of five paid months at $28 during the measured period, average observed revenue is $140.
At a 40% gross margin on that revenue, average gross profit is $56 per acquired customer. Only $6 remains after the $50 media cost, before creative production, agency or staff costs and any other expenses not included in gross margin. A revenue-to-media-cost ratio of 2.8 does not make this a comfortably profitable program.
Now inspect why customers leave and whether the acquired audience matches the offer. Better activation or retention could improve the economics, but that is a hypothesis to test. Raising a target CPA would increase the allowed acquisition cost and would not, by itself, repair the margin problem.
Connect campaign learning with product and brand decisions
A campaign can reveal a promising use case or a mismatch between the message and the product. Review acquisition quality with the product team: which audiences activate, what blocks them and whether retained customers resemble the people the ads attract. Share those findings as evidence, not as automatic instructions to change the roadmap.
Performance and brand work can overlap. A useful demonstration may generate an immediate inquiry and improve understanding of the product. Brand awareness and consideration can also be measured, although they answer different questions from purchase conversion. Neither discipline has a universal time-to-result or a guarantee that it makes the other cheaper.
Before increasing spend, check recent cohort quality, customer capacity and the cost of additional volume. Average historical CAC does not tell you what the next customers will cost. A controlled experiment or holdout can help estimate incremental impact where feasible; a platform attribution report alone cannot establish it.
FAQ
Is performance marketing always pay-per-sale?
No. The term can describe action-based partner payments, but it is also used for outcome-optimized advertising billed on clicks or impressions. Check the actual payment and bidding model.
Does a high ROAS prove profitability?
No. ROAS uses attributed revenue and ad spend. You still need product costs, the remaining acquisition expenses, refunds and a suitable observation period. Attribution also does not prove incremental sales.
How much should the first test cost?
Choose a limit from the expected event cost, the amount of evidence needed for the decision and the loss the business can afford. A small exploratory test may help refine the offer without being large enough to estimate stable acquisition economics.
Does performance marketing replace product work?
No. Advertising can help suitable people discover and evaluate an offer. It cannot supply a missing capability or ensure that customers receive lasting value after acquisition.
Sources
Discuss your product's next step
I help define product priorities, plan launches and investigate where users drop out.
