What performance marketing is and is not
The defining feature of performance marketing is the unit of judgement. A campaign is not assessed on how many people it reached or how it looked but on what each result cost and whether that cost leaves a margin. Cost per acquisition (CPACost of each conversion.) and return on ad spend (ROASRevenue generated per unit of ad spend.) are the two summary measures, and the whole practice is the management of the chain that produces them: budget buys impressions, a share become clicks, a share of clicks become results, and results are worth something. The Performance Marketing Association describes the industry around paying for outcomes rather than exposure, and the IAB's measurement standards exist so that the outcomes different sellers report can be compared.
Three things performance marketing is not. It is not a channel: search, social, display, video, affiliate and app networks are all run as performance channels when the buying is tied to results, and all of them can be run as brand channels when it is not. It is not a synonym for digital marketing, which also contains organic search, content, email and the rest. And it is not the opposite of brand building, though the two are often set against each other in budget meetings; the section on performance and brand below explains why they are complementary.
The most common misunderstanding is to take a platform's reported return as the result. A platform reports the conversions it touched under its own rules. The business's result is the sales it would not otherwise have had, net of the cost, which is a different and smaller number. Holding those two apart is the single habit that separates experienced performance marketers from new ones, and most of the guides below return to it.
The performance system
Performance marketing is a system of linked decisions, and problems in one stage usually show up as symptoms in another. The table lists each stage, the question it answers and the guide that covers it.
| Stage | The question it answers | Guide |
|---|---|---|
| Economics | What can a result cost, and how is that target derived from margin and lifetime value? | Performance economics |
| Structure | How are campaigns split so that budgets, targets and reports map to real decisions? | Campaign structure across channels |
| Conversion events | Which events and values does each platform optimise towards, and do they mean the same everywhere? | Conversion events across platforms |
| Channel roles | What is each channel for, and which platform guide covers its mechanics? | Channel playbooks |
| Planning | How is the budget split across channels by role, with targets and a test reserve? | Cross-channel performance planning |
| Audiences | When does remarketing earn its budget, and when is it only claiming credit? | Prospecting vs remarketing, Customer lists across ad platforms |
| Testing and scaling | What to test, in what order, and how to grow spend without efficiency collapsing | Testing and scaling |
| Budget | How does spend follow results inside limits the business has agreed? | Target-led paid media budget, Restoring spend after a cut |
| Reporting | What is reviewed daily, weekly and monthly, and how platform numbers are reconciled | Performance reporting |
| Proof | Did the spend cause the results, or would they have happened anyway? | Incrementality testing |
Read the stages in order the first time. Economics before structure, because the targets decide how finely campaigns need to be split; structure before planning, because a plan allocates budget to the units structure creates; testing before scaling, because scaling what has not been tested spends more to learn the same lesson. The media planning framework behind the planning row and the experimentation loop behind the testing row are both general tools that this hub applies to paid acquisition in particular.
Fundamentals: economics and structure
The fundamentals hub covers what every performance channel shares before any platform is opened. The first foundation is economics: how spend becomes revenue and then margin, why targets must come from the business's unit economics rather than from what a platform suggests, and how a blended view of acquisition cost (CACCost to acquire one new customer.) relates to the channel and campaign views that platforms report. Customer lifetime value (LTVTotal value a customer brings over time.) sets the ceiling on what a customer may cost; payback decides how much of that ceiling the business can afford to wait for. The performance economics guide walks through the chain, and the ROAS calculator and CPA calculator turn its rates into targets.
The second foundation is structure. Every platform organises spend into some hierarchy of campaigns, groups and ads, and every platform's documentation offers its own advice on how to arrange them. Underneath the differences the principle is the same: split where a decision has to be made differently, and consolidate everywhere else so that automated bidding has enough data to learn from. The campaign structure guide sets out that principle and the naming and tracking conventions that make a structure reportable, and hands the platform-specific detail to the Google Ads, Meta Ads and programmatic roots.
Further guides sit beside those two foundations. The conversion events guide covers the events and values each platform optimises towards, and the ad auctions page explains how search, social and programmatic auctions decide what a buyer pays and which parts of that price the buyer controls.
Channel playbooks
Channels have roles, not rankings. Search captures demand that already exists and is bounded by how much of it there is. Paid social creates demand among people who were not looking and is bounded by creative and audience size. Display and video buy reach and frequency at low cost per impression and are the most useful when the job is to be remembered rather than clicked. Affiliate and partner channels pay on results and borrow other people's audiences. The channel playbooks hub describes each role and points to the platform roots for mechanics: Google Ads for search, Shopping, YouTube and the Google display network; Meta Ads for Facebook and Instagram; and programmatic for display and video bought through demand-side platforms.
The core guides beneath the hub are about the decisions between channels rather than within them. The cross-channel performance planning guide turns roles into a budget with targets per channel, a flighting calendar and a reserve for tests. The prospecting vs remarketing comparison addresses the most persistent allocation error in the field: a remarketing layer that reports superb efficiency because it reaches people who were going to buy anyway, while the prospecting that fed it is starved. A further guide, using customer lists across ad platforms, runs the business's customer data as one programme across platforms: one definition of a customer, consent, segments that exclude, seed or re-engage, refreshes and match rates. Another further guide, stopping a paid advertising channel, covers the evidence an exit needs, what leans on the channel, the wind-down and what to keep for a return. Where an agency, consultant or freelancer is hired to run paid media across platforms, the guide to hiring and evaluating a performance marketing agency explains what stays in the business's name and how to weigh proposals.
Optimisation: testing, scaling, reporting
Once campaigns run, the work becomes a loop. Form a hypothesis about a lever, test it fairly, read the result without fooling yourself, roll out what won, and go round again. The optimisation hub describes that loop and the two levers that move results most on almost every account: creative and the landing page. The testing and scaling guide orders the levers by expected effect and cost, explains how to design a fair test when platforms overlap, and sets out the four scaling paths, with the guardrails that catch diminishing returns before they become losses.
Reporting is the other half of the loop. The performance reporting guide builds a cadence of daily checks, weekly optimisation and a monthly review, and shows how to reconcile platform numbers with analytics so that the targets rest on figures the business trusts. The bigger questions, whether a channel adds sales at all and what the mix should be, are answered by IncrementalityConversions caused by marketing. tests and modelling rather than by dashboards; the incrementality testing guide in the analytics root covers the test designs, and the attribution models framework explains why no attribution rule can answer that question on its own.
The target-led budget tutorial turns the targets into an agreement with finance that lets spend follow results inside limits: the target, a stopping rule, cash and payback limits, review points, platform settings and reporting. When money returns after a budget cut, the tutorial on restoring paid media spend after a cut brings it back in order, changing targets and budgets apart and reading the marginal cost at each step.
Performance and brand
Performance marketing harvests demand more efficiently than it creates it. Search can only answer searches that happen; remarketing can only follow visits that occurred; even paid social prospecting converts far better among people who already recognise the name. That means a business running performance alone tends to see efficiency fall as the pool of existing demand is used up, and to find that the cost of each result rises with every increase in budget. The usual response is to cut, which protects the reported efficiency and shrinks the business.
The alternative is to treat brand and performance as two parts of one system: brand activity makes more people willing to buy, performance activity converts them at an acceptable cost, and the measure of the whole is growth at an acceptable blended acquisition cost rather than the efficiency of either half. The brand vs performance guide in the strategy section sets out how to split the budget between the two and what evidence supports a longer-term view. For the performance marketer, the practical implication is to watch the leading indicators that brand activity moves (brand search volume, direct traffic, conversion rates on prospecting) and to report them alongside cost per result, so that the budget conversation is about the system and not about one dashboard.
A first performance plan
A first plan does not need every channel. It needs a target that comes from the economics, one channel where demand exists, tracking that was tested before launch, and a date for an honest review.
- Derive the target. From gross margin and an estimate of repeat purchase, set the most a customer may cost and the most a first order may cost. Write both down with the assumptions.
- Pick the channel by role. Where people already search for the product, start with search. Where they do not, start with paid social and expect a longer learning period. The cross-channel planning guide covers the second and third channels once the first is stable.
- Build the structure for the decisions. One campaign per distinct objective or market; consolidate everything that shares a target.
- Test the tracking before the first pound is spent. A conversion that does not record is a campaign that cannot be judged. The conversion events guide covers which event and value each platform should be given.
- Agree the review ritual. Daily checks for breakage, weekly changes with reasons, a monthly review against the target. The testing and scaling guide explains what to change first when the review shows a gap.
A furniture retailer's first quarter of performance marketing
Larkwood Furniture, an online retailer of made-to-order sofas, sets a first-order CPA ceiling of £120 from a gross margin of 45% on an average order of £900, keeping a third of the margin as contribution. It starts with a Google Search campaign on category and product terms at £150 a day and a Meta prospecting campaign at £100 a day, with purchase tracking confirmed on both before launch. After eight weeks search delivers orders at £85 and Meta at £160; the blended CPA is £118. Rather than cut Meta, the team notes that brand searches have grown 30% since Meta launched and that search CPA has fallen, runs a two-week geographic holdout, and finds that pausing Meta in the holdout regions lifts search CPA to £110. The plan for the next quarter keeps both channels, shifts a tenth of search budget to Meta and adds a remarketing layer capped at 5% of spend.
Common performance marketing problems
Most problems in a performance programme present as one of a few symptoms. The table routes each to the guide that diagnoses it.
| Problem | Usual cause | Guide |
|---|---|---|
| Targets set from platform suggestions, not margin | No economics chain from spend to contribution | Performance economics |
| Reports cannot answer which market or product is profitable | Structure split by platform habit rather than by decision | Campaign structure across channels |
| Every channel claims the same conversions | No channel roles; attribution read as truth | Channel playbooks, Performance reporting |
| Remarketing looks brilliant, growth has stalled | Budget drifting to the audience that was going to buy anyway | Prospecting vs remarketing |
| Efficiency collapses when budget rises | Scaling without tests, guardrails or marginal view | Testing and scaling |
| Spend split by last year's plan | No roles, no evidence-based allocation, no test reserve | Cross-channel performance planning |
| Weekly meeting with no decisions | Report built around platform metrics, not the plan | Performance reporting |
| Nobody knows whether the spend adds sales | No holdout or incrementality test has ever run | Incrementality testing |
| A fixed monthly budget stops spend that would have paid | Budget set as a number rather than an agreement on a target, a stopping rule and a ceiling | Target-led paid media budget |
The fundamentals hub is the place to start when the problem is a target or a structure; the optimisation hub when campaigns run but the loop is not producing decisions.
