Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Performance Marketing › Channel Playbooks

Cross-Channel Performance Planning

How to plan performance across channels: assigning channel roles, allocating budget by role, targets per channel, flighting and pacing, and test budgets.
Guideintermediate8 min readUpdated

A plan is a set of bets with roles

Most performance budgets are set by inertia. Last year's channel split is carried forward, the channel with the best reported return gets a little more, and the one with the worst gets a little less. The result looks rational and is mostly driven by attribution: channels that harvest demand report well, so they grow until they saturate, while channels that create demand report badly and shrink, and the demand they were feeding quietly dries up.

A plan built on roles avoids this. The media planning framework sets out the structure: for each channel, the objective it serves, the role it plays, the budget it gets and the measure it is judged on. The IPA's work on effectiveness, which the institute publishes and promotes, has long argued that short-term activation and longer-term demand creation need to be planned and measured differently; a performance plan inherits that idea even when its horizon is short.

Thinking of each line as a bet also changes how the plan is reviewed. A bet has an expected outcome and a way to tell whether it paid off. When a channel misses, the question is whether the bet was wrong or the execution was, and the answer decides whether budget moves or the campaign changes. The wider choice of channels across all marketing, paid and unpaid, belongs to the channel mix guide; this page assumes the paid channels are chosen and plans how to run them together.

Assigning channel roles

Roles describe what a channel does in this business's plan, not what it does in general. The same platform can play different roles for different businesses, and occasionally two roles within one plan.

RoleWhat it doesTypical channelsJudged onNot judged on
Demand captureMeets people already lookingPaid search, Shopping, marketplace adsCost per result; share of available demand capturedGrowth beyond the demand that exists
Demand creationReaches people not yet looking and gives them a reasonPaid social, short-form video, online videoCost per new customer; incremental lift; downstream searchLast-click return
Reach and memoryBuilds Reach at low cost per thousandProgrammatic display and video, connected TVReach, frequency, attention, incrementalityDirect conversions
Re-engagementBrings back people who showed interestRemarketing on social and display, email-led audiencesIncremental conversions under a capReported return alone
Partner-ledBorrows audiences and pays on resultsAffiliates, marketplaces, partnershipsIncremental sales net of intercepted demandVolume of tracked sales

Assigning a role forces two useful conversations. The first is about measurement: a demand creation channel judged on last-click cost per acquisition (CPA) will always look poor, so the plan must say in advance how it will be judged. The second is about dependencies: demand capture channels draw on the demand the others create, so cutting creation channels shows up later as search volume falling, not as an immediate loss. The channel playbooks hub describes the roles in more depth and routes to each platform root.

Allocating budget by role and evidence

Once roles are set, budget follows two inputs: what the role needs to work, and how much is known about the channel's returns.

Demand capture is sized by demand. Search and Shopping budgets have a natural ceiling: the number of relevant queries at a price that meets the target. Fund them close to that ceiling, because the marginal result there is usually the cheapest in the plan, and stop when the next result costs more than the target allows.

Demand creation is sized by ambition and evidence. There is no natural ceiling, so the budget is a decision. Where incrementality tests or historical patterns show what a level of spend produces, fund to the level the economics support. Where evidence is thin, fund to a level large enough to be measured and plan the test that will produce the evidence.

Reach and re-engagement are sized by audience. Reach channels are planned by the reach and frequency they need to buy at a given price per thousand (CPM). Re-engagement is capped by the size of the audience it draws on, and the prospecting vs remarketing comparison covers how to set that cap.

The budget allocation guide covers the overall marketing budget; the marketing budget calculator and CPM calculator help turn targets and prices into channel figures. The allocator below lets you test how a split by role changes the volume and cost the plan should expect.

Budget allocator by channel role

Targets per channel and the blended target

The plan has one target that matters most: the blended acquisition cost or return that the economics require. Every channel target is a means of hitting it. Two rules connect them.

First, channel targets must differ by role. Demand capture channels can carry tight cost per result or return on ad spend (ROAS) targets because their results are close to the purchase and mostly attributable. Demand creation channels need looser click-based targets or a different measure altogether, such as cost per new customer confirmed in the business's own data. Re-engagement needs the strictest target, because a large share of what it reports would have happened anyway.

Second, the channel targets, weighted by planned spend, must reconcile to the blended target after allowing for the overlap between channels. Every platform claims conversions that others also claim, so if each channel exactly hits a target equal to the blended one, the blended result will miss. The adjustment comes from history: compare the sum of channel-reported results with actual new customers over a few months, and tighten channel targets by the observed overlap. The economics behind these numbers are set out in the performance economics guide.

Flighting, pacing and seasonality

A budget is also a shape over time. Flighting is the pattern of spend across the year: continuous, concentrated in bursts, or continuous with peaks. Pacing is how the spend within a period is released, day by day and week by week.

Performance spend should follow demand. In most categories demand is seasonal, and both the volume of searches and the price of attention rise around peak periods. Demand capture channels follow demand automatically if their budgets are not capped below it, so the main decision is whether to release more budget at peaks. Demand creation channels need to lead demand: the reach that persuades people to buy at a peak is bought in the weeks before it, when auctions are often cheaper, as cost per click (CPC) and price per thousand typically climb as competitors arrive.

Pacing matters most at the edges. Spending a monthly budget too fast leaves the end of the month dark; too slow leaves budget unspent and pushed into expensive final days. Platforms pace daily budgets on their own, and their help centres document how daily spend can vary around the setting while averaging out over the period. The plan should set monthly or weekly totals per channel and leave daily pacing to the platform, intervening only when the weekly check shows a channel drifting materially from its line. The monthly paid media review checklist is where each channel's spend, results and health are read against the plan before any money moves between lines.

Test budgets and the plan on one page

Every plan should hold back a share of budget to learn something it does not yet know: a new channel, a new audience, a new market, or an incrementality test on an existing channel. Without that reserve, the plan only ever confirms itself. Protect it explicitly, because when core channels look hungry, the test budget is the first thing raided. The 70-20-10 budget rule is one widely used convention for sizing that reserve beside a share for promising activity, with written rules for when money moves between them.

Each test line needs the same discipline as the rest of the plan: a hypothesis, a budget sized to give a readable answer, a duration, and a decision rule written in advance. The test design itself belongs to the testing and scaling guide.

The finished plan fits on one page. For each channel: role, budget for the period, monthly flighting, target, how it is judged, and the test attached to it, with the blended target and the test reserve at the foot. The media plan template provides that layout.

Illustrative example — not a real result

A pet insurance provider replaces a carried-forward split with a role-based plan

Pawsure, a pet insurer, had spent £1.2m a year split 60% search, 30% Meta and 10% display for three years. Search volume had flattened and its cost per policy had risen from £74 to £96. The new plan assigned roles: search for demand capture, funded to the point where the marginal policy cost £110; Meta for demand creation, judged on cost per new policyholder confirmed in the policy system; online video for reach in the eight weeks before the spring puppy season; remarketing capped at 8% of spend. A 7% test reserve funded a geographic holdout on Meta. The blended target was £85 per policy, with channel targets tightened by the 20% overlap seen in the previous year. After two quarters, branded search volume was up 15% in test regions and blended cost per policy was £88.

Common mistakes

  • Carrying last year's split forward. It encodes last year's attribution bias; start from roles and evidence instead.
  • One target for every channel. Demand creation judged on the same last-click target as search will always be cut, and search will later shrink with it.
  • Ignoring overlap when setting channel targets. If each channel exactly hits the blended target, the blended result will miss.
  • Capping demand capture below demand. The cheapest marginal results in the plan are usually the search queries left unfunded.
  • Starting reach at the peak. Demand creation bought only during the peak competes at the highest prices and too late to influence the purchase.
  • No protected test budget. A plan without a learning reserve can never discover a better plan.
  • Moving budget between channels weekly. Inter-channel changes belong to the monthly review; weekly changes belong inside channels.

Frequently asked questions

How much of a performance budget should go to tests?

Enough to produce a readable answer within the planning period, and no more than the business can afford to lose if every test fails. Many teams reserve a small, fixed share and protect it from being raided when the core channels look hungry. The size is better set by what each test needs to reach a conclusion than by a fixed rule of thumb.

Should every channel have the same CPA target?

No. Channels doing different jobs should be judged on different terms. A channel that captures existing demand can carry a tight target, while one that creates new demand will look more expensive on last-click figures and should be judged on cost per new customer and on incremental tests. What must hold is that the channel targets, weighted by spend, reconcile to the blended target the economics require.

How often should a cross-channel plan change?

The plan itself is usually set for a quarter or a season, with a monthly review that can move budget between channels within agreed limits. Daily and weekly changes belong inside channels, not between them. Rewriting the plan more often than the evidence changes turns it into a log of reactions rather than a set of bets.

Terms used on this page

Tools to try

Frameworks

Where this page is used

Part of these learning paths

Also related

Sources

  1. Institute of Practitioners in Advertising — IPA

Related