Budget Allocation & Pacing

Most budget decisions are made on the wrong number: the comfortable blended average, which numbs you to the place where money actually leaks. Because auctions price each new customer higher, the marginal dollar can be losing money while the average looks fine — and performance spend quietly drains a demand pool that brand is meant to refill. This module is budgeting on the margin, pacing for learning, and funding both engines on purpose.

Why allocation is a margin problem

Budget allocation is a margin problem disguised as an average problem. Because auctions price each additional customer higher than the last, the cost of your NEXT customer (marginal CAC) rises with scale while your blended average stays calm — so the average can look healthy while the last dollars lose money. Good allocation caps spend where marginal CAC crosses allowable, and funds brand and activation as two deliberate engines.

Most budget decisions are made on the wrong number. A team sees a comfortable blended CAC, concludes the channel is healthy, and pours in more — without noticing that the last slice of spend is acquiring customers far above what the business can afford. The average is an anesthetic; it numbs you to exactly the place where money starts leaking. This module is about budgeting on the number that actually governs scale: the marginal one.

There is a second axis the average hides too. Performance spend is only one of two engines — the short, measurable activation engine — and it draws down a pool of demand that brand-building (the long engine) is supposed to refill. Allocate everything to activation and the numbers look great until the pool drains and CACs climb, which is the Adidas story from module 1 told as a budgeting failure. The hero figure holds both ideas at once: cap at the margin, and fund both engines on purpose.

Planning from unit economics

Budget planning starts from the four numbers established in module 1 — gross margin, allowable CAC, payback ceiling, breakeven ROAS — signed off by finance. Everything downstream is that math wearing channel names. Annual planning sets the brand : activation balance and the channel mix; quarterly planning re-walks the caps as auctions and competition move the marginal-CAC curve.

A budget is not a number you negotiate; it is a conclusion you derive. Before any channel split, the economics have to be settled and owned by finance, because every later allocation decision is just that math applied to a campaign. Teams that skip this step end up defending spend with platform ROAS — the flattering number module 1 warned about — instead of with the margin reality the budget actually answers to.

Pacing: the three ways a budget dies

Pacing is how evenly a budget is spent across a period, and it matters for two reasons: front-loading exhausts the budget early and goes dark when end-of-period conversions are often cheapest, and — less obviously — jerky spend starves automated bidding of the steady daily signal it needs to stay out of the learning phase. Even pacing is learning hygiene, not a finance preference.

Within a single month a budget can die three different deaths, and only one pacing pattern avoids them. The chart below contrasts the steady ideal with the classic front-load: spend the budget too fast and the account goes dark in the final stretch, surrendering the cheap end-of-period conversions and jolting the bidder out of its rhythm.

Claim: Marginal CAC rises with scale because auctions price each additional customer higher than the last — so a healthy blended average can hide frontier acquisitions priced above allowable. Source: RGM analysis from client spend-curve audits. Context: Budget caps belong where the marginal line crosses allowable CAC, which is reached well before the average looks alarming.

Allocating to the margin, not the average

Allocate the next dollar to where its MARGINAL return is highest, and cap each channel where its marginal CAC crosses your allowable CAC — not where the channel’s average still looks profitable. The average pools cheap early demand with expensive frontier demand; the marginal line is where scaling decisions actually live. Find it with stepped spend tests and geo experiments.

The single most useful instrument in budgeting is not a forecast — it is the marginal-CAC dial. Drag it and watch the deception play out: the average CAC barely moves while the cost of the last dollar climbs past your allowable. The cap belongs at that crossing, and it is always reached well before the blended number looks alarming.

Always-on vs flighted

Always-on suits demand capture and channels where the bidder benefits from continuous learning; flighting (concentrated bursts) suits genuine launches and seasonal peaks where reach matters more than steady presence. The hidden cost of flighting a performance channel is that each restart re-triggers the learning period — so reserve bursts for cases where their value clearly exceeds the re-learning tax.

The always-on-versus-flighted question is usually argued on brand grounds, but in performance it is mostly a learning-cost question. Every time you go dark and come back, the bidder pays a re-learning toll. That makes always-on the default for capture channels and flighting a deliberate exception you take only when a burst’s concentrated value is worth restarting the clock.

Dynamic reallocation and the 60/40 balance

Dynamic reallocation moves budget toward what is working — but in steps under ~20% so it does not re-trigger the learning period, and toward marginal (not average) efficiency. Above this sits the deliberate brand : activation balance, anchored near Binet and Field’s 60:40, because pure activation drains the demand pool it draws from.

Reallocation should be continuous but never jerky. The instinct to yank budget toward this week’s winner in one big move buys a fortnight of volatility on top of whatever the move was for. And reallocation inside the activation budget is only half the job — the bigger lever is the balance between the two engines, which the dial below lets you feel.

Seasonality planning

Plan seasonality as scenarios with pre-committed moves, and get ahead of peaks: raise budgets and loosen targets BEFORE a known rush so the bidder exits its learning period before the demand arrives rather than during it. Lower spend deliberately after. The classic mistake is reacting to a peak in real time, when the learning lag means you optimize just as the demand recedes.

Seasonality punishes the reactive. By the time a peak is visibly underway, raising budgets triggers a learning period that resolves after the peak has passed — you pay the volatility and miss the demand. The fix is to treat known peaks as planned events: pre-load budget and loosen targets early so the machine is calm and confident when the rush hits.

Scenario planning

Scenario planning means building bad / base / best cases with the exact moves pre-committed at each trigger — what scales, what holds, what gets cut — decided while everyone is calm, tied to leading indicators (MER drift, marginal CAC crossing allowable) rather than gut feel. The point is to make the hard cut decision before the panic, so execution is fast and unpolitical.

Single-number forecasts are comfortable and useless; they tell you nothing about what to DO when reality misses. A scenario plan does the hard thinking up front — it names the triggers and the responses before the quarter starts, so a week-six miss becomes a playbook execution rather than a philosophical debate held under pressure.

Advanced playbook

Advanced allocation runs as a living system: marginal-CAC caps re-found quarterly, the brand : activation balance set by econometrics rather than habit, budget paced for learning hygiene, and a standing scenario plan with evidence-triggered moves. The annual mix is an MMM question and the channel caps are experiment questions — both re-walked as the curves move.

The senior move is to stop treating the budget as an annual document and start treating it as an instrument you re-tune against curves that never sit still. Auctions reprice, competitors enter, demand cycles — so the marginal-CAC line, the saturation points, and the optimal mix all drift. The build below is how that living budget goes up and stays honest.

Common mistakes

The classic allocation mistakes share one root: budgeting on the average and the forecast instead of the margin and the scenario. Scaling on blended CAC, front-loaded pacing, jerky over-20% moves, over-rotating into pure activation, reacting to seasonality in real time, and single-number forecasts are the recurring six.

• Scaling on blended CAC. The average hides the frontier; the last dollars can be losing money while the blend looks fine. Cap at the marginal crossing.
• Front-loading the budget. Exhausts spend early, goes dark when conversions are cheapest, and starves the bidder of steady signal. Pace even.
• Yanking budget in big jerks. Moves over ~20% re-trigger the learning period; the volatility costs more than the optimization gains. Step it.
• Over-rotating into pure activation. Performance drains the demand pool brand refills; the Adidas 77/23 profile drains it dry. Fund both engines.
• Reacting to seasonality in real time. The learning lag means you optimize as demand recedes. Lead each peak by a learning period.
• Single-number forecasts. They tell you nothing to DO on a miss. Pre-write bad/base/best with trigger-mapped moves.

Quick answers

How should I split my marketing budget?

Start from unit economics (margin, allowable CAC, payback), then set a deliberate balance between brand-building and activation — Binet and Field’s research points to roughly 60:40 on average, varying by category. Within the activation budget, allocate to where the next dollar’s MARGINAL return is highest (not the average), cap each channel where marginal CAC crosses allowable, and pace spend evenly across the period.

What is the 60/40 rule in marketing?

It is Binet and Field’s finding, from the IPA effectiveness databank, that campaign effectiveness tends to peak when about 60% of budget builds the brand (the long-term, compounding engine) and 40% drives short-term activation. It is an average across categories, not a law — your optimum is an econometrics/MMM question — but it is the standard caution against over-rotating into pure performance activation.

What is marginal CAC and why does it matter for budgeting?

Marginal CAC is the cost to acquire the NEXT customer at your current spend level, as opposed to the average cost across all customers. Because auctions price each additional customer higher, marginal CAC rises with scale while the average stays calm — so the average can look healthy while the last dollars are bought above your allowable. Budget caps belong where marginal CAC crosses allowable CAC, not where the average still looks fine.

What is budget pacing and why does it matter?

Pacing is how evenly you spend a budget across a period. Front-loading exhausts the budget early and goes dark when end-of-period conversions are often cheapest; back-loading underspends when you could be learning. Beyond wasting money, jerky pacing starves automated bidding of the steady daily signal it needs — so even pacing is learning hygiene, not just a finance preference.

Should I run always-on or flighted campaigns?

Always-on suits demand capture (search, retargeting) and channels where the bidder benefits from continuous learning; flighting suits genuine bursts (launches, seasonal peaks) where concentrated reach matters more than steady presence. The risk with flighting performance channels is that each restart re-triggers the learning period, so reserve it for cases where the burst’s value clearly exceeds the re-learning cost.

How do I plan budget for seasonality?

Plan it as scenarios, not a single forecast: build bad / base / best cases with pre-committed moves at each trigger, raise budgets and loosen targets ahead of known peaks (so the bidder exits learning before the rush rather than during it), and lower them deliberately after. Tie the triggers to leading indicators like MER drift and marginal CAC so the decision fires on evidence, not panic.

Operating checklist — score yourself

Use this as the operating standard for allocation and pacing. None of it is a spreadsheet ritual — it is the discipline of budgeting on the marginal truth instead of the average comfort, and funding both engines so the demand pool never runs dry.