What is Flighting?

A A$120,000 annual budget flighted into four six-week bursts is 168 active days, a daily rate of 120,000 ÷ 168 = A$714. Each burst spends 714 × 42 = A$29,988. Run continuously instead, the same total is 120,000 ÷ 365 = A$329 a day. The bursts buy more than double the daily weight when demand exists; the gaps mean the brand is absent for 197 days of the year, and automated bidding has to relearn at the start of each burst.

Flighting schedules advertising in defined periods of activity separated by periods of none. A campaign runs for six weeks, stops for six, and runs again.

The argument for it is concentration. A budget that would be unremarkable spread across a year can be substantial when compressed into the weeks where demand exists, and presence is not linear in spend — being noticeable during a period when people are buying is worth more than being barely present all year.

The argument against it is memory. Advertising works partly by accumulation, and what accumulates decays. A gap is a period during which the decay continues and nothing offsets it.

Which argument wins depends on the category, the size of the budget, and how established the brand already is, and the honest answer for most accounts is a compromise between the two.

Continuity runs at a constant weight throughout. It suits categories where demand is steady, where the purchase can happen any week, and where the budget is large enough to maintain a useful presence continuously.

Flighting runs bursts with genuine gaps. It suits categories where demand is genuinely seasonal — tax services, ski equipment, garden supplies — and situations where the budget cannot sustain meaningful weight across a full year.

Pulsing maintains a base level at all times and raises it during periods of higher demand. It is the compromise, and in performance accounts it is usually the right one: the base preserves the always-on segments and keeps the bidding models fed, while the peaks concentrate the money where it works hardest.

Three costs, and they are unequal.

Memory decay is the one most discussed and the hardest to observe. Brand recognition and recall fall during a gap at a rate that depends on how much was there to begin with. An established brand with distinctive assets decays slowly; a new brand with little in memory has little to lose and also little to return to.

Model decay is the one that bites performance accounts and is rarely planned for. Automated bid strategies build their predictions from recent conversion data. A campaign restarted after six weeks of nothing begins with a stale model and re-enters a learning period, during which delivery is erratic and cost is unrepresentative. On a short burst, that period can consume most of the flight.

Position loss is category-dependent. In auction-based channels, going dark means competitors take the impressions and, in some cases, the customers permanently. Where the purchase is one-off and the customer is then locked in for years, a gap is not a pause — it is a concession.

A burst has to be long enough to be worth starting.

For an automated campaign, that means long enough to clear the learning period with time remaining to deliver at settled performance. A fortnight is frequently too short. Four to six weeks gives the strategy room to stabilise and then perform.

For a brand message, it means long enough to build the frequency the message needs against the audience it is aimed at. A burst that ends after most people have seen the ad once has bought reach without repetition, which for most messages is the weaker half of the pair.

The temptation with a fixed annual budget is to cut more, shorter bursts so that the brand is never away for long. The arithmetic usually argues the other way: fewer, longer bursts clear the fixed costs of starting — the learning, the frequency build — a smaller number of times.

The single most useful modification to a flighting plan in a performance account is to refuse the full gap.

A low-weight always-on layer — brand defence on search, retargeting against recent site visitors, a small prospecting budget — costs a fraction of a burst and preserves three things: the conversion signal that keeps bid models current, the retargeting pools that would otherwise empty, and the defence of the brand's own name.

This converts pure flighting into pulsing, and in most accounts the cost of the base layer is repaid by the shorter learning period at the start of each peak alone.

The decision should rest on evidence the account already holds.

Search volume by month for the category's main terms. Past sales by week, ideally over more than one year so a one-off can be distinguished from a pattern. Conversion rate by period, which frequently varies more than traffic does. Competitive activity, where it can be observed.

What it should not rest on is the financial calendar, which is where flighting plans are most often set. Quarters are an accounting convention and demand does not know about them.

Where the data shows a genuine concentration of demand, flighting against it is straightforwardly correct. Where it shows demand spread fairly evenly, flighting is a way of being absent half the time for no reason.

A burst is a natural unit for a campaign idea, and this is the underused advantage of the structure.

Running the same execution across four bursts through a year gives up most of it. Each burst can carry its own creative, its own angle and its own offer, which addresses fatigue by construction rather than by rotation, and makes the return of the campaign an event rather than a resumption.

Planning the creative calendar to the flight calendar, rather than treating them separately, is what turns a budget-scheduling decision into an advertising one.