The annual budget usually starts life as a sentence ("last year plus 10%") and dies in a spreadsheet nobody looks at again. This lesson builds it a different way: from the actual seasonality of the business, with peaks prepared in good time, a reserve that prevents panic and a mechanism that lets the budget follow the results instead of being frozen in January.
The seasonality index
From 2-3 years of history (conversions and spend by month; failing that, search volumes from the Keyword Planner or Google Trends for your sector), work out the index for each month: that month's conversions ÷ the monthly average. An index of 1.4 in November and 0.6 in August describes your demand curve. Do it per campaign or family when the seasonality differs (heating vs air conditioning).
The monthly plan comes out of annual budget × the month's index ÷ 12:
more money where there is more demand, not the same every month.
Peaks: raise before, not during
A demand peak starting on 15 November needs its budget and its changes one or two weeks earlier: Smart Bidding takes time to calibrate, new creatives need history, and impression share is won before the competitors arrive. A standard plan for a peak:
| When | What |
|---|---|
| −4 weeks | Creatives, landing pages and feeds for the peak ready; negatives reviewed |
| −2 weeks | Raise budgets in steps towards the peak level; loosen targets if you are after volume |
| −3 days | Smart Bidding seasonality adjustment if the peak is sharp and short |
| During | Daily monitoring; do not touch the structure |
| +1 week | Bring budgets and targets back in steps; write down what you learnt for next year |
The reserve
10-15% of the annual budget left unallocated in January, for three uses: peaks that beat the forecast (never leave lost IS to budget in your best month), testing (new campaigns, new networks, lesson 4) and the unexpected (a new competitor, an opportunity, a crisis). Reserve that goes unused is allocated in the final quarter; reserve that gets used is justified by its result.
Budget tied to results, not to a percentage
The problem with "5% of sales goes to advertising" is that the budget chases the result instead of producing it. The alternative: budget by goal and marginal CPA:
- The year's business goal (customers, sales) and an acceptable CPA/ROAS.
- Curves per campaign (lesson 2): how many conversions at what marginal.
- The budget is whatever it takes to buy the profitable conversions up to the point where the marginal is still within target — not a euro more (the flat zone) and not a euro less (lost IS with a good CPA).
- Quarterly review: if campaigns are beating target with lost IS, the budget goes up; if the curve has flattened, it gets reallocated.
That way the budget grows when there are results to buy and eases off when there are not, without waiting for next year.
Presenting it to whoever signs it off
One page:
- The monthly plan (a 12-month table with index, budget, forecast conversions, forecast CPA).
- Assumptions (seasonality from history, estimated marginal CPA, reserve).
- Scenarios (−20% / base / +20%) with forecast conversions, from the Planner or from your curves.
- Review rules (quarterly, tied to attainment and IS): the plan changes with the data, and that is a strength, not a weakness.
- What is not being promised: forecasts are estimates; competition and measurement move them.
Annual planning mistakes
- The same budget every month against seasonal demand.
- Raising the peak's budget on the day of the peak.
- No reserve: funding the unexpected out of what works.
- Setting it in January and never touching it: December's market is not January's.
- Planning without clean measurement: forecasts built on false data.
💡 Ninja trick: the annual plan lives in a spreadsheet; its execution lives in the scripts. Smart Bidding (SBNS) accepts goals and budget limits per month per campaign (your seasonal plan) and moves the day-to-day within those limits; Guardian watches the pace against the month's plan. You review the plan each quarter; the machine runs it every night.
What you should remember
- A seasonality index per month (and per family) from your history: the monthly plan follows demand.
- Peaks: prepare at −4 weeks, raise at −2, come back down at +1.
- A 10-15% reserve for peaks, testing and the unexpected.
- Budget by goal and marginal, reviewed quarterly; not by a fixed percentage.
- Present the plan, the assumptions, the scenarios and the review rules.
That is the end of Module 9. The last one in the Intermediate level, Scripts I, explains what Google Ads scripts are, what they solve and how the Suite is installed.