What this template is for
Use the memo whenever money moves between budget lines after the plan was approved: from one channel to another, from a reserve into a channel, between campaigns that report to different owners, or out of marketing altogether when a line cannot spend. Small moves inside a single campaign, such as shifting budget between ad sets that share an objective, usually do not need one; the test is whether someone outside the channel team would expect to be asked.
The memo sits between two other documents. Above it, the media plan set the original split by channel and period and, ideally, a named reserve with a rule for releasing it. Beside it, the guide to tracking spend against budget keeps the running position of every line, which is where the memo takes its numbers from and where the approved move is recorded afterwards. The memo is not the place to stop poor ads inside a channel: that is a routine optimisation decision covered by the pausing underperforming ads SOP. A memo is about money changing lines. Asking for new money for an investment the plan did not include, such as a new channel, a tool or a hire, needs the fuller argument set out in the tutorial on writing a business case for a marketing investment.
The point of writing it down is not bureaucracy. A move made on platform-reported returns that later turn out to be mostly non-incremental is a common and expensive mistake, and the memo is where that risk is named before the money moves. It also gives the next review something to check: a move with an expected effect and a review date can be judged, while a move made on instinct can only be defended. Most memos are written at a reforecastA rebuild of the full-year view of the budget from actual spend and what marketing now knows, done at the points finance sets; not a fresh budget., and the tutorial on reforecasting a marketing budget shows how the full-year estimate that produces them is built.
The template
Copy the table into a document or a ticket and keep the field names. The first row is the hardest and the most important; if it cannot be written in one sentence, the proposal is not ready.
| Field | What to write |
|---|---|
| Decision requested | The move in one sentence: amount, from which line, to which line, from which date, and who must approve it |
| Current position | Budget, spend to date and forecast to period end for every line the move touches, plus the total envelope |
| Evidence | The marginal return or marginal cost per result for the source and the destination, the period it covers and how it was measured |
| Incrementality caveat | Whether the evidence is platform-reported, modelled or tested, and what that does to confidence in the move |
| Constraints | Commitments that cannot move, minimum spends, learning phases, pacing limits and dates that fix spend in place |
| Expected effect | The result expected from the move as a range, with the assumption that drives the range |
| Risks and guardrails | What could go wrong, and the metric thresholds that would reverse the move before the review date |
| Review date and owner | When the move is judged, by whom, and against which numbers |
| Approval record | The approver, the date, and the version of the budget sheet that was updated |
Keep the memo to one page. Supporting analysis — the charts behind the marginal estimate, the holdout readout, the pacing report — goes in an appendix or a link, so the approver reads the decision and the evidence summary first.
How to fill each section
Decision requested
Write the move as a sentence with every part present: "Move a stated amount from paid search prospecting to paid social prospecting from the start of next month, approved by the marketing director." A memo that asks to "rebalance towards social" is not a decision; it is an invitation to argue. Name the approver because moves above a certain size, or across owners, usually need someone senior to the channel managers, and the memo should go to that person directly. Who that approver is for each size and kind of move belongs in the team's delegation table, which the guide to setting spending authority explains how to agree with finance.
Current position
Copy the budget, actual spend and forecast for each affected line from the spend tracker, not from the ad platforms. Platform spend and invoiced cost can differ, and finance will reconcile against invoices. Include the total envelope so the approver can see that the move does not change the overall budget, or, if it does, by how much. The spend tracking guide covers the accruals and timing differences that make this row harder than it looks at month end.
Evidence
This is the row that justifies the move, and it should show marginal returnExtra revenue returned by the last increment of ad spend., not average return. A channel with a high average return can be near saturation, so that the next unit of money buys much less than the last; a channel with a modest average can still have room. State the estimate for both lines, the period it covers, and the method: a spend-response curve, a budget step test, a marketing mix model read, or a geographic holdout. The marginal returns and forecasting guide explains how to estimate the curve and where Diminishing ReturnsEach extra unit of spend or effort buys less extra result than the one before it. begin, and the budget allocation guide explains why the comparison that matters is between marginal returns rather than averages.
Incrementality caveat
Say plainly where the evidence comes from. Platform-reported returns credit the platform with conversions that might have happened anyway; modelled returns depend on the model; tested returns, from a holdout or a geographic split, are the strongest. A move supported only by platform numbers can still be right, but the memo should say so and size the move accordingly. The incrementality testing guide explains the designs that produce tested evidence and how long they take, which matters when the memo proposes a test rather than a full move.
When the evidence is a model read, the guide to evaluating a marketing mix model lists the checks, from missing controls to the width of its intervals, that decide whether the memo can cite it for a full move or only for a test.
Constraints
List everything that limits how fast or how far money can move. Committed spend — insertion orders, sponsorships, minimum spends agreed with a publisher — cannot be reallocated without cost. Automated bidding needs stable budgets: large budget changes can send campaigns back into a learning period, and platforms document how daily budgets and campaign-level budgets behave when they change. The budgets and pacing guide covers Budget PacingHow fast a campaign spends its budget across the flight compared with the plan. in Google Ads, and the same logic applies to campaign budgets on Meta. Where these limits apply, the memo should phase the move rather than make it in one step.
Expected effect
Give a range, not a point. The range comes from the uncertainty in the marginal estimates for both lines: the source line will lose some results when its budget falls, and the destination line will gain some when its budget rises, and the net is the difference. Write the assumption that drives the range — usually the shape of the destination's response curve beyond the spend levels already observed — so the reviewer can see what would have to be true for the move to work.
Risks and guardrails
Name what could go wrong and set thresholds that trigger a reversal before the review date. A guardrail is a metric and a level: cost per result on the destination line, total volume across both lines, or a quality measure such as refund rate or lead acceptance. Setting that level from the metric's own recent range, as a control chart does, keeps an ordinary bad week from triggering a reversal. Without a guardrail, a move that is going badly runs until the next scheduled meeting. With one, the channel owner can reverse it on the evidence without asking again.
Review date and owner
Set the date on which the move is judged and name the person who will judge it. Allow enough time for the destination line to stabilise after the change and for conversions with a long lag to arrive. The performance reporting guide explains the review cadence the memo should slot into, so that the move is read in a regular review rather than in a special one.
Approval record
Record who approved the move, on what date, and which version of the budget sheet now reflects it. This row turns the memo into an audit trailThe record, attached to the work item, of what was reviewed, which version, who reviewed it, when, what they decided and the evidence relied on.: at year end, the sum of approved memos explains the difference between the plan that was signed off and the money that was actually spent.
Filled example
Below, the memo is completed for an invented business. The figures, the test and the names are illustrative.
Vossberry Meal Kits: moving prospecting budget from display to paid social
| Field | Vossberry's memo |
|---|---|
| Decision requested | Move £18,000 of the Q3 prospecting budget from programmatic display to Meta prospecting, in two steps of £9,000 on 1 August and 15 August, approved by Hannah Okafor, Head of Growth |
| Current position | Q3 prospecting budget £96,000. Display: budget £30,000, spent £9,400 to 20 July, forecast £29,000. Meta prospecting: budget £42,000, spent £14,100, forecast £41,500. Paid search non-brand: budget £24,000, unchanged. Total envelope unchanged at £96,000 |
| Evidence | A four-week geographic holdout on display (eight matched regions dark, eight live) measured an incremental cost per first box of £71, against £38 for Meta prospecting from a budget step test in June, when weekly spend rose from £2,800 to £3,600 and cost per first box moved from £35 to £38 |
| Incrementality caveat | Display figure is tested (holdout); Meta figure comes from platform-reported conversions during the step test, cross-checked with new-customer counts in the order system, which rose in line. Meta's incremental cost is therefore likely higher than £38; the move is phased to allow for that |
| Constraints | £6,000 of the display budget is committed to a private marketplace deal to 31 August and cannot move; Meta campaign budgets will be raised by no more than 20% at a time, a week apart, to avoid resetting delivery |
| Expected effect | Losing £18,000 of display at £71 per first box gives up about 250 first boxes; spending it on Meta at an expected £42–£55 per first box gains about 330–430. Net gain of roughly 75–175 first boxes over the quarter, assuming Meta's cost per first box rises as spend grows beyond the June test levels |
| Risks and guardrails | Reverse the second step if Meta cost per first box exceeds £60 for seven consecutive days, or if total first boxes across Meta and display fall below the July weekly average for two weeks running. Watch first-box refund rate on the Meta cohort; reverse if it exceeds 6% |
| Review date and owner | 22 September, read by Tom Achterberg, Performance Lead, using order-system first boxes by region and the spend tracker |
| Approval record | Approved by Hannah Okafor on 24 July; budget sheet version Q3-v4 updated by finance on 25 July |
On review, Meta's cost per first box had risen to £49 at the higher spend, inside the expected range, and no guardrail was breached; the second step stayed in place and the display line ran only its committed deal to the end of August.
Common mistakes
- Average return instead of marginal return. Moving money towards the channel with the best average result often moves it into saturation. Compare what the next unit of money buys on each line.
- Platform numbers presented as proof. Platform-reported returns overstate incrementality to different degrees on different channels. Say which evidence is tested and size the move to the weakest link.
- Ignoring committed spend. Money already committed to deals or sponsorships is not available, and a memo that counts it as movable overstates what the move can achieve.
- One big step. A large budget change can restart learning and distort delivery for days. Phase the move, and say in the memo when each step happens.
- A point estimate. A single expected number invites a pass or fail argument at review. A range with its assumption tells the reviewer what to check.
- No reversal rule. Without guardrails, a move that is going badly waits for the next meeting. Set the thresholds in advance so the channel owner can act on them.
- No record in the budget sheet. An approved move that never reaches the spend tracker leaves the plan and the actuals irreconcilable at year end. Update the sheet and record its version on the memo.
The media plan that sets the original split and the monthly report that tracks it are in the templates library.
