
Campaign Setup
Part of Programmatic campaign optimisation
Balancing prospecting and retargeting budgets
Review prospecting and retargeting spend using audience opportunity, mature outcomes and the value of additional delivery.
undefined
Set the prospecting and retargeting split by comparing what the next dollar in each group is likely to achieve against the campaign goal. There is no dependable default percentage: base the working split on eligible opportunities, delivery and the additional actions each group can produce. Set an allowable cost per additional action from the campaign brief before moving funds.
Retargeting can show a lower reported cost per attributed action because it selects people who interacted earlier. Attribution alone does not show how many extra actions the ads caused, so do not use a lower reported cost as proof that retargeting deserves more budget. Test whether it adds actions before comparing its value with prospecting.
Define the prospecting and retargeting rules
Define prospecting as the campaign's eligible audience beyond its specified retargeting rule. Define retargeting by the prior action and recency window that make a person eligible. Choose a window that reflects how long the earlier interaction remains relevant and the campaign's observed time to action, then record the exact rule.
Record inclusion and exclusion rules, message, primary outcome, available inventory and authorised spend for each group. The groups may have different paths to action, so their reported rates need context. Use the same outcome definition when comparing them, and document any differences in how they are measured.
Check whether the retargeting list is current and whether its membership rule still describes people who need the message. Record the list size and the window used to build it, then check whether people remain eligible for the full period. Set a frequency cap to limit repeated exposure, and review delivery to ensure it does not prevent useful reach.
Treat reach and inventory estimates as planning signals, not a precise count of people available for this split. Check eligible list size and actual delivery before assigning more retargeting budget. Use observed delivery to decide whether there is enough volume for a useful test rather than relying on a universal minimum list-size threshold.
Keep the retargeting window separate from the attribution lookback: the first determines who can receive an ad, while the second determines which actions are counted against it. Choose and record both before comparing results. Use campaign evidence about audience relevance and time to action to inform those choices, rather than treating either setting as a default.
Key Metrics for Prospecting and Retargeting Evaluation
- Retargeting List Size
- Record size and recency window used to build list
- Frequency Cap
- Set to limit repeated exposure; review delivery regularly
- Attribution Lookback Window
- Separate from retargeting window; must be defined and recorded
- Incremental Action Test
- Use holdout audience to measure added actions beyond attributed ones
Compare the next useful opportunity
Read completed outcome counts and cost beside delivery and reach where reliable, and wait until outcomes have had time to mature under the agreed attribution window. Retargeting may encounter a small pool or repeated exposure; prospecting may need more time before an action. Compare each group's next available opportunities, not just its average reported rate.
A low retargeting CPA from a few attributed conversions is weak grounds for taking most of the prospecting budget. Equally, an allocation to prospecting needs review if it cannot find relevant inventory. Check whether the retargeting actions are incremental and whether prospecting has room to deliver before deciding which group should receive the next approved amount.
Use the same four checks for both groups. Record the evidence for each check against the campaign goal before changing the allocation.
For eligibility, estimate what relevant opportunities remain under each group's rules and flight. For delivery, check whether budget is spending and where requests are filtered or lost; for outcomes, count completed actions only when they are mature under the agreed definition. For additional spend, judge whether it is likely to buy useful new exposure or mostly repeat existing delivery.
These are decision checks, not a percentage formula. When reach reporting is unavailable or uncertain, use the delivery and frequency evidence that exists and label the gap. Inspect creative, inventory and the landing page before treating a weak prospecting result as a budget problem, and investigate targeting or repeated exposure before increasing retargeting spend.
Measure whether retargeting adds actions with an audience holdout. Randomly assign eligible people to a retargeting group or a group excluded from retargeting, keeping the audience rule, outcome definition and measurement period aligned. Compare completed actions between the groups to estimate additional actions, then compare the cost of that lift with the campaign's allowable cost per action.
Prospecting vs Retargeting: Key Decision Factors
- EligibilityEstimate remaining relevant opportunities under each group's rules and flight
- DeliveryCheck budget spending, ad requests filtered or lost, and frequency cap effectiveness
- OutcomesCount completed actions only after attribution window has elapsed and data is mature
- Additional SpendAssess whether new budget buys useful exposure or repeats existing delivery
Set and adjust the budget allocation
Document a working split, an authorised adjustment range and the evidence that will trigger a review. Protect any minimum required by the brief or a defined learning period. Set the review point so there is time to deliver and for outcomes to mature under the agreed attribution window.
Move only an approved test amount when the receiving group's next opportunities look more useful, then check whether its incremental outcome quality holds as delivery expands. If retargeting does not produce useful additional actions, do not justify its allocation with attributed conversions alone. If one role cannot deliver usefully under the brief, revise the allocation rather than preserving a percentage for its own sake.
Check automation before calling the split fixed; confirm how it may affect budgets and whether it is suitable for a split that must stay within set amounts. Verify the settings that can change delivery or spending before attributing a result to the planned allocation.
Inspect the automation settings and saved budgets that actually govern the two groups. In DV360, an insertion order can hold a budget, budget segments with date ranges and a frequencyCap field; the API example sets the cap to unlimited (Google for Developers). Separate insertion orders or platforms require separate checks.
Report the roles separately and explain the trade-off between attributed and additional actions. Set the attribution lookback and outcome definition before comparing groups, apply them consistently and count outcomes as mature only after the agreed window has elapsed. If evidence is sparse, retain a bounded allocation and state what further delivery or mature outcomes would justify a change.
How to Evaluate and Adjust Budget Allocation
- Document working split and review triggersInclude authorised adjustment range and evidence for change
- Protect minimum required by brief or learning periodEnsure sufficient data is collected before adjusting
- Move only approved test amountOnly shift budget if next opportunity appears more valuable
- Verify automation settings and saved budgetsCheck insertion orders, budget segments, and frequency caps in DV360
- Report roles separately with trade-off explanationClarify attributed vs additional actions and consistency in measurement



