AdRoll minimum deposit
what it unlocks and what a test really costs

Audit essential site events, consent coverage, product identifiers, customer exclusions, order status and refund handling. Estimate the engineering or operations work needed to make the audience definition reproducible. A tiny media charge cannot compensate for unreliable eligibility data.

Check whether the intended audience is large enough to deliver under the selected country, recency and channel conditions. If it is not, redesign the audience or evidence question before allocating spend rather than weakening the definition after launch.

Prepare a creative sequence appropriate to each journey stage. Budget variants for product viewers, cart abandoners or broader site visitors separately. Reusing one generic message can obscure whether poor performance came from media or journey mismatch.

Create a technical allowance to verify audience population, creative serving, destination, conversion transport and billing. Then create a separate contribution allowance sized for the eligible audience and the minimum incremental-order evidence the business needs.

Use daily and total media limits, frequency boundaries and a review schedule. Stop when audience eligibility breaks, repetition exceeds the approved range, order data cannot mature cleanly or spend reaches the loss limit before the evidence floor.

Keep reserve outside active campaign authority. Release it only after the technical path, audience composition and early holdout records are consistent. Pay-as-you-go access supports this staged control but does not replace it.

Match account charges or invoices to campaign spend, dates, currency and adjustments. Keep media charge, software or service fee and add-on cost in separate ledger rows. Resolve discrepancies before calculating the total test cost.

Add readiness and operator costs to a total evidence-cost view, while keeping them outside CPM. The record should show what the business spent on media, what it spent to make the test valid and what incremental-order evidence it obtained.

Archive or pause audiences and creatives according to the approved closure plan. Confirm that no unintended campaign remains eligible, preserve the holdout result and record the conditions required to reopen spend.

Name the journey problem before creating a campaign. Examples include measuring cart-recovery incrementality, checking product-feed continuity for recent viewers or testing whether a broader visitor audience is large enough for stable delivery. One worksheet should answer one problem.

Record the eligible audience count and expected replenishment. Calculate the maximum reachable users under the planned recency and frequency rules. This sets an upper boundary on useful impressions and prevents a budget from demanding more delivery than the audience can absorb responsibly.

Prepare the counterfactual before media release. Define holdout eligibility, allocation, contamination checks, order maturity and the smallest decision-relevant difference. If the design cannot distinguish contribution, reduce the claim and fund a diagnostic rather than a performance verdict.

Model charge scenarios from observed or bounded dynamic CPM ranges. Include low delivery, expected delivery and high competition, then add creative, data and operator costs outside media. The worksheet shows cash exposure without presenting any scenario as a promised rate.

Set checkpoints for technical continuity, audience composition, reach, frequency, order reconciliation and loss. Every checkpoint names the owner and action. Pause authority must be available when a feed, event or exclusion fails.

Close the worksheet with charge reconciliation, residual campaign state and evidence gained. The final line says whether to scale, revise audience or creative, repair data, retain the diagnostic only, or stop.

At the smallest level, ask only whether events populate the intended audience, exclusions work, creative serves to the correct destination and charges reconcile. Do not judge order economics from a diagnostic cell with insufficient eligible users or maturity.

At the next level, ask whether the audience can deliver without excessive repetition and whether platform exposure joins to paid orders. Use reach and frequency gates before opening a causal test.

At the contribution level, fund a counterfactual and mature-order window capable of detecting the smallest business-relevant difference. State when audience volume makes that goal infeasible.

At the scale level, release budget only for audience-stage cells that passed contribution and operational gates. Continue to protect customer exclusions, channel overlap and marginal frequency.

If any level fails, choose repair, redesign or stop. Do not add spend to compensate for broken data, tiny audiences or a counterfactual that cannot answer the question.

Record the decision produced at every level. Pay-as-you-go flexibility is most valuable when each small release purchases a specific piece of evidence.

Any request to raise daily or total spend identifies the audience-stage cell, current reach and frequency, counterfactual status, mature incremental orders, marginal contribution and remaining uncertainty. A deadline or unused account capacity is not an evidence-based reason.

Finance confirms charge reconciliation and package scope. Media operations confirms delivery controls. Analytics confirms the order and holdout joins. Merchandising confirms product and destination continuity. The exception proceeds only when the relevant owners approve their evidence.

Time-limited approval states the amount, duration and automatic reversion. If the additional spend changes audience composition or channel mix, create a new test interval rather than merging it into the old result.

Archive denied requests with the failed gate. This prevents repeated pressure to fund the same unresolved problem and shows what evidence would allow reconsideration.

Review cumulative exceptions each month. Repeated temporary increases can silently become the real budget policy, bypassing the original audience and loss controls. Consolidate only after the complete contribution and charge history supports a revised standing limit.

Attach the revision to a new controlled version of the launch worksheet and preserve the replaced limit. Finance and media owners should be able to trace every standing increase back to mature accepted incremental contribution evidence, not merely to earlier temporary approvals.

The AdRoll pricing page checked on 2026-08-11 states that self-service advertising is pay-as-you-go with no minimum spend. Managed services and add-on products can follow different commercial rules, so the self-service statement must stay attached to its product route.

No minimum spend removes an entry floor; it does not remove the need for audience readiness, an evidence budget, loss limits and current billing verification inside the account.