Method demonstration. This method demonstration uses fictional examples. It does not report client results or independently verified production behaviour.
Separate permission to spend from opportunity to serve
If a campaign is spending less than planned, first confirm that it is eligible and scheduled to run. Then check whether the selected inventory exists and whether the campaign can compete for it. Increasing the daily budget will not fix a rejected creative, an exhausted total cap or a targeting combination with no available traffic.
The aim is useful delivery within the agreed economics. Filling the budget is not a performance goal by itself. This is a review method for third-party traffic buying; platform controls and bidding modes differ, so use the actual account documentation for each setting.
Check the operating facts before touching the bid
Read the current campaign state, not only yesterday’s export. Note the account balance, approval status, daily and total caps, start and end dates, schedule timezone and delivery status. Compare the same elapsed eligible hours with an appropriate prior period, rather than today’s partial spend with yesterday’s full day.
PropellerAds’ own troubleshooting guide includes balance, budget limits, scheduling, targeting, inventory availability and bids among the checks for low delivery. It also warns that an empty retargeting audience can receive no impressions. These are platform-specific checks, not a promise that any one setting is the cause. Source: PropellerAds delivery troubleshooting.
Record the active bidding mode. A manual bid test and a conversion-based bidding system are different interventions. Before changing a conversion goal, verify that the intended event reaches the source and that the source is using it. Never send a fake live conversion to make delivery start.
Follow the first demonstrated constraint
| Evidence | Interpretation to check | Next step |
|---|---|---|
| Campaign cannot serve | Approval, account or schedule condition may block delivery. | Resolve the demonstrated condition within existing authorization; do not increase bids. |
| Eligible, but the selected slice has little supply | A narrow intersection or unavailable placements may limit volume. | Check available inventory and exclusions; propose one allowed targeting test. |
| Supply appears available, delivery remains low | Competition or a bid condition is a hypothesis, not yet a diagnosis. | Review bidding mode and economics before a bounded one-variable bid test. |
| Delivery began falling after a change | The new rule, goal, asset or targeting may matter. | Compare the change record and a matched window; isolate the change. |
| Reports are delayed or contradictory | The apparent gap may be incomplete observation. | Reconcile timestamps and spend before issuing another action. |
Where the platform provides a traffic estimator, treat it as a supply estimate rather than guaranteed delivery or conversion quality. PropellerAds also documents a drop-reason report comparing the previous two days; it is not an explanation of every intraday slowdown. Source: PropellerAds drop reasons.
Worked example: a permitted bid is still not a profitable bid
Fictional planning example. All amounts and outcomes below are invented, not an offer or a real account’s budget. Assume a manual CPC campaign with US$300 remaining in an approved test allowance. US$200 is already reserved for unfinished tests, leaving US$100 uncommitted. The active slice shows delivery, but less than planned; eligibility and destination checks passed.
The business has set a fictional ceiling of US$40 per approved outcome. If 1% of paid clicks eventually produce that outcome, a simple planning ceiling is 40 × 0.01 = US$0.40 per click. At 0.5%, it is only US$0.20. These calculations exclude other costs and do not predict the price or quality of additional inventory.
| Assumed mature conversion rate | US$0.30 cost per click | US$0.36 cost per click |
|---|---|---|
| 1.0% | Implied CPA: US$30 | Implied CPA: US$36 |
| 0.5% | Implied CPA: US$60 | Implied CPA: US$72 |
A proposed increase from US$0.30 to US$0.36 appears acceptable only under the stronger conversion assumption. If the rate is immature or new inventory may differ, the uncertainty is material. A bid ceiling is also not the same as realized average CPC.
I would leave the US$200 reservation intact. If authorized, a separate US$30 maximum-spend test could examine the bid hypothesis within the uncommitted allowance, with targeting and assets fixed. US$30 is a fictional risk allocation, not a universal threshold or a sample-size claim. The run may show whether delivery changes while remaining far too small to judge approved-outcome economics.
Make the test and the decision separate
The brief should identify the single setting, starting value, proposed value, permitted inventory, spending cap, eligible observation window, review owner and restoration rule. Check whether the platform changes learning or other settings automatically when that value changes. Preserve the starting configuration.
Observe impressions, clicks, realized cost, placement composition and spend after the change. Later, review mature approved outcomes using the agreed window. Higher delivery is evidence that delivery changed; it does not prove the extra traffic is equally valuable.
If the cap is reached before useful evidence arrives, stop or hold as specified. Do not expand geography, replace the creative and increase the bid together merely to force a spend target. If there is no economically acceptable inventory, underspending can be the correct decision.
Leave a diagnosis that another buyer can continue
A concise handoff states the observed shortfall and matched hours, the verified constraints, the one change if any, its authorization and readback, remaining allocation and next review. Label unverified causes as hypotheses. Do not write “bid too low” when all you know is “spend below plan.”
Continue with budget pacing when conversions arrive late for the allocation ledger, missing source conversions for signal checks, and verifying a campaign change before treating an action as complete.