Paid media
What should a PPC monitoring system check before raising an alert?
A PPC monitoring system should check data quality, spend against the agreed plan, meaningful performance changes and the commercial context around them. Every alert should explain what triggered it, show the evidence and identify the next useful action.
An alert earns attention when it helps an account manager make a better decision.
Check whether the data is trustworthy
Before flagging a performance issue, establish whether the underlying data is complete and comparable.
Has the source refreshed? Is the latest day complete? Have tracking definitions changed? Are recent conversions still arriving? Has a campaign been renamed, paused or moved?
A sudden fall in reported conversions warrants investigation. The first investigation may concern measurement.
Keep data-quality alerts distinct from performance alerts so the recipient knows what to check.
Compare spend with the agreed spending plan
A straight-line pacing calculation is useful when spend is intended to be relatively even.
Suppose an account has a £6,000 budget across a 30-day month. At the end of day twelve:
- Straight-line planned spend is £2,400.
- Actual spend is £3,000.
- Spend is £600, or 25%, ahead of that plan.
- The remaining £3,000 allows an average of approximately £166.67 per day across the remaining eighteen days.
Those calculations describe the position. They do not determine the correct response.
An account might intentionally spend more during an early-month promotion. The monitor needs that plan, including approved exceptions, to judge whether the variance matters.
Look beyond percentage changes
A 100% increase from one conversion to two has a different evidential weight from an increase from 500 to 1,000.
Pair relative movement with absolute volume and commercial impact. Define minimum data requirements appropriate to the account rather than applying a universal threshold.
Useful dimensions include:
| Dimension | Question |
|---|---|
| Magnitude | How large is the change in absolute terms? |
| Persistence | Has it continued across comparable periods? |
| Concentration | Is it account-wide or isolated? |
| Context | Does a known change explain it? |
| Consequence | What happens if nobody acts? |
A small account may need a longer comparison window to support a meaningful efficiency judgement. A broken purchase journey can require immediate investigation regardless of sample size.
Write an alert someone can act on
A fictional budget alert might read:
Account spend is ahead of the monthly plan.
£3,000 spent against £2,400 planned through day twelve: £600 above plan. Data refreshed this morning. No approved front-loaded promotion is recorded. The account manager should confirm whether the remaining daily allowance is appropriate before changing budgets.
That is enough to begin a decision. Include a link to the underlying report, the relevant dates and a record of any subsequent action.
Keep review and account changes separate
Monitoring can prepare recommendations while an account manager controls the response.
For an initial system, use read-only access and a review queue. A suggested budget adjustment should make its assumptions, affected campaigns and expected consequences visible.
If automated account changes are introduced later, define their scope, limits, monitoring and recovery process explicitly.
Measure the quality of the alerts
Review how often alerts identify a real issue, how many are dismissed, how quickly useful alerts are acted on and which problems the system missed.
Repeated alerts without new evidence can be grouped. Known exceptions should expire or be reviewed so they do not hide future problems indefinitely.
Peach & Stone combines paid media experience with systems that support consistent account oversight.
Next step: List the five account problems you most want to catch earlier. Use those to define the first monitoring rules and the evidence each rule needs.