What is PPC monitoring?

PPC monitoring is the continuous checking of a paid campaign while it is running, rather than the review of it afterwards. It covers two things a report cannot: the numbers the campaign produces — spend, conversions, cost per acquisition — and the destination the click lands on. The output is not a chart. It is an alert to a named person, raised while the money is still moving.

What should you monitor in a PPC campaign?

Four things in the numbers and six on the other side of the click. In the numbers: a metric that stopped arriving, a cost through a ceiling you set, a target drifting off pace, and two systems reporting different values for the same figure. Past the click: a soft 404, a redirect to somewhere else, a page rendering broken, conversions stopping while spend continues, a failure that only happens on a phone, and a page refusing visitors outright. Every one of the six returns HTTP 200 to a status-code monitor.

How often should you monitor PPC campaigns?

Set the interval by the spend rate, not by the reporting calendar. At $58.59 a minute across three campaigns — one real advertiser’s rate — a five-minute check caps a single failure at roughly $300 and an hourly check caps it at about $3,500. Daily checks make the cap the size of a day’s budget.

Is PPC monitoring the same as PPC reporting?

No, and the difference is who the output is for. A PPC report is assembled for somebody else to read on a schedule — a client, a director, a board — and it describes a period that has already closed. Monitoring has no schedule and no reader: it runs continuously and produces nothing at all until something is wrong, at which point it interrupts a person. A tool that emails a beautiful report every Monday can let a campaign spend into a dead page from Tuesday to Friday without a word.

The category

PPC monitoring

PPC monitoring is the continuous checking of a paid campaign while it is running — the numbers it produces and the page the click lands on — so that a campaign which has stopped working reaches a person instead of a month-end review.

Most of the category runs inside the ad account: disapproved ads, budget pacing, bid rules, settings that drifted. That work is worth doing and it is not what this page is about. Two kinds of failure live outside the account, and between them they are where the expensive ones are.

Four failures in the numbers

These run on whatever figures arrive, from any system that can report one. None of them needs to see inside an ad account. The intervals are the real schedules, read off the job table.

A number stopped arriving

every minute

Conversions were reporting every hour and then stopped, while spend carried on. Nobody gets an alert for a chart that simply ends.

Missing, zero and late are three different states, and every tool that treats them as one is why a dead campaign looks like a quiet week.

A cost passed the line you set

every minute

Cost per acquisition through its ceiling, spend rate above its cap — the threshold you would have checked on Monday, checked continuously.

A guardrail in a spreadsheet is a number somebody has to remember to look at. This one is a row in a database that pages when it breaks.

A target went off pace

every 15 minutes

Not that the goal was missed — that at the current rate it is going to be, while there is still a month left to do something about it.

Month-end is when a missed target becomes a fact. Pace is when it is still a decision.

Two systems disagree about the same number

every 15 minutes

The ad platform claims 140 conversions; your database has 96. One of them is wrong and the gap has been there for a week.

Nobody reconciles two dashboards by eye. A number you cannot trust is worse than one you do not have, because you will act on it.

Six failures past the click

The ad account can be in perfect health while every click it buys arrives somewhere useless. These are checks on the destination itself, fetched as a phone, every five minutes. Five of the six answer 200 OK; the sixth answers with a refusal that an uptime monitor files as downtime and an ad platform does not notice at all.

The page says it is gone

A CMS catch-all, a route with no match, a deleted product page. The visitor reads “page not found”; the server says 200 OK.

The status code is fine, so status-code monitoring reports it healthy.

The click lands somewhere else

A retired campaign URL that now bounces to a sign-in wall or the homepage. The visitor never sees the offer they clicked.

The 200 belongs to wherever it ended up, not to your landing page.

The page renders broken

The page comes back a fraction of its usual size: a template that lost its body, a build that shipped an empty shell, a CDN that dropped the assets. It looks wrong to a human and fine to a machine.

Every check that stops at the status code sees a healthy page.

Conversions stopped while spend carried on

The page is up, the page is whole, and nobody is converting. The funnel broke somewhere the page cannot show you.

Nothing about the page is wrong, so no page check will ever fire.

Fine on your laptop, broken for the click

A consent wall, an app-install interstitial or a geo-redirect that only fires for a phone — which is where most paid clicks come from.

A desktop check never sees it, because it never asks as a phone.

The page is refusing visitors

A 401, 403, 429 or 503 — a login wall left on by accident, a rate limiter, a WAF rule that started matching your own traffic.

This is the one an uptime monitor does see — and files as down. The fix is a rule, not a rollback.

Those six have a page of their own, with the reasoning behind each one: landing page monitoring. The most expensive is the soft 404.

PPC monitoring is not PPC reporting

They are usually sold together and they answer opposite questions. A report is assembled for somebody else to read on a schedule, and it describes a period that has already closed. Monitoring has no schedule and no reader: it produces nothing at all until something is wrong, and then it interrupts a person.

This matters more than it sounds. A tool that sends a client a beautiful report every Monday will happily let a campaign spend into a dead page from Tuesday to Friday, and then report it accurately the following Monday. The report was never wrong. It was just late by the entire length of the problem.

VampireAI does not build client-facing reports, and if a scheduled white-label PDF is what you are shopping for, one of the reporting tools is the right purchase. This is the other half of the job.

How often to monitor a PPC campaign

Set the interval by the spend rate, not by the reporting calendar. The only question that matters is how much money passes through the campaign between two checks.

Spend rate (INC-2291, three campaigns)$58.59/min

Checked every 5 minutescaps one failure at ~$293

Checked every hourcaps it at ~$3,515

Checked once a day288× the five-minute exposure

What actually happened over 213 minutes$12,480

The middle rows are arithmetic on the rate above them. The daily row is a multiple rather than a figure because a day of spend is not flat and pretending otherwise would put a decimal point on a guess. The last row is a real incident, and 207 of its 213 minutes were not detection — they were a person asleep.

What it needs from you

The destination checks need a URL and nothing else: no OAuth, no account access, no review process. The number checks need the numbers, and today there is exactly one way they arrive — you send them, one POST with a token, from a cron job, a warehouse query or twelve lines of Python.

That is worth being plain about rather than dressing up: direct pull-based connectors to the ad platforms are not built yet. What exists instead is an open door that takes any number from any system, which is why there is no OAuth queue and nothing to wait on us for. When the platform connectors land they will be a convenience on top of that door, not a replacement for it.

See it monitoring something

The demo is a working copy loaded with INC-2291 — the check that caught it six minutes in, the incident it opened, and the prepared decision that would have stopped the spend.

Open the demo →Send your own numbers