Your paid media isn't always the problem: what to check before cutting spend

The conversation usually arrives fully formed. Cost per enquiry has crept up over two quarters. Pipeline is softer than the board expected. Sales have started saying, out loud and in meetings, that the leads are poor. And paid media is sitting there in the budget: the biggest controllable number, the easiest thing to reduce, the one lever you can pull before Friday.

Sometimes cutting is right. More often, the spend is not the thing that changed. What changed is something the spend is now revealing, or something the spend has been quietly amplifying for months while everyone looked at the media report.

This matters more than it used to. Campaigns are largely run by automated bidding now, and automated bidding is a machine for buying more of whatever you told it was good. If the definition of "good" in your account is a form fill, and a third of those form fills are students, job applicants or bots, the system will do an excellent job of finding you more of those. The report will look busy. The pipeline will not move. Cutting the budget in that situation reduces the volume of the wrong thing without fixing the instruction that produced it.

So the question to answer before the budget meeting is not "is paid media working?" It is narrower and more useful: is the media failing, or is the media accurately reporting a failure that sits somewhere else?

Separate the symptom from the cause before you touch the budget

Three symptoms dominate these conversations, and each has several plausible causes that have nothing to do with how much you are spending.

Cost per enquiry is rising. Could be genuine auction pressure. Could also be budget split across too many campaigns to ever learn, brand terms being bought by your own prospecting campaign, or a conversion being counted twice.

Volume is fine but quality is poor. Rarely a spend problem. Usually a signal problem: the system is optimising to an action that is easy to complete and commercially meaningless.

Leads have dropped off a cliff with no obvious change. Check the plumbing first. Broken tracking, a failed form integration or a site migration will produce exactly this pattern, and no amount of budget will fix it.

The checks below run roughly in order of how much damage they cause when they go wrong.

Check what the account is actually optimising towards

This is the highest-leverage check and the one most often skipped, because it looks like housekeeping rather than strategy.

Open the conversion actions in the account and list every one marked as primary. In most underperforming accounts you will find a mix of things that are not commercially equivalent: a demo request sitting alongside a guide download, a newsletter signup, a contact page view and a phone-number click. Automated bidding treats them as the same objective. It will then spend your money finding the cheapest one.

What to look for

• How many conversion actions are set to primary, and whether any of them are genuinely low-intent
• Whether any single action is being counted more than once per enquiry
• The ratio between conversions reported in the ad platform and qualified opportunities in your system. If the platform reports 400 and sales worked 30, the platform is optimising towards a number nobody in the business recognises
• Whether anything from the sales stage is fed back into the platform at all

What to do about it

Reduce primary conversions to the actions that genuinely indicate commercial intent, and demote the rest to secondary so you can still see them. Then close the loop by sending qualified-lead or opportunity stages back into the platform as offline conversions, so bidding optimises against outcomes your sales team would recognise rather than against form submissions.

This is where the single biggest performance change usually comes from, and it involves no additional budget. It is also the part of PPC and paid search management most often left as it was configured on day one.

Cloudflare is a useful illustration of the difference between buying more traffic and buying better outcomes. International paid search and display campaigns were restructured away from match-type segmentation towards product-focused audience structures, and automated workflows were built to feed real-time sales lead data into daily optimisation. Across that programme, paid search conversions rose 239%, from around 2,000 to over 7,600 quarter on quarter, cost per enquiry fell 63%, from $150 to $56, and campaign coverage expanded from 9 to 18 international markets. The instructive part is the shape of it rather than the size: activity expanded while the cost of an enquiry fell, because the system was making decisions against better information.

Follow one lead all the way through before you trust any report

Marketing teams audit campaigns far more often than they audit the path a lead takes after it is created. Yet a broken handover produces the same symptoms as a failing campaign, and costs the same money.

Run the test yourself rather than asking whether it works. Click a live ad, complete the form as a realistic prospect, then follow that record through every system it touches.

What to look for

• Does a contact appear in your platform, with the campaign, source and medium still attached rather than showing as direct or unknown
• Do hidden fields carry through, or are they arriving blank
• Does the lead route to an owner, and how long before someone contacts them
• Are duplicates being created, splitting the history of the same buyer across two records
• Has consent handling or a cookie banner change silently stopped a proportion of conversions being recorded
• If a marketing platform sits between your site and your sales system, does the same enquiry exist in both, with the same value

Two failures are extremely common and almost invisible from the media report. The first is a form that submits successfully for the user but stops creating records after an integration update, so the business concludes that demand collapsed. The second is attribution decay, where every enquiry arrives in the sales system unlabelled, so paid media appears to contribute nothing and gets cut on the strength of its own broken reporting. Both are GA4 and tracking problems rather than media ones, and both are cheap to fix once found.

Look for the money you are spending against yourself

Account structure problems are less dramatic but they accumulate.

Fragmented budgets. A large number of small campaigns, each with a fraction of the budget, rarely gathers enough data to optimise. Everything sits permanently in learning. Consolidating into fewer, better-funded structures usually improves results at identical total spend.

Overlapping keywords and campaigns. When several campaigns compete for the same query, you are bidding against yourself and splitting the performance data that the platform needs. Check the search terms report for the same terms appearing across campaigns, and use negatives to route each query to one intended home.

Automated campaign types absorbing brand demand. Broad automated campaigns will happily serve against people searching for your company name. Those clicks convert beautifully and make the campaign look excellent, while the enquiries would very likely have arrived anyway. If performance improved suddenly without a strategic change, check whether brand traffic has been quietly reclassified.

Paid and organic overlap. If you already rank first for a term and nobody else is bidding on it, paying for the same click is a defensible choice in a competitive category and a waste in an uncontested one. Test it by pausing a defined set of those terms for a fortnight and watching total enquiries rather than paid enquiries.

Too much testing is its own performance problem

Experimentation is treated as an unambiguous good, so this one rarely gets raised. It should be.

Automated bidding needs a stable period to learn. Significant edits restart that process. Teams under pressure tend to make changes weekly, sometimes several at once, which means the account spends most of its life recalibrating and very little of it performing. The results then look erratic, which prompts more changes.

Practical discipline

• Change one meaningful variable at a time and let it run long enough to produce a readable result, sized against your actual conversion volume rather than a habitual two-week window
• Keep a shared log of what changed and when, so performance shifts can be traced to a decision rather than debated
• Before running a test, decide what result would make you act. A test you would ignore either way is not worth the learning period it costs
• Separate creative tests from structural ones. Running both at once tells you almost nothing

If your monthly conversion volume is low, which is normal in considered business-to-business sales, accept that you cannot run a meaningful test every fortnight. Test the things with the largest expected effect, and leave the rest alone.

When the ad is doing its job and the page is not

A campaign can only deliver a click. Everything that determines whether the click becomes an enquiry happens afterwards, and it is often the least-maintained part of the system.

What to look for

• Broken destinations. Site migrations and content tidy-ups routinely orphan landing pages that are still being paid for. Crawl every live destination in the account, including the pages behind every extension and asset
• Redirect chains that slow loading and sometimes strip campaign parameters
• Out-of-date content. Pages still promoting last year's event, a retired price, a superseded product name or a webinar that has already happened
• Message mismatch. The ad promises a specific answer to a specific problem and the page opens with a generic homepage-style proposition, so the visitor has to work out whether they are in the right place
• Form friction. Fields that are only there because a system needs them, not because a human would volunteer them at that stage

Shelter is a helpful example because the campaigns were not the constraint. The work focused on the journeys people landed in after clicking a paid ad rather than on the media buying itself, and strengthening those journeys produced a 14% increase in earnings per click. The mechanism is the point rather than the number: the same audience and the same ads produced materially better commercial results because what happened after the click changed.

If you find issues here, fix them before you judge the media. This is conversion rate optimisation work rather than media work, and judging a campaign on a page it was never given a fair chance with will tell you nothing useful.

Reaching the wrong people efficiently is still the wrong outcome

Targeting problems are easy to mistake for budget problems because the response looks identical: spend more, get more of the same.

Questions worth asking

• Can you describe, in a sentence, the person the campaign is meant to reach and the moment they are in? If the answer is a job title and a sector, the targeting is probably broader than you think
• Are audience expansion or broad-match settings quietly widening reach beyond the intended group, and does the search terms report support that concern
• Is geography aligned to where you can actually sell and service
• In business-to-business, are you relying on proxies for seniority that do not reliably identify buyers, and is the campaign reaching people who influence the decision as well as the person who signs

Then look at the creative, which is where technical categories most often lose. Netacea's paid campaigns were attracting traffic but not consistently from the most relevant decision makers, and the complexity of the product made it difficult to explain the real-world impact of bot attacks to non-technical audiences. The response was a restructured paid media approach with new audience frameworks, plus creative built to translate a technical problem into a commercial one: protecting revenue and customer experience. The reported outcome was better quality and relevance of leads entering the pipeline, from a change to targeting and message rather than to budget.

Why your platforms disagree, and which one should govern

Almost every one of these conversations eventually reaches the moment where the ad platform, the analytics tool and the sales system all report different numbers, and the disagreement is taken as evidence that paid media is overstating itself.

They will never match, and they are not meant to. The ad platform credits the click that led to the conversion within its own window. Analytics tools apply a different model. Your sales system records whatever was captured on the form and then attributes revenue to a deal that may have closed five months later.

What to do with that

Decide, in advance, which system governs which decision. The ad platform is a reasonable guide for in-campaign optimisation. Your sales system is the only honest source for whether the enquiries became revenue. Analytics sits between the two and is useful for understanding the path, not for arbitrating the result.

Then check the lag before drawing conclusions. In a considered sale, this quarter's spend produces next quarter's opportunities and the quarter after's revenue. Comparing this month's cost against this month's closed business will make every channel with a long cycle look like a failure, and will make the last thing someone clicked look like the hero.

A two-week check before the budget meeting

Run this in order. Most of it needs analyst time rather than agency time.

1. List every primary conversion action and mark which ones a sales director would recognise as commercially meaningful.
2. Compare platform conversions to qualified opportunities in your sales system for the last full quarter. Note the ratio. That ratio, not the cost per enquiry, is your real efficiency measure.
3. Submit a test enquiry from a live ad and follow the record end to end, timing how long until a human responds.
4. Crawl every landing destination in the account for errors, redirect chains and out-of-date content.
5. Pull the search terms report for the last 90 days and look for overlap between campaigns, brand terms in prospecting campaigns, and queries that reveal an audience you cannot sell to.
6. List every change made in the last quarter and check how many learning periods you interrupted.
7. Read your own landing pages immediately after the ads that lead to them, in sequence. Mismatches are obvious this way and almost invisible in a spreadsheet.
8. Only then compare cost per opportunity against what a customer is worth to you.

When cutting spend genuinely is the right decision

Not every paid media problem is a fixable one, and defending the budget on principle is as lazy as cutting it on instinct.

Reduce or stop when the category has no meaningful search demand and you are paying to interrupt people who will never buy this way. When cost per qualified opportunity exceeds what a customer is worth over their lifetime and the gap is structural rather than seasonal. When the spend is compensating for a sales follow-up problem that nobody wants to address, because faster follow-up is cheaper than more leads. When you have run these checks, fixed what they surfaced and performance genuinely has not moved.

Reduce differently, though. An across-the-board percentage cut removes your best-performing activity in the same proportion as your worst. Cut by evidence: stop the campaigns and terms that produce enquiries your sales team does not want, and leave the ones producing opportunities alone.

What to check before you cut paid media spend

Paid media is the most measurable part of most marketing programmes, which is exactly why it absorbs the blame for problems created elsewhere. It is also the only part that responds instantly to a decision, which makes it the easiest place to look decisive.

Before you cut the budget, find out what the budget has been buying, and what the system has been told to buy more of. In a great many cases, the answer is not a media problem at all. It is a definition problem, and it is far cheaper to fix.

If the issue looks less like the cost of the click and more like what happens to the enquiry once it exists, the problem is turning interest into sales across the whole journey rather than optimising one channel in isolation.

We will run these checks against your campaigns and tell you whether the spend is the problem. If it is not, we will tell you that too, and where we think the problem actually sits.