All articles

Conversion · 6 min read · March 5, 2026

Hidden Marketing Leaks: Where the Money Goes and How to Find It

Most marketing budgets lose a third of their value before a lead reaches you. The six leaks I find most often, how to spot each, and the order to fix them.

The short version

Six leaks drain most marketing budgets: paying for irrelevant searches, leads the tracking cannot see, slow landing pages, forms nobody works, reviews nobody asks for, and rankings for the wrong searches. Fix the tracking first, then the search terms report, then the rest. Each one is visible in your own accounts in minutes.

Leaks are quieter than failures

A campaign that fails is easy to see. The phone stops. A leak is different. The phone still rings, the report still shows traffic, and somewhere between the click and the booked job a third of what you paid for drains away without anyone noticing. I have found a leak on nearly every account we have taken over, and the pattern is consistent enough that I can list the usual suspects.

The fix for every one of them starts the same way: find it in your own accounts rather than in a report. This post goes through the six I find most often, what each looks like from the inside, and the order I fix them in. Kyle here, and this is the part of the job I enjoy most, because it is the fastest money a business can get back.

Leak one: paying for searches that are not customers

Open the Google Ads search terms report. It lists the searches that triggered your ads. On most accounts I inherit, a meaningful share of spend went to searches that were never going to book: job seekers, suppliers, DIY questions, cities you do not serve, services you do not offer. Google’s own guidance is that “if a search term isn’t relevant enough to the products or services you offer, add it as a negative keyword.”

On a foundation repair account in Colorado Springs, this report and a restructure around job types took roughly seventy thousand dollars a year out of the spend with no loss of inspections; the numbers are in the case study. It is the first thing I look at on any account with ads running, and it is the leak that is usually biggest.

How to read the search terms report without getting lost

The report can run to thousands of rows, which is why it goes unread. The way through it is to sort by cost and read from the top until you have covered about eighty percent of the spend, which is usually a few hundred rows at most. For each one ask a single question: would a customer who is about to book type this. Job seekers, suppliers, students, people in cities you do not serve, people looking for free advice or a product you do not sell, all get added as negatives. Anything ambiguous gets left alone for now and checked again next month against whether it produced a call.

Then look at the other end: searches that did book a job but only got shown a few times. Those are usually being starved by broader terms soaking up the budget, and they deserve their own ad group with their own landing page. That is how an account gets restructured around job types rather than around whatever the previous setup happened to be, and it is the difference between cutting waste once and keeping it out.

One warning. Do not cut aggressively in week one and then judge the account in week two. Give each round of negatives two to three weeks, watch cost per lead rather than spend, and keep notes on what you removed and why, so the next person, or you in six months, can see the reasoning.

Leak two: leads the tracking cannot see

If calls are not tracked, the account is blind to most of what it produces, and an account that cannot see calls optimises toward forms, which are a minority of leads for a service business. Google Ads conversion tracking lets you define what counts, explicitly including a phone call; if that is not set up, the system is learning from the wrong signal. In Analytics, check the key events list: if it is empty or counts page views, every report built on it is wrong in both directions.

This leak does not waste money directly. It wastes every decision made on the data, which is worse. It is why our 12 data checks start with what counts as a lead.

Leak three: the slow page between the click and the call

A paid click that lands on a page taking six seconds to load on a phone is a click that often bounces before the phone number appears. Google’s Core Web Vitals guidance sets the bar for a good experience: main content loading within 2.5 seconds, response to a tap under 200 milliseconds, and a layout that does not jump around. Test your busiest landing page on a phone connection rather than your office wifi.

The fix is usually unglamorous: images that were uploaded at full size, scripts nobody uses, a theme doing too much. We treat it as part of conversion work rather than as a separate technical project, because the point is the phone number showing up before the visitor leaves.

Speed also compounds with the first leak. If a third of your clicks are the wrong searches and a third of the right ones bounce before the page loads, the campaign is effectively paying three times the stated cost per lead. Owners rarely see it framed that way because the two problems live in different reports, one in Google Ads and one in a speed test, and nobody puts them side by side. I do, on every audit, because the combined number is the one that decides whether ads are worth running at all.

Leaks four and five: the form nobody works, the review nobody asks for

Leak four is after the lead arrives. A form fill that sits in an inbox for a day is a lead that called the next company. Check where your forms go, who sees them, and how fast the first reply goes out. This is a process leak more than a marketing one, but it sits on top of every dollar of marketing spend, so it counts.

Leak five is the review ask. BrightLocal’s Local Consumer Review Survey for 2026 found 97% of consumers read reviews for local businesses and 80% say they are likely to use a business that responds to all of its reviews. Reviews drive the prominence part of Google’s local ranking and they drive the decision once someone has found you. A business that does good work and does not ask is leaking both.

The leak that hides in a good month

Leaks are hardest to spot when things are going well. A busy spring hides the fact that half the ad spend is buying searches that would have found you anyway, and a full calendar hides the form fills that never got a reply. The owners who find leaks early are the ones who look at the reports in the good months with the same suspicion they bring to the bad ones.

A practical way to do that is to check two ratios every month regardless of how the month felt: cost per lead by campaign, and leads that became jobs by source. If either moves in the wrong direction during a good month, something is leaking underneath the volume, and it will show up fully when the season turns. Our 12 data checks post has the full list of what to read and where to find it.

Leak six: the page that ranks for the wrong thing

The last leak hides inside a green report. A site can rank well for searches nobody makes, or for your own business name, and produce no new customers. Open Search Console, sort by clicks, and read the top searches. If they are all your name, the work has not reached the people who do not know you yet. Faith wrote the full version in ranking for the wrong keywords.

The order to fix all six: tracking first, because nothing else can be judged without it. Search terms second, because it saves money this week. Then the landing page, the follow-up, the review ask, and the keyword plan. If you want a second pair of eyes on your accounts, book the free diagnostic. Finding the leak is most of what it is.

Frequently asked questions

What is the most common marketing leak?

Paying for searches that were never going to book. The Google Ads search terms report shows them, and Google’s own advice is to add irrelevant ones as negative keywords.

How do I know if my calls are being tracked?

Check whether the phone number on your site is a tracking number and whether calls show up as conversions in Google Ads or key events in Analytics. If neither, they are not.

How fast should my landing page load?

Google’s Core Web Vitals guidance asks for the main content within 2.5 seconds and a response to taps under 200 milliseconds. Test on a phone connection rather than office wifi.

Which leak should I fix first?

Tracking. Until you know what counts as a lead and calls are counted, you cannot tell whether any other fix worked.

Kyle Rogers

Kyle Rogers

Founder and Managing Director

Kyle has spent twelve years running local search and paid media for home service, legal and healthcare businesses. He builds every account plan himself, reads the map pack grids weekly, and built the automation that checks client sites and Business Profiles every day.

Keep reading

Ready to fix and scale your marketing?

We map what's broken, what's working and what will move the business next. The diagnostic is free and yours to keep.