The 100 Business Leak Test · Part 1
I tested 100 Texas businesses that pay for ads. 74 had a leak.
I checked 100 Central Texas businesses that pay for ads, the way a new customer would: from the ad, to the website, to their Google reviews. I wrote down every place a customer who was ready to buy could slip away. This page shares the totals, with no names, so any owner can hold their own business up against them.
What's leaking
Two out of three left some of their newest Google reviews unanswered. More than a quarter had a website that made it hard to reach them from a phone. And 34 of the 79 whose ads lead to a website were sending people somewhere that couldn't report back who became a customer.
What it can cost
A leak doesn't lower the ad bill. It lowers what comes back from it. One business I worked with thought its Facebook ads cost about $150 per customer. After I fixed their tracking, the real number was about $1,200. Fixing it also showed which channel brought in customers worth more over time, even though it cost more up front. They moved their budget there and scaled to more than $1 million a month in ad spend.
How I found it
I followed each business's ads to the page they send people to, then looked at it the way a new customer would: the tracking that loaded, what it took to reach them on a phone and their ten newest Google reviews. All of it is visible from the outside, and no business is named.
Where the 100 leaked, link by link
A link is drawn broken where half or more of the businesses leaked, and cracked where a third or more did.
Why I ran this test
The marketing usually isn't the problem.
For most of my career I've been brought into businesses as the marketing guy. The job was supposed to be more leads. Over and over, the bigger problem turned out to be somewhere else: ads nobody could measure, a website that lost people on a phone, leads that waited a day or two for a call back.
I call the whole path the Revenue Chain. A customer moves through it link by link, from the first time they see you to the review they leave and the friend they send. Your business is only as strong as the weakest link, and the leak is rarely where the owner is looking.
I wanted to know how common those breaks are here at home, so I picked 100 Central Texas businesses that were spending real money on ads when I checked.
Three questions I wanted answered
- When they pay for an ad, can the ad tell them who became a customer?That's the difference between knowing which ads make money and hoping.
- When someone lands on their website from a phone, how easy is it to reach them?Facebook and Instagram are mostly used on phones, so most of the people clicking these ads are on one, ready to call or ask a question.
- When happy customers talk, is anyone listening?Google reviews are where a referral gets checked before the phone rings.
Who I tested
Businesses with 39,769 Google reviews between them.
These aren't startups. I found them through Meta's public Ad Library and Google's sponsored results: 96 were running Facebook and Instagram ads and 4 were running Google ads. I kept established businesses, 93 of them with at least 50 Google reviews, and left out franchise locations, anyone I know personally, my own clients and other marketing agencies. The typical business in the test has about 180 Google reviews. Twelve have more than 1,000.
44 in the San Antonio area, 24 in the Austin area and 32 in the Hill Country and the I-35 towns between, from Kerrville and Fredericksburg to New Braunfels, San Marcos and Kyle.
How the leaks added up
26 leaked in more than one place.
Only 26 businesses held on every link I could grade. For a lot of the rest, the fix isn't one change. And a customer who never makes it through doesn't care which link broke.
Each business, by how many of its three tested links leaked. Red marks a link that was cracked or broken.
Finding 1 · Measurement
Their ads were learning from clicks, not customers.
A tracking pixel is a small piece of code on a web page. When someone clicks an ad and then calls, books or fills out a form, the pixel tells the ad platform. That's how Facebook or Google learns which people to show the ad to next, and how an owner finds out which ads actually paid off. Without it, the platform can only judge an ad by clicks and likes. The owner is left guessing.
I didn't stop at the homepage. I opened every business's active ads and followed them to the page they actually send people to, then checked that page. That's the page that has to report back.
How the 79 whose ads lead to a website graded on measurement
Someone had set up tracking, just not for the ads they were running. Ten of the cracked businesses had a Google Ads tag on the page while the money was going to Facebook and Instagram. Two of the broken ones had a pixel installed, but a cookie banner kept it from ever loading.
Fifteen never sent anyone to their website at all. Every one of their ads kept people on Facebook or Instagram, with Meta's own forms, Messenger or a call button. Meta sees those leads itself, so I didn't grade them. Six more weren't graded because I couldn't check the page their ads lead to, or their ads stopped during the test.
Bigger didn't mean better. You might expect the most established businesses to have this handled. They leaked more often than the smallest ones.
Leaking on measurement, by number of Google reviews
of businesses with fewer than 100 Google reviews (5 of 16)
with 100 to 299 reviews (15 of 31)
with 300 reviews or more (13 of 28)
Finding 2 · Capture
Websites held up best. Forms were where they cracked.
I visited every site the way a customer would on a phone, looking for a number to tap, a clear next step and a short way to ask a question. This was the strongest link in the test, and no site failed outright. When something was in the way, it was usually the form.
How all 100 graded on capture
What got in the way
- 12asked for more than five pieces of information on their main form before someone could ask a single question.
- 8had no tap-to-call link where a phone visitor would look for one.
- 4had no clear next step, like Book, Schedule or Get a quote.
- 3were slow for real visitors, by Google's own data from people who visited.
Every field is a chance to lose someone. On a phone, each one is one more thing to type before they hear back from you. The best forms in the test asked for a name, a number and a short message, and got the rest on the call.
Speed mattered less than I expected. I only counted a site as slow when Google had data from people who actually visited it, because lab tests can make a fast site look slow. By that measure, a slow site was the rarest problem in the test.
Finding 3 · Referrals
Customers kept talking. Owners kept missing some of it.
Reviews are how a referral gets checked. Someone hears your name from a friend, looks you up, and reads the newest reviews and how you answered them. I sorted every business's Google reviews by newest and counted how many of the ten most recent had a reply from the owner.
How many of their ten newest reviews the owner answered
answered every one of their ten newest reviews
answered some and left the rest
didn't answer any of them
How 98 graded on referrals, counting fresh reviews and star rating too
Size didn't help here either. Of the 33 businesses with 300 or more reviews, 12 answered fewer than half of their newest ones.
Some had gone quiet. Ten hadn't had a new Google review in two months or more, and one hadn't had one in two years, while still paying for ads to bring in new customers.
By industry
No industry got through clean.
Every group had at least one link where a third or more were leaking, and the weak link was different for each. Dental, medical and med spa practices had the best websites in the test and the worst tracking: 57% of those whose ads lead to a website were sending people to pages that couldn't report back. Home service companies had the most website problems, and 10 of their 12 came down to a long form. Everyone else, from venues to law firms, leaked most in their reviews. Builders and remodelers leaked most in their reviews, and half of them ran ads that never sent anyone to a website.
Each link shows the share of that group leaking there. Measurement counts only businesses whose ads lead to a website. A link is drawn broken where half or more of the group leaked, cracked where a third or more did. Everyone else: 8 venues and lodging, 4 law firms, 3 auto shops, 3 business services and 2 pet resorts.
If you own a business
Three things you can check this week.
None of these take more than a few minutes, and none of them need a new tool.
- Ask whoever runs your ads which ad brought in your last ten customers.If the answer is a guess, your ads can't report back, and you're paying for them to learn from clicks instead of customers.
- Pull up your own website on your phone and try to reach yourself.Count the taps and the fields. If it takes more than a couple of each, you're losing people who were ready to talk.
- Read your ten newest Google reviews.Answer the ones you missed, good and bad. Then ask your last five happy customers for one.
What comes next
Part 2: what a customer sees before they call.
Part 2 looks at trust: how these same businesses handle their bad reviews, whether a website visitor ever sees their good ones, what their Google listing tells people and what a customer finds at 9pm on a Saturday. It comes out Monday, November 23.
The rest of the chain, like how many inquiries turn into customers, what a customer is worth and how many come back, lives inside the business. In my experience that's where the expensive leaks are, and it's the work I do with owners one on one.
Why I'm doing this
I'd rather show you than tell you.
I'm Preston Cone. I grew up in Comfort and moved back to the Hill Country a few years ago. I've spent more than a decade in marketing: inside Facebook, building and running a seven-figure agency, and starting a software company. Along the way I've worked with more than 200 businesses.
Most of what I end up fixing isn't the marketing. It's what happens after someone clicks, calls or fills out a form. I wanted to show local owners what that looks like with real numbers from businesses like theirs, instead of sending one more pitch for more leads.
If something on this page sounds like your business, that's what the call is for. It's 30 minutes, with no slides. Bring what you know about your marketing and your numbers, and I'll tell you where I'd look first. If it's something your own team can fix, I'll tell you that too.
Find your own leak
Where is your business leaking?
Book a 30-minute call. Bring what you know about your marketing and your numbers, and I'll tell you where I'd look first.
How I ran the test Method, dates and grading
- When
- November 2026.
- Who
- 100 established businesses in the San Antonio area (44), the Austin area (24) and the Hill Country and I-35 towns (32): 36 home services, 32 dental, medical and med spa practices, 12 builders and remodelers, 8 venues and lodging, 4 law firms, 3 auto shops, 3 business services and 2 pet resorts.
- How they were picked
- Each one was running ads when I checked: 96 on Facebook and Instagram, found in Meta's public Ad Library, and 4 on Google, confirmed in Google's Ads Transparency Center. I matched every ad to the business by its website, not just its name. I kept established businesses, kept a mix of trades, and left out franchise locations, whose websites and tracking are usually run by a corporate office, along with businesses I know personally, current and past clients, and marketing agencies. A few of the independents have more than one location; for those I used the location in the area I was testing.
- Measurement
- I opened each business's active ads and followed them to the page they send people to. On that page I recorded which tracking loaded and fired on a normal visit, scrolling like a customer, including tags inside a tag manager. Ads that keep people on Facebook or Instagram, like Meta's own forms, Messenger or call buttons, weren't graded, because the platform sees those leads directly. Neither were businesses whose landing page I couldn't check or whose ads stopped during the test: 21 in all.
- Capture
- On each business's main website I counted form fields, tap-to-call links and next-step buttons, and pulled Google's real-visitor speed data where Google has it.
- Referrals
- I sorted each business's Google reviews by newest and counted how many of the ten newest had a reply from the owner. Two businesses couldn't be graded because their listing doesn't show reviews in a way I could sort.
- Grading
-
Link Holding Cracked Broken Measurement The page their ads send people to reports back to the platform they pay Tracking on that page, but not for the platform they're paying Basic analytics only, or nothing Capture Short form (5 fields or fewer), tap-to-call, a clear next step, not slow for real visitors One or two of those missing Three or more missing, or no way to act on a phone Referrals A new review in the last month, 4.5 stars or better, owner answers at least half of the newest ten One or two of those slipping Nothing new in about three months, under 4.0 stars, or all three slipping - What this is
- A test of 100 businesses, not a scientific study. Everything in it is something anyone could see from the outside on the days I checked. Some tracking can't be seen from the outside, like results sent to an ad platform straight from a business's own systems, so a business doing that could show here as leaking when it isn't. No business is named. Any owner in the test can ask me for their own result, privately.