Straight answers: problems
Why the Reports Stop Before the Bank, and the One Number That Needs a Name on It
For the owner with a team who pays someone to run the ads and still can't tell from the report if they work.
"I spend thousands on ads. How do I even know it's working?" The agency's report lands in the first week of the month, and you open it on your phone in the truck. Clicks are up, cost per conversion is down, and every arrow is green. You walk into the office and ask how many jobs came from Google last month. Your office manager says she would have to look. Your salesperson says the leads have been weak. The agency says it was a strong month, and nobody can show you what that month put in the bank.
Nobody in that conversation is lying. Everyone is looking at a different piece, and the only person expected to add the pieces together is you, who did not set up the account and cannot tell a good report from a nice one.
Below is what each way of checking really tells you, and how to put one number with one name on it, so you can see every week whether the ad money is coming back. You keep your agency, and you do not have to learn the ad account. The number has limits, and we will show you those too.
When you cannot tell, you still have to decide. You decide whether to spend more going into your busy season, whether to cut the ads when a month feels slow, and whether to renew the agency. You make those calls on a feeling, a bad week or a good story from one customer. You could be cutting the ads that fill your schedule, or paying for months for ads that bring in calls and very few jobs, and nothing in the building would tell you either way.
Then there is your own time. You are the judge of work you were never trained to judge, so you read the report twice, you ask the agency questions you are not sure how to ask, and you nod at the answers.
If you want to see what those hours cost at your own numbers, read What Is Your Hour Worth as a Business Owner?
Every one of these is a reasonable place to look, and a good company uses several of them. Each one answers a real question about the ads. None of them answers yours, which is whether the money you spent came back to the bank, and none of them makes anyone in your company responsible for finding out.
Google Ads and Meta both give you a results column, and it is useful to the person running the ads. Look at what goes into it. Google says a conversion is whatever activity the account marks as valuable, such as a purchase, a sign-up or a phone call. So someone picked what counts in your account, and it is fair to ask who picked it and what they picked. Google says a call can count by AI review of the recording or by lasting longer than a length you set, and it admits that a long call can still be a wrong number or a robocall.
All of this is honest counting by the rules of the account. Those rules can count a call that booked nothing, because the booked job lives in your company, and Google can only receive it if someone in your company sends it.
The good agencies work hard on this. AgencyAnalytics, which sells reporting tools to agencies, tells them to trim the report to 5 to 10 numbers tied to the client's goals and to skip the ones that only look good. That is the right advice, and plenty of agencies follow it. The trouble is what they can see. An agency sees the clicks, the calls and the forms. It does not see your schedule or your bank account unless someone sends them. So a fair report can still end before the part you care about. You may have had the same problem with hiring, where you pay someone and you still have to judge the work.
If so, read Why Hiring More People Didn't Fix It.
Google Analytics shows what people did on your website: where they came from, what they looked at and which actions you flagged as important. Google says attribution means assigning credit for actions along a path, and that is the job this tool does. It is good at it. It also stops when the person leaves your website. Our read, not Google's: a job booked over the phone and paid by check never shows up there unless someone sends it in from somewhere else.
This one is a real step up, and if you do not have it, it is worth a look. CallRail describes giving each campaign its own phone number, so you can see which campaigns, keywords and channels make the phone ring, with transcripts of the calls. For a company that books by phone, that closes a big gap. CallRail lists linking closed deals back to their source as a separate feature, so a ringing phone does not become a dollar figure until someone assigns a value or connects the deals. The report also has to be read by someone.
Keep asking this. It is cheap, and it catches word of mouth, which no ad tool sees. Blend, a B2B agency, points out that buyers tend to remember the most recent or most significant thing, which may not be how they first heard of you. CallRail makes the same point, and CallRail sells the alternative, so weigh that. Our own view is that it only works if your staff ask every single time and write the answer down. A customer who says "Google" has told you the channel and not the campaign. A blank box has told you nothing, and on a busy day the box stays blank.
ROAS is return on ad spend. Google describes its target as the average conversion value you want for each dollar of ads. The math is simple, and we use the same division on this page. A calculator only divides whatever you type in. Our reasoning: if the value going in is a guess at what a job is worth, the answer is a guess too, and it says nothing about what you kept after you paid for the job.
Every owner thinks of this one. It is the closest thing to a real test, and it makes sense to think of it, because an ad getting credit for a conversion and an ad causing it are different things. Google and Meta both accept that difference. Google's Conversion Lift compares people who see the ads with people who do not, but it is not open to every account and it needs a Google representative. Meta's lift tests keep one group from seeing the ads, and some can show spend minimums. The rest is our judgment. A month with the ads off can also be a slow season, a storm or a big referral, and you pay for the test in lost calls.
Even the large companies struggle here. Two economists, Lewis and Rao, studied 25 large ad experiments at major retailers and brokerages, and the typical test still left a range on the return more than 100 percentage points wide. A pause tells you something, and it will not settle the question.
If the report stops short of the bank, start with your own page. The Company Scoreboard lays out each step from your ad money to your bank, with a name on it.
Build your Company ScoreboardPicture two companies with the same agency, the same ad money and the same report. In the first, the report goes to the owner, who has twelve other things to read, so it gets a glance. When a month feels slow, he calls the agency and the agency sends more charts. In the second, the office manager owns one number: the money banked from jobs that came from the ads, for each dollar the ads cost. She looks at it every Monday. When it drops, she tells the owner what happened before he asks, and she has already called the agency with a question it can answer.
Most owners cannot tell if their ads work because nobody in the company owns that number. That is our argument from working with owners, and no study says it.
The reports stop at the platform or the form because that is as far as the people writing them can see. The rest of the trail runs through your phones, your schedule, your invoices and your bank. Google even has goals for leads that get qualified or closed in your own records, and they sit empty unless someone in your company sends the information. So the person who ties it together has to work for you.
This number has limits, and you should know them before you trust it. It does not prove the ads caused the jobs. A customer who saw your truck, heard from a neighbor and then clicked an ad shows up as an ad job. The number tells you whether the money is coming back, and it gives one person a reason to chase it when it is not. It also runs on call tracking and the answers your office writes down when a customer calls, and both go bad quietly when nobody checks them. That is why it needs a name on it. A badly set up account can still sink your ads even when someone owns the number, and then the number is what tells that person to go look.
You still will not read the ad account yourself, and you do not need to. You read one number every week, and one person can tell you why it moved.
If the ad report is one of a dozen things still waiting on your judgment, read Why Everything Still Comes Back to You, Even With a Team.
B10M works on the seven parts of a company that make it grow: leadership, focus, hiring, marketing, sales, accountability and innovation. You work on one at a time, never all seven.
The way of choosing which one is called The One-Play Method, and it has three steps: find the lowest one, move it up one level, repeat.
On this page, the lowest one is the ad number nobody owns. Moving it up one level is small on purpose: one person owns the number, it goes on the page the team looks at, and that person speaks up when it is off. You keep your agency and your ad accounts. It will not tell you what caused every job. It does mean someone other than you checks the ad money every Monday, and then you look for the next lowest one.
The Company Scoreboard puts the path from your ad money to your bank on one page. Type a few lines about your company and your team, and say where your work comes from. It shows each step with a number and a name, and it picks the two numbers to start tracking this week.
Build your Company ScoreboardHere is each way of checking by what it tells you, what it leaves out and who still can't tell whether the money came back.
| Option | What it tells you | What it leaves | Who still can't tell |
|---|---|---|---|
| The platform's own numbers | Which ads got credit for the actions the account chose to count | Whether those actions became booked jobs and money in the bank | You, and the person running the ads, unless someone sends the booked jobs in |
| The agency's monthly report | What the agency did and how its numbers moved | Your schedule and your bank, which the agency cannot see unless someone sends them | You, since the report ends before the bank |
| Google Analytics | What people did on your website | What happened once they left it, including phone bookings and checks | Anyone without a way to send those booked jobs in |
| Call tracking | Which campaign made the phone ring | Whether the call became a job, unless someone assigns a value or connects the deals | You, unless someone reads it and has the job data |
| "How did you hear about us?" | What the customer remembers, including word of mouth | Which campaign it was, and every time staff did not ask | You, on the days the box stays blank |
| A ROAS calculator | The division, done correctly | Whether the value you typed in is real, and what you kept after the job | You, if the job value was a guess |
| Turning the ads off | What happened in the months the ads were off | Slow seasons, storms and referrals, plus the calls you gave up to find out | You, because one pause cannot settle it |
| One owned number (B10M) | Whether the ad money is coming back, every week, and who is chasing it | Whether the ads caused the jobs. It also depends on clean call tracking and what your office writes down, finds nothing wrong in the account by itself and takes weeks before the habit holds | Everyone, on what caused the jobs, and you, if those answers go bad and nobody says so |
The platform's own numbers tell you which ads got credit for the actions the account chose to count. They leave out whether those actions became booked jobs and money in the bank, so you and the person running the ads can't tell unless someone sends the booked jobs in. The agency's monthly report tells you what the agency did and how its numbers moved. It leaves out your schedule and your bank, which the agency cannot see unless someone sends them, so you can't tell, since the report ends before the bank. Google Analytics tells you what people did on your website. It leaves out what happened once they left, including phone bookings and checks, so anyone without a way to send those booked jobs in can't tell. Call tracking tells you which campaign made the phone ring. It leaves out whether the call became a job, unless someone assigns a value or connects the deals, so you can't tell unless someone reads it and has the job data. Asking "How did you hear about us?" tells you what the customer remembers, including word of mouth. It leaves out which campaign it was and every time staff did not ask, so you can't tell on the days the box stays blank. A ROAS calculator does the division correctly. It leaves out whether the value you typed in is real and what you kept after the job, so you can't tell if the job value was a guess. Turning the ads off tells you what happened in the months they were off. It leaves out slow seasons, storms and referrals, plus the calls you gave up to find out, so you still can't tell, because one pause cannot settle it. One owned number tells you whether the ad money is coming back, every week, and who is chasing it. It leaves out whether the ads caused the jobs. It also depends on clean call tracking and what your office writes down, finds nothing wrong in the account by itself and takes weeks before the habit holds. So everyone still can't tell what caused the jobs, and you can't tell if those answers go bad and nobody says so.
Sample
Nate and his company are made up. He is not a client, and nothing here is a result. The page is here to show what an ad line looks like when someone owns it.
Nate owns a roofing company with an office manager, two salespeople, install crews and an agency that runs his Google ads. He used to get the agency's report and forward it to nobody. This is his one page, step by step, from the first lead to money in the bank:
The rule at the bottom of the page is the part Nate wanted. By noon every Monday, Renee sends him the ad line, what moved it and what she is doing about it, and she sends it before he asks. If calls are up and signed jobs are not, she listens to some of the calls and takes what she heard to the agency. If the ad line is off two Mondays in a row, she gets the agency on the phone and Nate hears the answer from her. If intake answers are missing, she says so first, because a missing answer makes the line wrong.
The ad line read "Not tracked yet" for the first stretch, and that was fine. Roofing jobs take a while to get paid, so the early work was making sure the office asked every caller and wrote the answer down.
Nate reads one line on Monday and asks one question. The agency still runs the ads, and now one person in his company can talk to it about jobs and money.
To see your own page instead of Nate's, type a few lines about your company, what you sell and who is on your team, and the Company Scoreboard builds it. You can type in the numbers you know and fix what it got wrong in plain words. It also shows who to call when a number is off.
Build your Company ScoreboardThe method is proven over 30 years in business and 17 years helping small companies.
Max Koz is the man behind B10M. Here is his record with clients, in his words: 45 of 50 had significant positive movement.
Those results are those owners' results. They are not a forecast for your company, and we are not promising them.
We wrote this page for two reasons, and you should know both. We want it to be useful even if we never talk: pick one person, give her the ad number and look at it together every week. And if you want help, here is where: build your Company Scoreboard first, and if you still want help after that, apply for a 15-minute call with Max.
Choosing the person and the number takes one conversation. After that she does the counting, and your part is a short look on Monday. Waiting for a calmer month means making the next few ad decisions on a feeling.
EOS works for a lot of companies. When it fades, the owner is usually the only one keeping the meetings and the scorecard alive. This asks for less: one number, one name, and a rule that she brings it to you before you ask. If that one fades too, you will see it on Monday, and it cost you very little to find out.
Keep it short. AgencyAnalytics tells agencies to trim reports to 5 to 10 numbers tied to the client's goals, and that is a good thing to ask for. Also ask what counts as a conversion in your account and who decided that. Then expect the report to stop before the bank, and send the agency your booked-job numbers so it can stop guessing.
No, and do not fire anyone over one number. It could be the ads, but it could also be the way your office answers the phone, your prices or a slow season. Upgrow, a paid search agency, is right that an agency's work can be judged in the account itself, by the change history and the search terms, and bad setup can sink good ads. A bad number tells someone to go look, and the account is one of the places to look.
We won't give you a number of months we can't back up, because it depends on how long your jobs take to close and get paid. Our judgment is that you give it long enough for ad jobs to reach the bank, and agree on that length before the clock starts.
In our judgment, your company should. The change history and the search terms anyone would want to review live in that account. Your owning it does not mean you log in. It means the login belongs to your company, and the person who owns the number knows where it is.
Yes, more slowly. An ad lead this month can become money in the bank a few months from now, so one week's number will not tell you much about that week's ads. Have her watch a longer stretch for the money and use booked jobs as the earlier sign. The number will be rough, and it is still better than going on a feeling the way you do today.
You can, and it will work until you get busy or go away, and then nobody is looking. It also means one more thing has to go through you, which is how you ended up reading ad reports in your truck.
If you want to know what happens to the rest of the company when you step away, read How Long Would Your Company Run Without You?
Checked November 3, 2026. Facts about other companies come from their own pages and are in our words. Commonly repeated numbers about wasted ad spend are not used, because they trace back to old seller surveys or to no study we could find. Everything else on this page is our judgment, not a study and not a promise.
A conversion is whatever activity the advertiser marks as valuable, such as a purchase, sign-up or phone call.
Calls can count by AI review of recordings or by lasting longer than a minimum length you set.
A long call can still be a wrong number or a robocall.
Goals exist for leads further qualified in your CRM and leads that reached a closing step, usually outside Google Ads.
The average conversion value, such as revenue, you want per dollar of ad spend.
Compares people who see ads with people who do not. Not available to all accounts and needs a Google representative.
Lift tests create test and holdout groups, and some may show spend minimums.
Attribution means assigning credit for user actions along a path.
Tells agencies to filter to 5 to 10 KPIs tied to client goals; most clients want revenue, leads or engagement rather than vanity metrics.
Unique phone numbers per campaign show which campaigns, keywords and channels produce calls and texts.
The answer to "how did you hear about us" is typically the last touchpoint. CallRail sells the alternative.
Buyers recall the most recent or significant touchpoints, which may not be how they first heard about you.
An agency can be judged in the account: change history, search terms, placements, structure. Upgrow is a paid search agency.
25 large advertising experiments at major retailers and brokerages. The median confidence interval on return was over 100 percentage points wide.
