Before you raise the ad budget, find out how much of it is being wasted.
Every agency you've fired, no matter the question, answered the same way: spend more. We work in the opposite order. Prove what the spend is actually buying, cut what the evidence convicts, move that money to what's working, and raise the budget last, once raising it has been earned. Until then, more budget buys more waste, and you keep renting your growth from a landlord who can turn it off with a switch.
The dashboard glows. The bank account does not.
Owners who call us about ads are almost never new to ads. They've run them, hired for them, and been shown winning reports that never turned into money they could touch.
So the first conversation tends to open with one of these.
Notice what none of these are: a traffic problem. Every one of them is a truth problem wearing a traffic costume, which is why the work starts with the numbers, not the ads.
Ads do not bring in business. They minimize waste.
That sounds backward until you sit with it. Somewhere out there, people are already looking for a way out of the exact pain your product solves, and paid traffic puts you in front of them while they look.
The platform will happily sell you that introduction all day long. What it won't tell you is how many of those introductions were never real.
What decides whether ads make you money is how much you pay for them: the fake clicks, the wrong audiences, the keywords that sound like buyers but never sign, the winning campaign that only wins on the dashboard.
One brand showed us reporting that said it was doing well. The reconciled number said otherwise: well under break-even, once the free merchandise and the promotion costs riding on each sale were counted.
Nobody had lied, exactly. The dashboard was just answering an easier question than the one that pays rent. Recalibrating what the reported return actually means, against the full cost of selling, is usually the first honest number an owner has seen.
So the promise underneath this page is narrow and checkable: every dollar you spend to be found should buy a real person, and we should be able to prove which ones did.
Raising the budget is the last lever we pull.
There are four levers in any ad account, and the industry habitually reaches for the fourth one first, because it's the easiest to sell. We pull them in order. The order is the discipline.
Four levers, pulled in sequence.
Lever two has a rule attached: nothing gets cut on a gut feeling. One client was ready to kill a category that looked expensive on the surface. When the full chain was laid out, from budget through consults to cash collected, that category was returning roughly nine times its cost. The full record is published.
Inside the first 30 days What we're actually doing while the levers move. The analyses, the competitive homework, and the quiet wins.
The account gets studied before it gets opinions. We run performance analyses across time of day, devices, and audiences, and an n-gram analysis, a word-by-word read across your search terms and ad copy, to surface patterns your reports are too aggregated to show.
Then we do the competitive homework: what your competitors' ads say and how they're positioned, what their customers' reviews praise and complain about most, what their creatives and promotions look like. And yes, we sign up for their newsletters, because we want every card on the table.
Along the way, the obvious leaks get fixed as they surface, without waiting for a grand plan. And what paid traffic teaches us rarely stays in the ad account: winning language and winning segments get tested on your website too, where the same lesson tends to lift results a second time.
One more thing runs through all four levers. When a build calls for it, campaigns, ads, and channels get segmented by intent, with copy and creative matched to how aware each buyer is: the one searching their symptom, the one comparing fixes, the one comparing you against a rival by name.
Different people need different ads, and pretending they're one average person is its own kind of waste.
Whose clicks are you paying for?
Here's an uncomfortable mechanic of paid search: your competitors can click your ads, bots can fill your forms, and the platform gets paid either way.
On one account we instrumented, a single seven-day window absorbed roughly $65,000 to $69,000 in fake clicks and produced about 17 to 18 form fills, none of them real people. The firm was sure the answer was more budget. That record is published in full.
So protection isn't an add-on here. Every engagement runs third-party fraud detection, tuned deeply through a direct relationship with the provider, sending the right signals to block the junk at the door.
And the same evidence does double duty, because a documented fake click is a refund request. We file those requests with the ad platform, and sometimes the platform pays.
One reframe before you panic about fraud in your market. Heavy fraud is usually a sign of a market worth fighting for, because nobody bothers to fake clicks where there's no money.
So the goal stays simple: be the one advertiser in a loud market who isn't paying for ghosts.
You'll never need to make the angry phone call.
You know the call: the one where you've finally had enough of not knowing, and you dial your agency ready to demand answers.
That call exists because of how most reporting works: a monthly PDF, a wall of green numbers, and silence until something breaks. We run the opposite cadence, because we'd rather be the ones reaching out.
The watch
Every edit gets a maturation window, generally a week or two, before we judge it. If performance falls off, we can see it week to week and revert the change while it's still a detail, not a quarter's story.
The conversation
The reports matter less than the conversation around them: what moved, what we believe, what we're challenging in what you told us, and what needs deciding. Assumptions built on a half-explained detail are where accounts go wrong, so we ask.
The planning
Sales, launches, and seasonal pushes get planned before they're urgent, so nobody is building a campaign in a panic ten days out. Less chaos for your team is a deliverable in its own right.
Surprises still happen in ad accounts. They always will.
The difference a process makes is that surprises arrive small, get named at the next conversation or an immediate call, and end up as refined knowledge about your market instead of as a bad quarter nobody can explain.
Your name is on everything.
Some agencies hold the ad account so leaving costs you your own history. We think that arrangement tells you everything about how they expect the relationship to end.
Here, everything we build is built in your name. It stays with you, whatever happens.
That last one matters more than the logins. A year of honest testing produces something no platform can switch off: a documented map of which messages, audiences, and promotions actually move your buyers.
The rented part of paid traffic is the clicks. The owned part is what you learn buying them, and we make sure the learning lands on your side of the table.
Who else could run this, and when they honestly should.
Some businesses honestly do better behind one of these doors, so open them all before you decide.
Another ads agency Ask one question: how do they get paid? Percentage of spend pays the agency when you spend, not when you win.
Many agencies charge a percentage of your ad spend, which means their revenue grows when your budget grows, whether or not your results do.
We refuse that model on principle: it pays the agency to recommend the one move we make only when the evidence supports it.
The other tell is in the forecasts. Agencies win accounts with generous projections, and then the dashboard has to keep the promise.
We do the reverse and deliberately underestimate our own projections, so the plan carries enough budget for real testing and the surprises land on the upside. If an agency's forecast has never once embarrassed the agency, the forecast was graded by the people who wrote it.
The platform's own reps Free advice, from the party selling you the media. The optimization score isn't measuring what you think it measures.
The account manager calling from the ad platform is polite, knowledgeable, and financially incentivized on your account's optimization percentage. That score measures how thoroughly you've adopted the platform's newest tools and automations, not whether you get the best outcome.
In practice that mostly means giving every bell and whistle a full run on your money, at a premium, while the recommendations reliably end where they always end: raise the budget.
There's no malice in any of it. You're simply getting the landlord's opinion on how much rent you should pay.
An in-house media buyer Sometimes the right call. Here's the honest test. Incredible ones exist. Many just inherit the platform's advice.
There are genuinely excellent in-house buyers, and if yours keeps learning and has a track record of scaling and maintaining accounts, that hire can be a great call. We'll tell you so.
The common failure mode is quieter: one person, one account, no basis for comparison, so the platform's account manager becomes their main advisor, and the platform's incentives quietly become your strategy.
What we bring is concentration of pattern: we're testing across far more campaigns in a shorter time than any single account allows, so we recognize the pattern earlier and can hand your buyer the alternative play, the expansion, the edge. We work alongside in-house buyers too, not only in place of them.
Doing it yourself You ran the ads. The ads ran fine. Something still leaked. The tools are not the gap. The testing volume is.
A capable founder can absolutely run their own campaigns, and plenty do respectably. The constraint is exposure, not intelligence: your account shows you one market's worth of evidence, at the pace of your own budget, while the judgment this work runs on is built from watching far more campaigns succeed and fail than any one account can show.
There's also a compounding cost nobody bills you for: every hour inside the ad manager is an hour of the founder's attention spent below its best use.
If you're going to do it yourself, do it with honest tracking and fraud protection at minimum. Those two alone prevent the most expensive mistakes.
When more traffic is the wrong prescription.
Here's where we're supposed to warn you off ads, and we're not going to. Paid traffic is almost always an answer for a product that genuinely solves a felt problem, because it's one of the fastest ways to stand in front of people already looking. When paid traffic fails, the channel is rarely the disease. One of these usually is, and each gets checked in the diagnosis before a dollar of new spend moves.
The campaign is being scored on the wrong unit. A practice selling a $15,000 machine through ads and judging the campaign on purchases will kill a working campaign, because nobody buys a machine like that from a click. That sale closes through a demo or a consultation, so the ads must be scored on qualified opportunities created. Rescore it, and the prescription often stands.
The economics can't carry the testing. The work has to be worth it: if the payback math on a new customer can't absorb a real test budget, more traffic just accelerates the loss. You'll hear that from us in the diagnosis, with the arithmetic attached, not in month four.
The leak is after the click. If the page or the follow-up loses the buyers the ads deliver, fixing traffic first just pays to send more buyers to the same leak. Fixing the leak is conversion work. It lives under this roof, and the effort shifts with one conversation.
The operation can't absorb the win. Success has logistics. We've watched a sold-out best seller trigger a reorder the cash flow wasn't ready for, then sit out of stock while the market share it had just won was handed straight back. If scaling would break fulfillment, capacity, or cash, we scale slower on purpose and tell you why.
Read the full files, not the highlight reel.
Every number below sits inside a published record, de-identified like everything we publish, so you can judge the frame around it yourself.
More clicks meant more cases, so the fix had to be more budget.
One seven-day window absorbed roughly $65,000 to $69,000 in fake clicks and produced about 17 to 18 form fills, none of them real.
Fraud protection went on, the waste stopped being funded, and the budget stopped being the suspect.
The way up was volume: spend more, rank higher, answer more calls.
The account was bleeding waste first. Cleaned up, the same channel could run on arithmetic instead of luck.
Roughly $100k a month in search spend produced 80 to 90 signed cases a month at about $1,100 to $1,250 each, for over a year, until the partners sold the practice.
The dashboard said patients were booking for almost nothing.
Refreshes and phone clicks were being counted as bookings. The real cost per booked patient was around $250.
Only real booked visits got counted, and the honest number settled around $76.
Results from specific engagements, measured by us and told with their context intact. They are records, not promises of what your account will do.
First the diagnosis. Then the prescription. Then the spend.
Nothing on this page is sold at the door. The examination comes first, the prescription decides what runs, and that order holds for everyone.
Paid Traffic
Your channels run by our hands, in your name. Waste cut first. Budget raised last.
Where every engagement starts, and the only way in. A senior diagnosis on your real numbers that tells you exactly what we see and exactly what we would do if you hired us. The findings are yours either way: run them with us, with your own team, or with nobody.
Only if the diagnosis prescribes it. One paid channel, run end to end: strategy, copy, creative, testing, analytics, and fraud protection, with a reasonable volume of creative production included. Two channels, Google and Meta together, run $15,000 a month, because each channel gets its own full focus.
Ad spend is separate, in your accounts, and for a single channel a starting commitment around $10,000 to $15,000 a month is typical. Complex product catalogs are scoped separately.
We never charge a percentage of ad spend. The incentive runs backward.
- Your name on the accounts, the pixel, the audiences, the tracking.
- Fraud protection on by default, with documented fakes filed for refunds.
- Projections deliberately underestimated, so surprises land on the upside.
- Every tested rule about your market documented, dated, and yours.
When it works What happens when the account starts winning? Budgets grow on proof, and channels multiply on purpose.
Two things, in order. First, the budget conversation changes sides: instead of an agency asking you for more, the evidence decides how fast the budget grows. Clients have started with a few thousand a month and grown their spend several times over once the proof arrived, and one started a single channel small and recently ran it at many times the opening commitment, consistently.
Second, winning gets exported. A campaign proven on one platform gets cloned where the same buyer is cheaper to reach, often at a fraction of the original setup cost, and the strongest lessons get carried into new channels deliberately, not as a hedge but as an expansion.
One caution stays taped to the dashboard the whole way: platforms sell success back at a premium. As your account wins, your costs drift up, which is exactly why the scaling decision runs on reconciled numbers and payback math instead of momentum.
Asked before anyone hands over an ad account.
What does paid traffic management cost?
With us, the entry point is a $2,500 diagnosis that decides whether paid traffic is even the right treatment. If it is, managing one paid channel runs $10,000 a month, and two channels, such as Google and Meta together, run $15,000, with each channel getting its own dedicated strategy, copy, creative, testing, and fraud protection.
Ad spend is separate and stays in your own accounts; for one channel, plan on a starting commitment of roughly $10,000 to $15,000 a month. We don't charge a percentage of ad spend.
Why don't you charge a percentage of ad spend like other agencies?
Because the incentive runs backward. An agency paid on spend earns more every time it convinces you to raise the budget, which is precisely the recommendation that should require the most evidence. A flat retainer means the only way we win is the account performing, and it keeps our advice clean when the right answer is to spend less.
The one honest exception: accounts with enormous active catalogs, hundreds or thousands of products, carry real extra workload and get scoped separately.
How do I know the leads are real people?
Two disciplines, both on by default. First, ruthless attribution: a paid lead only counts as a paid lead if it verifiably came through the ad platform, because generous counting is how agencies claim credit for leads you'd have gotten anyway.
Second, fraud detection tuned through a direct relationship with the provider checks the traffic itself, blocks the junk, and documents what it caught, and the documentation doubles as a refund request to the platform. Sometimes the platform pays it.
We tried ads twice and lost money both times. Why would this go differently?
Usually one of three things was true, and none of them is "ads don't work for your business." The tracking was never verified, so decisions ran on fiction. Or the campaign was scored on the wrong unit, judged on instant purchases when your sale actually closes through a call or a demo. Or everything ran on one big bet instead of isolated tests, so one wrong guess took the whole budget with it.
The diagnosis checks all three against your actual account history before we recommend anything, and if the honest answer is that the economics can't carry it yet, you'll hear that instead of a pitch.
My current agency's numbers look great. What would you even fix?
Maybe nothing, and the diagnosis will say so if so. But dashboards usually flatter, because they rarely subtract refunds, cancellations, fulfillment costs, or the promotions that bought the revenue.
We reconcile the reported return against what it truly costs you to sell, and we check what the report isn't showing: whether campaigns are being scored on the right unit and whether any of the volume is fake. Owners are sometimes losing money inside numbers that look great, and sometimes sitting on campaigns performing better than the reports show. Both are worth knowing.
What happens when costs spike or the platform changes the rules?
We assume it will happen, because it always does: platforms sell success back at a premium, algorithms shift, policies change. The account is watched week to week, so a spike gets caught while it's small and recent changes can be reverted while they're still reversible. The deeper protection is structural: isolated tests instead of one big bet, proven campaigns cloned to additional channels so no single landlord owns your pipeline, and documented learning that transfers to whatever channel comes next.
Do I own the ad accounts if we part ways?
Yes. Your name is on everything from day one: the ad accounts, the pixel, which is the platform's memory of your buyers, its audiences, the analytics, the tag manager, any server-side tracking we build, and the documented results of every test we ran.
That's the agreement, stated up front, because an agency that holds your account hostage is telling you how it expects the relationship to end.
What happens in the first month?
The levers, in order. The tracking gets verified and segmented so the numbers tell the truth, which sometimes means scoping a server-side build. Proven waste gets cut, after due diligence confirms it's truly waste and not a quiet contributor to sales elsewhere. That money moves to what's demonstrably working.
Meanwhile the account gets studied: performance analyses by time, device, and audience, a word-pattern read across your search terms, and competitive homework on your rivals' ads, reviews, and promotions. Budget increases come later, once they're arithmetic instead of hope.
When would you tell me not to buy paid traffic?
When the payback math can't carry a real test budget, when your operation can't absorb the win, whether that's fulfillment, sales capacity, or the cash to reorder inventory, or when the leak is after the click and the honest first move is fixing the page or the follow-up.
Each of those is checked in the diagnosis. If one of them is your situation, you'll get that finding and the fix order instead of a traffic pitch, and the diagnosis is yours to act on with anyone.
Stop guessing what the spend is doing. Start knowing.
We've met owners who checked their dashboards three times a day, because not knowing felt dangerous. A year into an honest account, that habit fades on its own. The numbers reconcile, the surprises arrive small and named, and the budget conversation becomes something you decide from evidence instead of something done to you.
You also get sharper at this yourself, because the point of every report we walk through is that you make better decisions with us in the room, and eventually without us.
The first question this page raised is still open: how much of your current spend is waste? That number exists, it's sitting in your account right now, and the Second Opinion is how it gets found. The findings are yours to keep, whoever you run them with.